Merge commit '04d14c9b68b03e8dbc6e3003c1ee06892dd32576'
authorMichael Niedermayer <michaelni@gmx.at>
Wed, 8 Oct 2014 19:58:42 +0000 (21:58 +0200)
committerMichael Niedermayer <michaelni@gmx.at>
Wed, 8 Oct 2014 20:00:09 +0000 (22:00 +0200)
* commit '04d14c9b68b03e8dbc6e3003c1ee06892dd32576':
  vc1: Split the decoder in components

Conflicts:
libavcodec/Makefile
libavcodec/vc1dec.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>
1  2 
libavcodec/Makefile
libavcodec/vc1.h
libavcodec/vc1_block.c
libavcodec/vc1_loopfilter.c
libavcodec/vc1_mc.c
libavcodec/vc1_pred.c
libavcodec/vc1_pred.h
libavcodec/vc1data.h
libavcodec/vc1dec.c

@@@ -473,9 -390,10 +473,11 @@@ OBJS-$(CONFIG_V410_ENCODER)            
  OBJS-$(CONFIG_V210X_DECODER)           += v210x.o
  OBJS-$(CONFIG_VB_DECODER)              += vb.o
  OBJS-$(CONFIG_VBLE_DECODER)            += vble.o
- OBJS-$(CONFIG_VC1_DECODER)             += vc1dec.o vc1.o vc1data.o vc1dsp.o \
+ OBJS-$(CONFIG_VC1_DECODER)             += vc1dec.o vc1_block.o vc1_loopfilter.o \
+                                           vc1_mc.o vc1_pred.o vc1.o vc1data.o \
+                                           vc1dsp.o \
 -                                          msmpeg4dec.o msmpeg4.o msmpeg4data.o
 +                                          msmpeg4dec.o msmpeg4.o msmpeg4data.o \
 +                                          wmv2dsp.o
  OBJS-$(CONFIG_VCR1_DECODER)            += vcr1.o
  OBJS-$(CONFIG_VMDAUDIO_DECODER)        += vmdaudio.o
  OBJS-$(CONFIG_VMDVIDEO_DECODER)        += vmdvideo.o
Simple merge
index 0000000,acd5145..6f84228
mode 000000,100644..100644
--- /dev/null
@@@ -1,0 -1,3054 +1,3052 @@@
 - * This file is part of Libav.
+ /*
+  * VC-1 and WMV3 decoder
+  * Copyright (c) 2011 Mashiat Sarker Shakkhar
+  * Copyright (c) 2006-2007 Konstantin Shishkov
+  * Partly based on vc9.c (c) 2005 Anonymous, Alex Beregszaszi, Michael Niedermayer
+  *
 - * Libav is free software; you can redistribute it and/or
++ * This file is part of FFmpeg.
+  *
 - * Libav is distributed in the hope that it will be useful,
++ * FFmpeg is free software; you can redistribute it and/or
+  * modify it under the terms of the GNU Lesser General Public
+  * License as published by the Free Software Foundation; either
+  * version 2.1 of the License, or (at your option) any later version.
+  *
 - * License along with Libav; if not, write to the Free Software
++ * FFmpeg is distributed in the hope that it will be useful,
+  * but WITHOUT ANY WARRANTY; without even the implied warranty of
+  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+  * Lesser General Public License for more details.
+  *
+  * You should have received a copy of the GNU Lesser General Public
 -/**
 - * Imode types
 - * @{
 - */
 -enum Imode {
 -    IMODE_RAW,
 -    IMODE_NORM2,
 -    IMODE_DIFF2,
 -    IMODE_NORM6,
 -    IMODE_DIFF6,
 -    IMODE_ROWSKIP,
 -    IMODE_COLSKIP
 -};
 -/** @} */ //imode defines
++ * License along with FFmpeg; if not, write to the Free Software
+  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+  */
+ /**
+  * @file
+  * VC-1 and WMV3 block decoding routines
+  */
+ #include "avcodec.h"
+ #include "mpegutils.h"
+ #include "mpegvideo.h"
+ #include "msmpeg4data.h"
+ #include "unary.h"
+ #include "vc1.h"
+ #include "vc1_pred.h"
+ #include "vc1acdata.h"
+ #include "vc1data.h"
+ #define MB_INTRA_VLC_BITS 9
+ #define DC_VLC_BITS 9
+ // offset tables for interlaced picture MVDATA decoding
+ static const int offset_table1[9] = {  0,  1,  2,  4,  8, 16, 32,  64, 128 };
+ static const int offset_table2[9] = {  0,  1,  3,  7, 15, 31, 63, 127, 255 };
+ /***********************************************************************/
+ /**
+  * @name VC-1 Bitplane decoding
+  * @see 8.7, p56
+  * @{
+  */
 -    int16_t *dc_val;
+ static void init_block_index(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     ff_init_block_index(s);
+     if (v->field_mode && !(v->second_field ^ v->tff)) {
+         s->dest[0] += s->current_picture_ptr->f->linesize[0];
+         s->dest[1] += s->current_picture_ptr->f->linesize[1];
+         s->dest[2] += s->current_picture_ptr->f->linesize[2];
+     }
+ }
+ /** @} */ //Bitplane group
+ static void vc1_put_signed_blocks_clamped(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     int topleft_mb_pos, top_mb_pos;
+     int stride_y, fieldtx = 0;
+     int v_dist;
+     /* The put pixels loop is always one MB row behind the decoding loop,
+      * because we can only put pixels when overlap filtering is done, and
+      * for filtering of the bottom edge of a MB, we need the next MB row
+      * present as well.
+      * Within the row, the put pixels loop is also one MB col behind the
+      * decoding loop. The reason for this is again, because for filtering
+      * of the right MB edge, we need the next MB present. */
+     if (!s->first_slice_line) {
+         if (s->mb_x) {
+             topleft_mb_pos = (s->mb_y - 1) * s->mb_stride + s->mb_x - 1;
+             if (v->fcm == ILACE_FRAME)
+                 fieldtx = v->fieldtx_plane[topleft_mb_pos];
+             stride_y       = s->linesize << fieldtx;
+             v_dist         = (16 - fieldtx) >> (fieldtx == 0);
+             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][0],
+                                               s->dest[0] - 16 * s->linesize - 16,
+                                               stride_y);
+             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][1],
+                                               s->dest[0] - 16 * s->linesize - 8,
+                                               stride_y);
+             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][2],
+                                               s->dest[0] - v_dist * s->linesize - 16,
+                                               stride_y);
+             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][3],
+                                               s->dest[0] - v_dist * s->linesize - 8,
+                                               stride_y);
+             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][4],
+                                               s->dest[1] - 8 * s->uvlinesize - 8,
+                                               s->uvlinesize);
+             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][5],
+                                               s->dest[2] - 8 * s->uvlinesize - 8,
+                                               s->uvlinesize);
+         }
+         if (s->mb_x == s->mb_width - 1) {
+             top_mb_pos = (s->mb_y - 1) * s->mb_stride + s->mb_x;
+             if (v->fcm == ILACE_FRAME)
+                 fieldtx = v->fieldtx_plane[top_mb_pos];
+             stride_y   = s->linesize << fieldtx;
+             v_dist     = fieldtx ? 15 : 8;
+             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][0],
+                                               s->dest[0] - 16 * s->linesize,
+                                               stride_y);
+             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][1],
+                                               s->dest[0] - 16 * s->linesize + 8,
+                                               stride_y);
+             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][2],
+                                               s->dest[0] - v_dist * s->linesize,
+                                               stride_y);
+             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][3],
+                                               s->dest[0] - v_dist * s->linesize + 8,
+                                               stride_y);
+             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][4],
+                                               s->dest[1] - 8 * s->uvlinesize,
+                                               s->uvlinesize);
+             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][5],
+                                               s->dest[2] - 8 * s->uvlinesize,
+                                               s->uvlinesize);
+         }
+     }
+ #define inc_blk_idx(idx) do { \
+         idx++; \
+         if (idx >= v->n_allocated_blks) \
+             idx = 0; \
+     } while (0)
+     inc_blk_idx(v->topleft_blk_idx);
+     inc_blk_idx(v->top_blk_idx);
+     inc_blk_idx(v->left_blk_idx);
+     inc_blk_idx(v->cur_blk_idx);
+ }
+ /***********************************************************************/
+ /**
+  * @name VC-1 Block-level functions
+  * @see 7.1.4, p91 and 8.1.1.7, p(1)04
+  * @{
+  */
+ /**
+  * @def GET_MQUANT
+  * @brief Get macroblock-level quantizer scale
+  */
+ #define GET_MQUANT()                                           \
+     if (v->dquantfrm) {                                        \
+         int edges = 0;                                         \
+         if (v->dqprofile == DQPROFILE_ALL_MBS) {               \
+             if (v->dqbilevel) {                                \
+                 mquant = (get_bits1(gb)) ? v->altpq : v->pq;   \
+             } else {                                           \
+                 mqdiff = get_bits(gb, 3);                      \
+                 if (mqdiff != 7)                               \
+                     mquant = v->pq + mqdiff;                   \
+                 else                                           \
+                     mquant = get_bits(gb, 5);                  \
+             }                                                  \
+         }                                                      \
+         if (v->dqprofile == DQPROFILE_SINGLE_EDGE)             \
+             edges = 1 << v->dqsbedge;                          \
+         else if (v->dqprofile == DQPROFILE_DOUBLE_EDGES)       \
+             edges = (3 << v->dqsbedge) % 15;                   \
+         else if (v->dqprofile == DQPROFILE_FOUR_EDGES)         \
+             edges = 15;                                        \
+         if ((edges&1) && !s->mb_x)                             \
+             mquant = v->altpq;                                 \
+         if ((edges&2) && s->first_slice_line)                  \
+             mquant = v->altpq;                                 \
+         if ((edges&4) && s->mb_x == (s->mb_width - 1))         \
+             mquant = v->altpq;                                 \
+         if ((edges&8) && s->mb_y == (s->mb_height - 1))        \
+             mquant = v->altpq;                                 \
+         if (!mquant || mquant > 31) {                          \
+             av_log(v->s.avctx, AV_LOG_ERROR,                   \
+                    "Overriding invalid mquant %d\n", mquant);  \
+             mquant = 1;                                        \
+         }                                                      \
+     }
+ /**
+  * @def GET_MVDATA(_dmv_x, _dmv_y)
+  * @brief Get MV differentials
+  * @see MVDATA decoding from 8.3.5.2, p(1)20
+  * @param _dmv_x Horizontal differential for decoded MV
+  * @param _dmv_y Vertical differential for decoded MV
+  */
+ #define GET_MVDATA(_dmv_x, _dmv_y)                                      \
+     index = 1 + get_vlc2(gb, ff_vc1_mv_diff_vlc[s->mv_table_index].table, \
+                          VC1_MV_DIFF_VLC_BITS, 2);                      \
+     if (index > 36) {                                                   \
+         mb_has_coeffs = 1;                                              \
+         index -= 37;                                                    \
+     } else                                                              \
+         mb_has_coeffs = 0;                                              \
+     s->mb_intra = 0;                                                    \
+     if (!index) {                                                       \
+         _dmv_x = _dmv_y = 0;                                            \
+     } else if (index == 35) {                                           \
+         _dmv_x = get_bits(gb, v->k_x - 1 + s->quarter_sample);          \
+         _dmv_y = get_bits(gb, v->k_y - 1 + s->quarter_sample);          \
+     } else if (index == 36) {                                           \
+         _dmv_x = 0;                                                     \
+         _dmv_y = 0;                                                     \
+         s->mb_intra = 1;                                                \
+     } else {                                                            \
+         index1 = index % 6;                                             \
+         if (!s->quarter_sample && index1 == 5) val = 1;                 \
+         else                                   val = 0;                 \
+         if (size_table[index1] - val > 0)                               \
+             val = get_bits(gb, size_table[index1] - val);               \
+         else                                   val = 0;                 \
+         sign = 0 - (val&1);                                             \
+         _dmv_x = (sign ^ ((val>>1) + offset_table[index1])) - sign;     \
+                                                                         \
+         index1 = index / 6;                                             \
+         if (!s->quarter_sample && index1 == 5) val = 1;                 \
+         else                                   val = 0;                 \
+         if (size_table[index1] - val > 0)                               \
+             val = get_bits(gb, size_table[index1] - val);               \
+         else                                   val = 0;                 \
+         sign = 0 - (val & 1);                                           \
+         _dmv_y = (sign ^ ((val >> 1) + offset_table[index1])) - sign;   \
+     }
+ static av_always_inline void get_mvdata_interlaced(VC1Context *v, int *dmv_x,
+                                                    int *dmv_y, int *pred_flag)
+ {
+     int index, index1;
+     int extend_x = 0, extend_y = 0;
+     GetBitContext *gb = &v->s.gb;
+     int bits, esc;
+     int val, sign;
+     const int* offs_tab;
+     if (v->numref) {
+         bits = VC1_2REF_MVDATA_VLC_BITS;
+         esc  = 125;
+     } else {
+         bits = VC1_1REF_MVDATA_VLC_BITS;
+         esc  = 71;
+     }
+     switch (v->dmvrange) {
+     case 1:
+         extend_x = 1;
+         break;
+     case 2:
+         extend_y = 1;
+         break;
+     case 3:
+         extend_x = extend_y = 1;
+         break;
+     }
+     index = get_vlc2(gb, v->imv_vlc->table, bits, 3);
+     if (index == esc) {
+         *dmv_x = get_bits(gb, v->k_x);
+         *dmv_y = get_bits(gb, v->k_y);
+         if (v->numref) {
+             if (pred_flag) {
+                 *pred_flag = *dmv_y & 1;
+                 *dmv_y     = (*dmv_y + *pred_flag) >> 1;
+             } else {
+                 *dmv_y     = (*dmv_y + (*dmv_y & 1)) >> 1;
+             }
+         }
+     }
+     else {
++        av_assert0(index < esc);
+         if (extend_x)
+             offs_tab = offset_table2;
+         else
+             offs_tab = offset_table1;
+         index1 = (index + 1) % 9;
+         if (index1 != 0) {
+             val    = get_bits(gb, index1 + extend_x);
+             sign   = 0 -(val & 1);
+             *dmv_x = (sign ^ ((val >> 1) + offs_tab[index1])) - sign;
+         } else
+             *dmv_x = 0;
+         if (extend_y)
+             offs_tab = offset_table2;
+         else
+             offs_tab = offset_table1;
+         index1 = (index + 1) / 9;
+         if (index1 > v->numref) {
+             val    = get_bits(gb, (index1 + (extend_y << v->numref)) >> v->numref);
+             sign   = 0 - (val & 1);
+             *dmv_y = (sign ^ ((val >> 1) + offs_tab[index1 >> v->numref])) - sign;
+         } else
+             *dmv_y = 0;
+         if (v->numref && pred_flag)
+             *pred_flag = index1 & 1;
+     }
+ }
+ /** Reconstruct motion vector for B-frame and do motion compensation
+  */
+ static inline void vc1_b_mc(VC1Context *v, int dmv_x[2], int dmv_y[2],
+                             int direct, int mode)
+ {
+     if (direct) {
+         ff_vc1_mc_1mv(v, 0);
+         ff_vc1_interp_mc(v);
+         return;
+     }
+     if (mode == BMV_TYPE_INTERPOLATED) {
+         ff_vc1_mc_1mv(v, 0);
+         ff_vc1_interp_mc(v);
+         return;
+     }
+     ff_vc1_mc_1mv(v, (mode == BMV_TYPE_BACKWARD));
+ }
+ /** Get predicted DC value for I-frames only
+  * prediction dir: left=0, top=1
+  * @param s MpegEncContext
+  * @param overlap flag indicating that overlap filtering is used
+  * @param pq integer part of picture quantizer
+  * @param[in] n block index in the current MB
+  * @param dc_val_ptr Pointer to DC predictor
+  * @param dir_ptr Prediction direction for use in AC prediction
+  */
+ static inline int vc1_i_pred_dc(MpegEncContext *s, int overlap, int pq, int n,
+                                 int16_t **dc_val_ptr, int *dir_ptr)
+ {
+     int a, b, c, wrap, pred, scale;
+     int16_t *dc_val;
+     static const uint16_t dcpred[32] = {
+         -1, 1024,  512,  341,  256,  205,  171,  146,  128,
+              114,  102,   93,   85,   79,   73,   68,   64,
+               60,   57,   54,   51,   49,   47,   45,   43,
+               41,   39,   38,   37,   35,   34,   33
+     };
+     /* find prediction - wmv3_dc_scale always used here in fact */
+     if (n < 4) scale = s->y_dc_scale;
+     else       scale = s->c_dc_scale;
+     wrap   = s->block_wrap[n];
+     dc_val = s->dc_val[0] + s->block_index[n];
+     /* B A
+      * C X
+      */
+     c = dc_val[ - 1];
+     b = dc_val[ - 1 - wrap];
+     a = dc_val[ - wrap];
+     if (pq < 9 || !overlap) {
+         /* Set outer values */
+         if (s->first_slice_line && (n != 2 && n != 3))
+             b = a = dcpred[scale];
+         if (s->mb_x == 0 && (n != 1 && n != 3))
+             b = c = dcpred[scale];
+     } else {
+         /* Set outer values */
+         if (s->first_slice_line && (n != 2 && n != 3))
+             b = a = 0;
+         if (s->mb_x == 0 && (n != 1 && n != 3))
+             b = c = 0;
+     }
+     if (abs(a - b) <= abs(b - c)) {
+         pred     = c;
+         *dir_ptr = 1; // left
+     } else {
+         pred     = a;
+         *dir_ptr = 0; // top
+     }
+     /* update predictor */
+     *dc_val_ptr = &dc_val[0];
+     return pred;
+ }
+ /** Get predicted DC value
+  * prediction dir: left=0, top=1
+  * @param s MpegEncContext
+  * @param overlap flag indicating that overlap filtering is used
+  * @param pq integer part of picture quantizer
+  * @param[in] n block index in the current MB
+  * @param a_avail flag indicating top block availability
+  * @param c_avail flag indicating left block availability
+  * @param dc_val_ptr Pointer to DC predictor
+  * @param dir_ptr Prediction direction for use in AC prediction
+  */
+ static inline int ff_vc1_pred_dc(MpegEncContext *s, int overlap, int pq, int n,
+                               int a_avail, int c_avail,
+                               int16_t **dc_val_ptr, int *dir_ptr)
+ {
+     int a, b, c, wrap, pred;
+     int16_t *dc_val;
+     int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     int q1, q2 = 0;
+     int dqscale_index;
+     wrap = s->block_wrap[n];
+     dc_val = s->dc_val[0] + s->block_index[n];
+     /* B A
+      * C X
+      */
+     c = dc_val[ - 1];
+     b = dc_val[ - 1 - wrap];
+     a = dc_val[ - wrap];
+     /* scale predictors if needed */
+     q1 = s->current_picture.qscale_table[mb_pos];
+     dqscale_index = s->y_dc_scale_table[q1] - 1;
+     if (dqscale_index < 0)
+         return 0;
+     if (c_avail && (n != 1 && n != 3)) {
+         q2 = s->current_picture.qscale_table[mb_pos - 1];
+         if (q2 && q2 != q1)
+             c = (c * s->y_dc_scale_table[q2] * ff_vc1_dqscale[dqscale_index] + 0x20000) >> 18;
+     }
+     if (a_avail && (n != 2 && n != 3)) {
+         q2 = s->current_picture.qscale_table[mb_pos - s->mb_stride];
+         if (q2 && q2 != q1)
+             a = (a * s->y_dc_scale_table[q2] * ff_vc1_dqscale[dqscale_index] + 0x20000) >> 18;
+     }
+     if (a_avail && c_avail && (n != 3)) {
+         int off = mb_pos;
+         if (n != 1)
+             off--;
+         if (n != 2)
+             off -= s->mb_stride;
+         q2 = s->current_picture.qscale_table[off];
+         if (q2 && q2 != q1)
+             b = (b * s->y_dc_scale_table[q2] * ff_vc1_dqscale[dqscale_index] + 0x20000) >> 18;
+     }
+     if (a_avail && c_avail) {
+         if (abs(a - b) <= abs(b - c)) {
+             pred     = c;
+             *dir_ptr = 1; // left
+         } else {
+             pred     = a;
+             *dir_ptr = 0; // top
+         }
+     } else if (a_avail) {
+         pred     = a;
+         *dir_ptr = 0; // top
+     } else if (c_avail) {
+         pred     = c;
+         *dir_ptr = 1; // left
+     } else {
+         pred     = 0;
+         *dir_ptr = 1; // left
+     }
+     /* update predictor */
+     *dc_val_ptr = &dc_val[0];
+     return pred;
+ }
+ /** @} */ // Block group
+ /**
+  * @name VC1 Macroblock-level functions in Simple/Main Profiles
+  * @see 7.1.4, p91 and 8.1.1.7, p(1)04
+  * @{
+  */
+ static inline int vc1_coded_block_pred(MpegEncContext * s, int n,
+                                        uint8_t **coded_block_ptr)
+ {
+     int xy, wrap, pred, a, b, c;
+     xy   = s->block_index[n];
+     wrap = s->b8_stride;
+     /* B C
+      * A X
+      */
+     a = s->coded_block[xy - 1       ];
+     b = s->coded_block[xy - 1 - wrap];
+     c = s->coded_block[xy     - wrap];
+     if (b == c) {
+         pred = a;
+     } else {
+         pred = c;
+     }
+     /* store value */
+     *coded_block_ptr = &s->coded_block[xy];
+     return pred;
+ }
+ /**
+  * Decode one AC coefficient
+  * @param v The VC1 context
+  * @param last Last coefficient
+  * @param skip How much zero coefficients to skip
+  * @param value Decoded AC coefficient value
+  * @param codingset set of VLC to decode data
+  * @see 8.1.3.4
+  */
+ static void vc1_decode_ac_coeff(VC1Context *v, int *last, int *skip,
+                                 int *value, int codingset)
+ {
+     GetBitContext *gb = &v->s.gb;
+     int index, escape, run = 0, level = 0, lst = 0;
+     index = get_vlc2(gb, ff_vc1_ac_coeff_table[codingset].table, AC_VLC_BITS, 3);
+     if (index != ff_vc1_ac_sizes[codingset] - 1) {
+         run   = vc1_index_decode_table[codingset][index][0];
+         level = vc1_index_decode_table[codingset][index][1];
+         lst   = index >= vc1_last_decode_table[codingset] || get_bits_left(gb) < 0;
+         if (get_bits1(gb))
+             level = -level;
+     } else {
+         escape = decode210(gb);
+         if (escape != 2) {
+             index = get_vlc2(gb, ff_vc1_ac_coeff_table[codingset].table, AC_VLC_BITS, 3);
+             run   = vc1_index_decode_table[codingset][index][0];
+             level = vc1_index_decode_table[codingset][index][1];
+             lst   = index >= vc1_last_decode_table[codingset];
+             if (escape == 0) {
+                 if (lst)
+                     level += vc1_last_delta_level_table[codingset][run];
+                 else
+                     level += vc1_delta_level_table[codingset][run];
+             } else {
+                 if (lst)
+                     run += vc1_last_delta_run_table[codingset][level] + 1;
+                 else
+                     run += vc1_delta_run_table[codingset][level] + 1;
+             }
+             if (get_bits1(gb))
+                 level = -level;
+         } else {
+             int sign;
+             lst = get_bits1(gb);
+             if (v->s.esc3_level_length == 0) {
+                 if (v->pq < 8 || v->dquantfrm) { // table 59
+                     v->s.esc3_level_length = get_bits(gb, 3);
+                     if (!v->s.esc3_level_length)
+                         v->s.esc3_level_length = get_bits(gb, 2) + 8;
+                 } else { // table 60
+                     v->s.esc3_level_length = get_unary(gb, 1, 6) + 2;
+                 }
+                 v->s.esc3_run_length = 3 + get_bits(gb, 2);
+             }
+             run   = get_bits(gb, v->s.esc3_run_length);
+             sign  = get_bits1(gb);
+             level = get_bits(gb, v->s.esc3_level_length);
+             if (sign)
+                 level = -level;
+         }
+     }
+     *last  = lst;
+     *skip  = run;
+     *value = level;
+ }
+ /** Decode intra block in intra frames - should be faster than decode_intra_block
+  * @param v VC1Context
+  * @param block block to decode
+  * @param[in] n subblock index
+  * @param coded are AC coeffs present or not
+  * @param codingset set of VLC to decode data
+  */
+ static int vc1_decode_i_block(VC1Context *v, int16_t block[64], int n,
+                               int coded, int codingset)
+ {
+     GetBitContext *gb = &v->s.gb;
+     MpegEncContext *s = &v->s;
+     int dc_pred_dir = 0; /* Direction of the DC prediction used */
+     int i;
+     int16_t *dc_val;
+     int16_t *ac_val, *ac_val2;
+     int dcdiff;
+     /* Get DC differential */
+     if (n < 4) {
+         dcdiff = get_vlc2(&s->gb, ff_msmp4_dc_luma_vlc[s->dc_table_index].table, DC_VLC_BITS, 3);
+     } else {
+         dcdiff = get_vlc2(&s->gb, ff_msmp4_dc_chroma_vlc[s->dc_table_index].table, DC_VLC_BITS, 3);
+     }
+     if (dcdiff < 0) {
+         av_log(s->avctx, AV_LOG_ERROR, "Illegal DC VLC\n");
+         return -1;
+     }
+     if (dcdiff) {
+         if (dcdiff == 119 /* ESC index value */) {
+             /* TODO: Optimize */
+             if (v->pq == 1)      dcdiff = get_bits(gb, 10);
+             else if (v->pq == 2) dcdiff = get_bits(gb, 9);
+             else                 dcdiff = get_bits(gb, 8);
+         } else {
+             if (v->pq == 1)
+                 dcdiff = (dcdiff << 2) + get_bits(gb, 2) - 3;
+             else if (v->pq == 2)
+                 dcdiff = (dcdiff << 1) + get_bits1(gb)   - 1;
+         }
+         if (get_bits1(gb))
+             dcdiff = -dcdiff;
+     }
+     /* Prediction */
+     dcdiff += vc1_i_pred_dc(&v->s, v->overlap, v->pq, n, &dc_val, &dc_pred_dir);
+     *dc_val = dcdiff;
+     /* Store the quantized DC coeff, used for prediction */
+     if (n < 4) {
+         block[0] = dcdiff * s->y_dc_scale;
+     } else {
+         block[0] = dcdiff * s->c_dc_scale;
+     }
+     /* Skip ? */
+     if (!coded) {
+         goto not_coded;
+     }
+     // AC Decoding
+     i = 1;
+     {
+         int last = 0, skip, value;
+         const uint8_t *zz_table;
+         int scale;
+         int k;
+         scale = v->pq * 2 + v->halfpq;
+         if (v->s.ac_pred) {
+             if (!dc_pred_dir)
+                 zz_table = v->zz_8x8[2];
+             else
+                 zz_table = v->zz_8x8[3];
+         } else
+             zz_table = v->zz_8x8[1];
+         ac_val  = s->ac_val[0][0] + s->block_index[n] * 16;
+         ac_val2 = ac_val;
+         if (dc_pred_dir) // left
+             ac_val -= 16;
+         else // top
+             ac_val -= 16 * s->block_wrap[n];
+         while (!last) {
+             vc1_decode_ac_coeff(v, &last, &skip, &value, codingset);
+             i += skip;
+             if (i > 63)
+                 break;
+             block[zz_table[i++]] = value;
+         }
+         /* apply AC prediction if needed */
+         if (s->ac_pred) {
+             if (dc_pred_dir) { // left
+                 for (k = 1; k < 8; k++)
+                     block[k << v->left_blk_sh] += ac_val[k];
+             } else { // top
+                 for (k = 1; k < 8; k++)
+                     block[k << v->top_blk_sh] += ac_val[k + 8];
+             }
+         }
+         /* save AC coeffs for further prediction */
+         for (k = 1; k < 8; k++) {
+             ac_val2[k]     = block[k << v->left_blk_sh];
+             ac_val2[k + 8] = block[k << v->top_blk_sh];
+         }
+         /* scale AC coeffs */
+         for (k = 1; k < 64; k++)
+             if (block[k]) {
+                 block[k] *= scale;
+                 if (!v->pquantizer)
+                     block[k] += (block[k] < 0) ? -v->pq : v->pq;
+             }
+         if (s->ac_pred) i = 63;
+     }
+ not_coded:
+     if (!coded) {
+         int k, scale;
+         ac_val  = s->ac_val[0][0] + s->block_index[n] * 16;
+         ac_val2 = ac_val;
+         i = 0;
+         scale = v->pq * 2 + v->halfpq;
+         memset(ac_val2, 0, 16 * 2);
+         if (dc_pred_dir) { // left
+             ac_val -= 16;
+             if (s->ac_pred)
+                 memcpy(ac_val2, ac_val, 8 * 2);
+         } else { // top
+             ac_val -= 16 * s->block_wrap[n];
+             if (s->ac_pred)
+                 memcpy(ac_val2 + 8, ac_val + 8, 8 * 2);
+         }
+         /* apply AC prediction if needed */
+         if (s->ac_pred) {
+             if (dc_pred_dir) { //left
+                 for (k = 1; k < 8; k++) {
+                     block[k << v->left_blk_sh] = ac_val[k] * scale;
+                     if (!v->pquantizer && block[k << v->left_blk_sh])
+                         block[k << v->left_blk_sh] += (block[k << v->left_blk_sh] < 0) ? -v->pq : v->pq;
+                 }
+             } else { // top
+                 for (k = 1; k < 8; k++) {
+                     block[k << v->top_blk_sh] = ac_val[k + 8] * scale;
+                     if (!v->pquantizer && block[k << v->top_blk_sh])
+                         block[k << v->top_blk_sh] += (block[k << v->top_blk_sh] < 0) ? -v->pq : v->pq;
+                 }
+             }
+             i = 63;
+         }
+     }
+     s->block_last_index[n] = i;
+     return 0;
+ }
+ /** Decode intra block in intra frames - should be faster than decode_intra_block
+  * @param v VC1Context
+  * @param block block to decode
+  * @param[in] n subblock number
+  * @param coded are AC coeffs present or not
+  * @param codingset set of VLC to decode data
+  * @param mquant quantizer value for this macroblock
+  */
+ static int vc1_decode_i_block_adv(VC1Context *v, int16_t block[64], int n,
+                                   int coded, int codingset, int mquant)
+ {
+     GetBitContext *gb = &v->s.gb;
+     MpegEncContext *s = &v->s;
+     int dc_pred_dir = 0; /* Direction of the DC prediction used */
+     int i;
 -    int16_t *dc_val;
++    int16_t *dc_val = NULL;
+     int16_t *ac_val, *ac_val2;
+     int dcdiff;
+     int a_avail = v->a_avail, c_avail = v->c_avail;
+     int use_pred = s->ac_pred;
+     int scale;
+     int q1, q2 = 0;
+     int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     /* Get DC differential */
+     if (n < 4) {
+         dcdiff = get_vlc2(&s->gb, ff_msmp4_dc_luma_vlc[s->dc_table_index].table, DC_VLC_BITS, 3);
+     } else {
+         dcdiff = get_vlc2(&s->gb, ff_msmp4_dc_chroma_vlc[s->dc_table_index].table, DC_VLC_BITS, 3);
+     }
+     if (dcdiff < 0) {
+         av_log(s->avctx, AV_LOG_ERROR, "Illegal DC VLC\n");
+         return -1;
+     }
+     if (dcdiff) {
+         if (dcdiff == 119 /* ESC index value */) {
+             /* TODO: Optimize */
+             if (mquant == 1)      dcdiff = get_bits(gb, 10);
+             else if (mquant == 2) dcdiff = get_bits(gb, 9);
+             else                  dcdiff = get_bits(gb, 8);
+         } else {
+             if (mquant == 1)
+                 dcdiff = (dcdiff << 2) + get_bits(gb, 2) - 3;
+             else if (mquant == 2)
+                 dcdiff = (dcdiff << 1) + get_bits1(gb)   - 1;
+         }
+         if (get_bits1(gb))
+             dcdiff = -dcdiff;
+     }
+     /* Prediction */
+     dcdiff += ff_vc1_pred_dc(&v->s, v->overlap, mquant, n, v->a_avail, v->c_avail, &dc_val, &dc_pred_dir);
+     *dc_val = dcdiff;
+     /* Store the quantized DC coeff, used for prediction */
+     if (n < 4) {
+         block[0] = dcdiff * s->y_dc_scale;
+     } else {
+         block[0] = dcdiff * s->c_dc_scale;
+     }
+     //AC Decoding
+     i = 1;
+     /* check if AC is needed at all */
+     if (!a_avail && !c_avail)
+         use_pred = 0;
+     ac_val  = s->ac_val[0][0] + s->block_index[n] * 16;
+     ac_val2 = ac_val;
+     scale = mquant * 2 + ((mquant == v->pq) ? v->halfpq : 0);
+     if (dc_pred_dir) // left
+         ac_val -= 16;
+     else // top
+         ac_val -= 16 * s->block_wrap[n];
+     q1 = s->current_picture.qscale_table[mb_pos];
+     if ( dc_pred_dir && c_avail && mb_pos)
+         q2 = s->current_picture.qscale_table[mb_pos - 1];
+     if (!dc_pred_dir && a_avail && mb_pos >= s->mb_stride)
+         q2 = s->current_picture.qscale_table[mb_pos - s->mb_stride];
+     if ( dc_pred_dir && n == 1)
+         q2 = q1;
+     if (!dc_pred_dir && n == 2)
+         q2 = q1;
+     if (n == 3)
+         q2 = q1;
+     if (coded) {
+         int last = 0, skip, value;
+         const uint8_t *zz_table;
+         int k;
+         if (v->s.ac_pred) {
+             if (!use_pred && v->fcm == ILACE_FRAME) {
+                 zz_table = v->zzi_8x8;
+             } else {
+                 if (!dc_pred_dir) // top
+                     zz_table = v->zz_8x8[2];
+                 else // left
+                     zz_table = v->zz_8x8[3];
+             }
+         } else {
+             if (v->fcm != ILACE_FRAME)
+                 zz_table = v->zz_8x8[1];
+             else
+                 zz_table = v->zzi_8x8;
+         }
+         while (!last) {
+             vc1_decode_ac_coeff(v, &last, &skip, &value, codingset);
+             i += skip;
+             if (i > 63)
+                 break;
+             block[zz_table[i++]] = value;
+         }
+         /* apply AC prediction if needed */
+         if (use_pred) {
+             /* scale predictors if needed*/
+             if (q2 && q1 != q2) {
+                 q1 = q1 * 2 + ((q1 == v->pq) ? v->halfpq : 0) - 1;
+                 q2 = q2 * 2 + ((q2 == v->pq) ? v->halfpq : 0) - 1;
+                 if (q1 < 1)
+                     return AVERROR_INVALIDDATA;
+                 if (dc_pred_dir) { // left
+                     for (k = 1; k < 8; k++)
+                         block[k << v->left_blk_sh] += (ac_val[k] * q2 * ff_vc1_dqscale[q1 - 1] + 0x20000) >> 18;
+                 } else { // top
+                     for (k = 1; k < 8; k++)
+                         block[k << v->top_blk_sh] += (ac_val[k + 8] * q2 * ff_vc1_dqscale[q1 - 1] + 0x20000) >> 18;
+                 }
+             } else {
+                 if (dc_pred_dir) { //left
+                     for (k = 1; k < 8; k++)
+                         block[k << v->left_blk_sh] += ac_val[k];
+                 } else { //top
+                     for (k = 1; k < 8; k++)
+                         block[k << v->top_blk_sh] += ac_val[k + 8];
+                 }
+             }
+         }
+         /* save AC coeffs for further prediction */
+         for (k = 1; k < 8; k++) {
+             ac_val2[k    ] = block[k << v->left_blk_sh];
+             ac_val2[k + 8] = block[k << v->top_blk_sh];
+         }
+         /* scale AC coeffs */
+         for (k = 1; k < 64; k++)
+             if (block[k]) {
+                 block[k] *= scale;
+                 if (!v->pquantizer)
+                     block[k] += (block[k] < 0) ? -mquant : mquant;
+             }
+         if (use_pred) i = 63;
+     } else { // no AC coeffs
+         int k;
+         memset(ac_val2, 0, 16 * 2);
+         if (dc_pred_dir) { // left
+             if (use_pred) {
+                 memcpy(ac_val2, ac_val, 8 * 2);
+                 if (q2 && q1 != q2) {
+                     q1 = q1 * 2 + ((q1 == v->pq) ? v->halfpq : 0) - 1;
+                     q2 = q2 * 2 + ((q2 == v->pq) ? v->halfpq : 0) - 1;
+                     if (q1 < 1)
+                         return AVERROR_INVALIDDATA;
+                     for (k = 1; k < 8; k++)
+                         ac_val2[k] = (ac_val2[k] * q2 * ff_vc1_dqscale[q1 - 1] + 0x20000) >> 18;
+                 }
+             }
+         } else { // top
+             if (use_pred) {
+                 memcpy(ac_val2 + 8, ac_val + 8, 8 * 2);
+                 if (q2 && q1 != q2) {
+                     q1 = q1 * 2 + ((q1 == v->pq) ? v->halfpq : 0) - 1;
+                     q2 = q2 * 2 + ((q2 == v->pq) ? v->halfpq : 0) - 1;
+                     if (q1 < 1)
+                         return AVERROR_INVALIDDATA;
+                     for (k = 1; k < 8; k++)
+                         ac_val2[k + 8] = (ac_val2[k + 8] * q2 * ff_vc1_dqscale[q1 - 1] + 0x20000) >> 18;
+                 }
+             }
+         }
+         /* apply AC prediction if needed */
+         if (use_pred) {
+             if (dc_pred_dir) { // left
+                 for (k = 1; k < 8; k++) {
+                     block[k << v->left_blk_sh] = ac_val2[k] * scale;
+                     if (!v->pquantizer && block[k << v->left_blk_sh])
+                         block[k << v->left_blk_sh] += (block[k << v->left_blk_sh] < 0) ? -mquant : mquant;
+                 }
+             } else { // top
+                 for (k = 1; k < 8; k++) {
+                     block[k << v->top_blk_sh] = ac_val2[k + 8] * scale;
+                     if (!v->pquantizer && block[k << v->top_blk_sh])
+                         block[k << v->top_blk_sh] += (block[k << v->top_blk_sh] < 0) ? -mquant : mquant;
+                 }
+             }
+             i = 63;
+         }
+     }
+     s->block_last_index[n] = i;
+     return 0;
+ }
+ /** Decode intra block in inter frames - more generic version than vc1_decode_i_block
+  * @param v VC1Context
+  * @param block block to decode
+  * @param[in] n subblock index
+  * @param coded are AC coeffs present or not
+  * @param mquant block quantizer
+  * @param codingset set of VLC to decode data
+  */
+ static int vc1_decode_intra_block(VC1Context *v, int16_t block[64], int n,
+                                   int coded, int mquant, int codingset)
+ {
+     GetBitContext *gb = &v->s.gb;
+     MpegEncContext *s = &v->s;
+     int dc_pred_dir = 0; /* Direction of the DC prediction used */
+     int i;
 -    mquant = v->pq; /* Loosy initialization */
++    int16_t *dc_val = NULL;
+     int16_t *ac_val, *ac_val2;
+     int dcdiff;
+     int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     int a_avail = v->a_avail, c_avail = v->c_avail;
+     int use_pred = s->ac_pred;
+     int scale;
+     int q1, q2 = 0;
+     s->bdsp.clear_block(block);
+     /* XXX: Guard against dumb values of mquant */
+     mquant = (mquant < 1) ? 0 : ((mquant > 31) ? 31 : mquant);
+     /* Set DC scale - y and c use the same */
+     s->y_dc_scale = s->y_dc_scale_table[mquant];
+     s->c_dc_scale = s->c_dc_scale_table[mquant];
+     /* Get DC differential */
+     if (n < 4) {
+         dcdiff = get_vlc2(&s->gb, ff_msmp4_dc_luma_vlc[s->dc_table_index].table, DC_VLC_BITS, 3);
+     } else {
+         dcdiff = get_vlc2(&s->gb, ff_msmp4_dc_chroma_vlc[s->dc_table_index].table, DC_VLC_BITS, 3);
+     }
+     if (dcdiff < 0) {
+         av_log(s->avctx, AV_LOG_ERROR, "Illegal DC VLC\n");
+         return -1;
+     }
+     if (dcdiff) {
+         if (dcdiff == 119 /* ESC index value */) {
+             /* TODO: Optimize */
+             if (mquant == 1)      dcdiff = get_bits(gb, 10);
+             else if (mquant == 2) dcdiff = get_bits(gb, 9);
+             else                  dcdiff = get_bits(gb, 8);
+         } else {
+             if (mquant == 1)
+                 dcdiff = (dcdiff << 2) + get_bits(gb, 2) - 3;
+             else if (mquant == 2)
+                 dcdiff = (dcdiff << 1) + get_bits1(gb)   - 1;
+         }
+         if (get_bits1(gb))
+             dcdiff = -dcdiff;
+     }
+     /* Prediction */
+     dcdiff += ff_vc1_pred_dc(&v->s, v->overlap, mquant, n, a_avail, c_avail, &dc_val, &dc_pred_dir);
+     *dc_val = dcdiff;
+     /* Store the quantized DC coeff, used for prediction */
+     if (n < 4) {
+         block[0] = dcdiff * s->y_dc_scale;
+     } else {
+         block[0] = dcdiff * s->c_dc_scale;
+     }
+     //AC Decoding
+     i = 1;
+     /* check if AC is needed at all and adjust direction if needed */
+     if (!a_avail) dc_pred_dir = 1;
+     if (!c_avail) dc_pred_dir = 0;
+     if (!a_avail && !c_avail) use_pred = 0;
+     ac_val = s->ac_val[0][0] + s->block_index[n] * 16;
+     ac_val2 = ac_val;
+     scale = mquant * 2 + v->halfpq;
+     if (dc_pred_dir) //left
+         ac_val -= 16;
+     else //top
+         ac_val -= 16 * s->block_wrap[n];
+     q1 = s->current_picture.qscale_table[mb_pos];
+     if (dc_pred_dir && c_avail && mb_pos)
+         q2 = s->current_picture.qscale_table[mb_pos - 1];
+     if (!dc_pred_dir && a_avail && mb_pos >= s->mb_stride)
+         q2 = s->current_picture.qscale_table[mb_pos - s->mb_stride];
+     if ( dc_pred_dir && n == 1)
+         q2 = q1;
+     if (!dc_pred_dir && n == 2)
+         q2 = q1;
+     if (n == 3) q2 = q1;
+     if (coded) {
+         int last = 0, skip, value;
+         int k;
+         while (!last) {
+             vc1_decode_ac_coeff(v, &last, &skip, &value, codingset);
+             i += skip;
+             if (i > 63)
+                 break;
+             if (v->fcm == PROGRESSIVE)
+                 block[v->zz_8x8[0][i++]] = value;
+             else {
+                 if (use_pred && (v->fcm == ILACE_FRAME)) {
+                     if (!dc_pred_dir) // top
+                         block[v->zz_8x8[2][i++]] = value;
+                     else // left
+                         block[v->zz_8x8[3][i++]] = value;
+                 } else {
+                     block[v->zzi_8x8[i++]] = value;
+                 }
+             }
+         }
+         /* apply AC prediction if needed */
+         if (use_pred) {
+             /* scale predictors if needed*/
+             if (q2 && q1 != q2) {
+                 q1 = q1 * 2 + ((q1 == v->pq) ? v->halfpq : 0) - 1;
+                 q2 = q2 * 2 + ((q2 == v->pq) ? v->halfpq : 0) - 1;
+                 if (q1 < 1)
+                     return AVERROR_INVALIDDATA;
+                 if (dc_pred_dir) { // left
+                     for (k = 1; k < 8; k++)
+                         block[k << v->left_blk_sh] += (ac_val[k] * q2 * ff_vc1_dqscale[q1 - 1] + 0x20000) >> 18;
+                 } else { //top
+                     for (k = 1; k < 8; k++)
+                         block[k << v->top_blk_sh] += (ac_val[k + 8] * q2 * ff_vc1_dqscale[q1 - 1] + 0x20000) >> 18;
+                 }
+             } else {
+                 if (dc_pred_dir) { // left
+                     for (k = 1; k < 8; k++)
+                         block[k << v->left_blk_sh] += ac_val[k];
+                 } else { // top
+                     for (k = 1; k < 8; k++)
+                         block[k << v->top_blk_sh] += ac_val[k + 8];
+                 }
+             }
+         }
+         /* save AC coeffs for further prediction */
+         for (k = 1; k < 8; k++) {
+             ac_val2[k    ] = block[k << v->left_blk_sh];
+             ac_val2[k + 8] = block[k << v->top_blk_sh];
+         }
+         /* scale AC coeffs */
+         for (k = 1; k < 64; k++)
+             if (block[k]) {
+                 block[k] *= scale;
+                 if (!v->pquantizer)
+                     block[k] += (block[k] < 0) ? -mquant : mquant;
+             }
+         if (use_pred) i = 63;
+     } else { // no AC coeffs
+         int k;
+         memset(ac_val2, 0, 16 * 2);
+         if (dc_pred_dir) { // left
+             if (use_pred) {
+                 memcpy(ac_val2, ac_val, 8 * 2);
+                 if (q2 && q1 != q2) {
+                     q1 = q1 * 2 + ((q1 == v->pq) ? v->halfpq : 0) - 1;
+                     q2 = q2 * 2 + ((q2 == v->pq) ? v->halfpq : 0) - 1;
+                     if (q1 < 1)
+                         return AVERROR_INVALIDDATA;
+                     for (k = 1; k < 8; k++)
+                         ac_val2[k] = (ac_val2[k] * q2 * ff_vc1_dqscale[q1 - 1] + 0x20000) >> 18;
+                 }
+             }
+         } else { // top
+             if (use_pred) {
+                 memcpy(ac_val2 + 8, ac_val + 8, 8 * 2);
+                 if (q2 && q1 != q2) {
+                     q1 = q1 * 2 + ((q1 == v->pq) ? v->halfpq : 0) - 1;
+                     q2 = q2 * 2 + ((q2 == v->pq) ? v->halfpq : 0) - 1;
+                     if (q1 < 1)
+                         return AVERROR_INVALIDDATA;
+                     for (k = 1; k < 8; k++)
+                         ac_val2[k + 8] = (ac_val2[k + 8] * q2 * ff_vc1_dqscale[q1 - 1] + 0x20000) >> 18;
+                 }
+             }
+         }
+         /* apply AC prediction if needed */
+         if (use_pred) {
+             if (dc_pred_dir) { // left
+                 for (k = 1; k < 8; k++) {
+                     block[k << v->left_blk_sh] = ac_val2[k] * scale;
+                     if (!v->pquantizer && block[k << v->left_blk_sh])
+                         block[k << v->left_blk_sh] += (block[k << v->left_blk_sh] < 0) ? -mquant : mquant;
+                 }
+             } else { // top
+                 for (k = 1; k < 8; k++) {
+                     block[k << v->top_blk_sh] = ac_val2[k + 8] * scale;
+                     if (!v->pquantizer && block[k << v->top_blk_sh])
+                         block[k << v->top_blk_sh] += (block[k << v->top_blk_sh] < 0) ? -mquant : mquant;
+                 }
+             }
+             i = 63;
+         }
+     }
+     s->block_last_index[n] = i;
+     return 0;
+ }
+ /** Decode P block
+  */
+ static int vc1_decode_p_block(VC1Context *v, int16_t block[64], int n,
+                               int mquant, int ttmb, int first_block,
+                               uint8_t *dst, int linesize, int skip_block,
+                               int *ttmb_out)
+ {
+     MpegEncContext *s = &v->s;
+     GetBitContext *gb = &s->gb;
+     int i, j;
+     int subblkpat = 0;
+     int scale, off, idx, last, skip, value;
+     int ttblk = ttmb & 7;
+     int pat = 0;
+     s->bdsp.clear_block(block);
+     if (ttmb == -1) {
+         ttblk = ff_vc1_ttblk_to_tt[v->tt_index][get_vlc2(gb, ff_vc1_ttblk_vlc[v->tt_index].table, VC1_TTBLK_VLC_BITS, 1)];
+     }
+     if (ttblk == TT_4X4) {
+         subblkpat = ~(get_vlc2(gb, ff_vc1_subblkpat_vlc[v->tt_index].table, VC1_SUBBLKPAT_VLC_BITS, 1) + 1);
+     }
+     if ((ttblk != TT_8X8 && ttblk != TT_4X4)
+         && ((v->ttmbf || (ttmb != -1 && (ttmb & 8) && !first_block))
+             || (!v->res_rtm_flag && !first_block))) {
+         subblkpat = decode012(gb);
+         if (subblkpat)
+             subblkpat ^= 3; // swap decoded pattern bits
+         if (ttblk == TT_8X4_TOP || ttblk == TT_8X4_BOTTOM)
+             ttblk = TT_8X4;
+         if (ttblk == TT_4X8_RIGHT || ttblk == TT_4X8_LEFT)
+             ttblk = TT_4X8;
+     }
+     scale = 2 * mquant + ((v->pq == mquant) ? v->halfpq : 0);
+     // convert transforms like 8X4_TOP to generic TT and SUBBLKPAT
+     if (ttblk == TT_8X4_TOP || ttblk == TT_8X4_BOTTOM) {
+         subblkpat = 2 - (ttblk == TT_8X4_TOP);
+         ttblk     = TT_8X4;
+     }
+     if (ttblk == TT_4X8_RIGHT || ttblk == TT_4X8_LEFT) {
+         subblkpat = 2 - (ttblk == TT_4X8_LEFT);
+         ttblk     = TT_4X8;
+     }
+     switch (ttblk) {
+     case TT_8X8:
+         pat  = 0xF;
+         i    = 0;
+         last = 0;
+         while (!last) {
+             vc1_decode_ac_coeff(v, &last, &skip, &value, v->codingset2);
+             i += skip;
+             if (i > 63)
+                 break;
+             if (!v->fcm)
+                 idx = v->zz_8x8[0][i++];
+             else
+                 idx = v->zzi_8x8[i++];
+             block[idx] = value * scale;
+             if (!v->pquantizer)
+                 block[idx] += (block[idx] < 0) ? -mquant : mquant;
+         }
+         if (!skip_block) {
+             if (i == 1)
+                 v->vc1dsp.vc1_inv_trans_8x8_dc(dst, linesize, block);
+             else {
+                 v->vc1dsp.vc1_inv_trans_8x8(block);
+                 s->idsp.add_pixels_clamped(block, dst, linesize);
+             }
+         }
+         break;
+     case TT_4X4:
+         pat = ~subblkpat & 0xF;
+         for (j = 0; j < 4; j++) {
+             last = subblkpat & (1 << (3 - j));
+             i    = 0;
+             off  = (j & 1) * 4 + (j & 2) * 16;
+             while (!last) {
+                 vc1_decode_ac_coeff(v, &last, &skip, &value, v->codingset2);
+                 i += skip;
+                 if (i > 15)
+                     break;
+                 if (!v->fcm)
+                     idx = ff_vc1_simple_progressive_4x4_zz[i++];
+                 else
+                     idx = ff_vc1_adv_interlaced_4x4_zz[i++];
+                 block[idx + off] = value * scale;
+                 if (!v->pquantizer)
+                     block[idx + off] += (block[idx + off] < 0) ? -mquant : mquant;
+             }
+             if (!(subblkpat & (1 << (3 - j))) && !skip_block) {
+                 if (i == 1)
+                     v->vc1dsp.vc1_inv_trans_4x4_dc(dst + (j & 1) * 4 + (j & 2) * 2 * linesize, linesize, block + off);
+                 else
+                     v->vc1dsp.vc1_inv_trans_4x4(dst + (j & 1) * 4 + (j & 2) *  2 * linesize, linesize, block + off);
+             }
+         }
+         break;
+     case TT_8X4:
+         pat = ~((subblkpat & 2) * 6 + (subblkpat & 1) * 3) & 0xF;
+         for (j = 0; j < 2; j++) {
+             last = subblkpat & (1 << (1 - j));
+             i    = 0;
+             off  = j * 32;
+             while (!last) {
+                 vc1_decode_ac_coeff(v, &last, &skip, &value, v->codingset2);
+                 i += skip;
+                 if (i > 31)
+                     break;
+                 if (!v->fcm)
+                     idx = v->zz_8x4[i++] + off;
+                 else
+                     idx = ff_vc1_adv_interlaced_8x4_zz[i++] + off;
+                 block[idx] = value * scale;
+                 if (!v->pquantizer)
+                     block[idx] += (block[idx] < 0) ? -mquant : mquant;
+             }
+             if (!(subblkpat & (1 << (1 - j))) && !skip_block) {
+                 if (i == 1)
+                     v->vc1dsp.vc1_inv_trans_8x4_dc(dst + j * 4 * linesize, linesize, block + off);
+                 else
+                     v->vc1dsp.vc1_inv_trans_8x4(dst + j * 4 * linesize, linesize, block + off);
+             }
+         }
+         break;
+     case TT_4X8:
+         pat = ~(subblkpat * 5) & 0xF;
+         for (j = 0; j < 2; j++) {
+             last = subblkpat & (1 << (1 - j));
+             i    = 0;
+             off  = j * 4;
+             while (!last) {
+                 vc1_decode_ac_coeff(v, &last, &skip, &value, v->codingset2);
+                 i += skip;
+                 if (i > 31)
+                     break;
+                 if (!v->fcm)
+                     idx = v->zz_4x8[i++] + off;
+                 else
+                     idx = ff_vc1_adv_interlaced_4x8_zz[i++] + off;
+                 block[idx] = value * scale;
+                 if (!v->pquantizer)
+                     block[idx] += (block[idx] < 0) ? -mquant : mquant;
+             }
+             if (!(subblkpat & (1 << (1 - j))) && !skip_block) {
+                 if (i == 1)
+                     v->vc1dsp.vc1_inv_trans_4x8_dc(dst + j * 4, linesize, block + off);
+                 else
+                     v->vc1dsp.vc1_inv_trans_4x8(dst + j*4, linesize, block + off);
+             }
+         }
+         break;
+     }
+     if (ttmb_out)
+         *ttmb_out |= ttblk << (n * 4);
+     return pat;
+ }
+ /** @} */ // Macroblock group
+ static const int size_table  [6] = { 0, 2, 3, 4,  5,  8 };
+ static const int offset_table[6] = { 0, 1, 3, 7, 15, 31 };
+ /** Decode one P-frame MB
+  */
+ static int vc1_decode_p_mb(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     GetBitContext *gb = &s->gb;
+     int i, j;
+     int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     int cbp; /* cbp decoding stuff */
+     int mqdiff, mquant; /* MB quantization */
+     int ttmb = v->ttfrm; /* MB Transform type */
+     int mb_has_coeffs = 1; /* last_flag */
+     int dmv_x, dmv_y; /* Differential MV components */
+     int index, index1; /* LUT indexes */
+     int val, sign; /* temp values */
+     int first_block = 1;
+     int dst_idx, off;
+     int skipped, fourmv;
+     int block_cbp = 0, pat, block_tt = 0, block_intra = 0;
+     mquant = v->pq; /* lossy initialization */
+     if (v->mv_type_is_raw)
+         fourmv = get_bits1(gb);
+     else
+         fourmv = v->mv_type_mb_plane[mb_pos];
+     if (v->skip_is_raw)
+         skipped = get_bits1(gb);
+     else
+         skipped = v->s.mbskip_table[mb_pos];
+     if (!fourmv) { /* 1MV mode */
+         if (!skipped) {
+             GET_MVDATA(dmv_x, dmv_y);
+             if (s->mb_intra) {
+                 s->current_picture.motion_val[1][s->block_index[0]][0] = 0;
+                 s->current_picture.motion_val[1][s->block_index[0]][1] = 0;
+             }
+             s->current_picture.mb_type[mb_pos] = s->mb_intra ? MB_TYPE_INTRA : MB_TYPE_16x16;
+             ff_vc1_pred_mv(v, 0, dmv_x, dmv_y, 1, v->range_x, v->range_y, v->mb_type[0], 0, 0);
+             /* FIXME Set DC val for inter block ? */
+             if (s->mb_intra && !mb_has_coeffs) {
+                 GET_MQUANT();
+                 s->ac_pred = get_bits1(gb);
+                 cbp        = 0;
+             } else if (mb_has_coeffs) {
+                 if (s->mb_intra)
+                     s->ac_pred = get_bits1(gb);
+                 cbp = get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_CBPCY_P_VLC_BITS, 2);
+                 GET_MQUANT();
+             } else {
+                 mquant = v->pq;
+                 cbp    = 0;
+             }
+             s->current_picture.qscale_table[mb_pos] = mquant;
+             if (!v->ttmbf && !s->mb_intra && mb_has_coeffs)
+                 ttmb = get_vlc2(gb, ff_vc1_ttmb_vlc[v->tt_index].table,
+                                 VC1_TTMB_VLC_BITS, 2);
+             if (!s->mb_intra) ff_vc1_mc_1mv(v, 0);
+             dst_idx = 0;
+             for (i = 0; i < 6; i++) {
+                 s->dc_val[0][s->block_index[i]] = 0;
+                 dst_idx += i >> 2;
+                 val = ((cbp >> (5 - i)) & 1);
+                 off = (i & 4) ? 0 : ((i & 1) * 8 + (i & 2) * 4 * s->linesize);
+                 v->mb_type[0][s->block_index[i]] = s->mb_intra;
+                 if (s->mb_intra) {
+                     /* check if prediction blocks A and C are available */
+                     v->a_avail = v->c_avail = 0;
+                     if (i == 2 || i == 3 || !s->first_slice_line)
+                         v->a_avail = v->mb_type[0][s->block_index[i] - s->block_wrap[i]];
+                     if (i == 1 || i == 3 || s->mb_x)
+                         v->c_avail = v->mb_type[0][s->block_index[i] - 1];
+                     vc1_decode_intra_block(v, s->block[i], i, val, mquant,
+                                            (i & 4) ? v->codingset2 : v->codingset);
+                     if ((i>3) && (s->flags & CODEC_FLAG_GRAY))
+                         continue;
+                     v->vc1dsp.vc1_inv_trans_8x8(s->block[i]);
+                     if (v->rangeredfrm)
+                         for (j = 0; j < 64; j++)
+                             s->block[i][j] <<= 1;
+                     s->idsp.put_signed_pixels_clamped(s->block[i],
+                                                       s->dest[dst_idx] + off,
+                                                       i & 4 ? s->uvlinesize
+                                                             : s->linesize);
+                     if (v->pq >= 9 && v->overlap) {
+                         if (v->c_avail)
+                             v->vc1dsp.vc1_h_overlap(s->dest[dst_idx] + off, i & 4 ? s->uvlinesize : s->linesize);
+                         if (v->a_avail)
+                             v->vc1dsp.vc1_v_overlap(s->dest[dst_idx] + off, i & 4 ? s->uvlinesize : s->linesize);
+                     }
+                     block_cbp   |= 0xF << (i << 2);
+                     block_intra |= 1 << i;
+                 } else if (val) {
+                     pat = vc1_decode_p_block(v, s->block[i], i, mquant, ttmb, first_block,
+                                              s->dest[dst_idx] + off, (i & 4) ? s->uvlinesize : s->linesize,
+                                              (i & 4) && (s->flags & CODEC_FLAG_GRAY), &block_tt);
+                     block_cbp |= pat << (i << 2);
+                     if (!v->ttmbf && ttmb < 8)
+                         ttmb = -1;
+                     first_block = 0;
+                 }
+             }
+         } else { // skipped
+             s->mb_intra = 0;
+             for (i = 0; i < 6; i++) {
+                 v->mb_type[0][s->block_index[i]] = 0;
+                 s->dc_val[0][s->block_index[i]]  = 0;
+             }
+             s->current_picture.mb_type[mb_pos]      = MB_TYPE_SKIP;
+             s->current_picture.qscale_table[mb_pos] = 0;
+             ff_vc1_pred_mv(v, 0, 0, 0, 1, v->range_x, v->range_y, v->mb_type[0], 0, 0);
+             ff_vc1_mc_1mv(v, 0);
+         }
+     } else { // 4MV mode
+         if (!skipped /* unskipped MB */) {
+             int intra_count = 0, coded_inter = 0;
+             int is_intra[6], is_coded[6];
+             /* Get CBPCY */
+             cbp = get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_CBPCY_P_VLC_BITS, 2);
+             for (i = 0; i < 6; i++) {
+                 val = ((cbp >> (5 - i)) & 1);
+                 s->dc_val[0][s->block_index[i]] = 0;
+                 s->mb_intra                     = 0;
+                 if (i < 4) {
+                     dmv_x = dmv_y = 0;
+                     s->mb_intra   = 0;
+                     mb_has_coeffs = 0;
+                     if (val) {
+                         GET_MVDATA(dmv_x, dmv_y);
+                     }
+                     ff_vc1_pred_mv(v, i, dmv_x, dmv_y, 0, v->range_x, v->range_y, v->mb_type[0], 0, 0);
+                     if (!s->mb_intra)
+                         ff_vc1_mc_4mv_luma(v, i, 0, 0);
+                     intra_count += s->mb_intra;
+                     is_intra[i]  = s->mb_intra;
+                     is_coded[i]  = mb_has_coeffs;
+                 }
+                 if (i & 4) {
+                     is_intra[i] = (intra_count >= 3);
+                     is_coded[i] = val;
+                 }
+                 if (i == 4)
+                     ff_vc1_mc_4mv_chroma(v, 0);
+                 v->mb_type[0][s->block_index[i]] = is_intra[i];
+                 if (!coded_inter)
+                     coded_inter = !is_intra[i] & is_coded[i];
+             }
+             // if there are no coded blocks then don't do anything more
+             dst_idx = 0;
+             if (!intra_count && !coded_inter)
+                 goto end;
+             GET_MQUANT();
+             s->current_picture.qscale_table[mb_pos] = mquant;
+             /* test if block is intra and has pred */
+             {
+                 int intrapred = 0;
+                 for (i = 0; i < 6; i++)
+                     if (is_intra[i]) {
+                         if (((!s->first_slice_line || (i == 2 || i == 3)) && v->mb_type[0][s->block_index[i] - s->block_wrap[i]])
+                             || ((s->mb_x || (i == 1 || i == 3)) && v->mb_type[0][s->block_index[i] - 1])) {
+                             intrapred = 1;
+                             break;
+                         }
+                     }
+                 if (intrapred)
+                     s->ac_pred = get_bits1(gb);
+                 else
+                     s->ac_pred = 0;
+             }
+             if (!v->ttmbf && coded_inter)
+                 ttmb = get_vlc2(gb, ff_vc1_ttmb_vlc[v->tt_index].table, VC1_TTMB_VLC_BITS, 2);
+             for (i = 0; i < 6; i++) {
+                 dst_idx    += i >> 2;
+                 off         = (i & 4) ? 0 : ((i & 1) * 8 + (i & 2) * 4 * s->linesize);
+                 s->mb_intra = is_intra[i];
+                 if (is_intra[i]) {
+                     /* check if prediction blocks A and C are available */
+                     v->a_avail = v->c_avail = 0;
+                     if (i == 2 || i == 3 || !s->first_slice_line)
+                         v->a_avail = v->mb_type[0][s->block_index[i] - s->block_wrap[i]];
+                     if (i == 1 || i == 3 || s->mb_x)
+                         v->c_avail = v->mb_type[0][s->block_index[i] - 1];
+                     vc1_decode_intra_block(v, s->block[i], i, is_coded[i], mquant,
+                                            (i & 4) ? v->codingset2 : v->codingset);
+                     if ((i>3) && (s->flags & CODEC_FLAG_GRAY))
+                         continue;
+                     v->vc1dsp.vc1_inv_trans_8x8(s->block[i]);
+                     if (v->rangeredfrm)
+                         for (j = 0; j < 64; j++)
+                             s->block[i][j] <<= 1;
+                     s->idsp.put_signed_pixels_clamped(s->block[i],
+                                                       s->dest[dst_idx] + off,
+                                                       (i & 4) ? s->uvlinesize
+                                                               : s->linesize);
+                     if (v->pq >= 9 && v->overlap) {
+                         if (v->c_avail)
+                             v->vc1dsp.vc1_h_overlap(s->dest[dst_idx] + off, i & 4 ? s->uvlinesize : s->linesize);
+                         if (v->a_avail)
+                             v->vc1dsp.vc1_v_overlap(s->dest[dst_idx] + off, i & 4 ? s->uvlinesize : s->linesize);
+                     }
+                     block_cbp   |= 0xF << (i << 2);
+                     block_intra |= 1 << i;
+                 } else if (is_coded[i]) {
+                     pat = vc1_decode_p_block(v, s->block[i], i, mquant, ttmb,
+                                              first_block, s->dest[dst_idx] + off,
+                                              (i & 4) ? s->uvlinesize : s->linesize,
+                                              (i & 4) && (s->flags & CODEC_FLAG_GRAY),
+                                              &block_tt);
+                     block_cbp |= pat << (i << 2);
+                     if (!v->ttmbf && ttmb < 8)
+                         ttmb = -1;
+                     first_block = 0;
+                 }
+             }
+         } else { // skipped MB
+             s->mb_intra                               = 0;
+             s->current_picture.qscale_table[mb_pos] = 0;
+             for (i = 0; i < 6; i++) {
+                 v->mb_type[0][s->block_index[i]] = 0;
+                 s->dc_val[0][s->block_index[i]]  = 0;
+             }
+             for (i = 0; i < 4; i++) {
+                 ff_vc1_pred_mv(v, i, 0, 0, 0, v->range_x, v->range_y, v->mb_type[0], 0, 0);
+                 ff_vc1_mc_4mv_luma(v, i, 0, 0);
+             }
+             ff_vc1_mc_4mv_chroma(v, 0);
+             s->current_picture.qscale_table[mb_pos] = 0;
+         }
+     }
+ end:
+     v->cbp[s->mb_x]      = block_cbp;
+     v->ttblk[s->mb_x]    = block_tt;
+     v->is_intra[s->mb_x] = block_intra;
+     return 0;
+ }
+ /* Decode one macroblock in an interlaced frame p picture */
+ static int vc1_decode_p_mb_intfr(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     GetBitContext *gb = &s->gb;
+     int i;
+     int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     int cbp = 0; /* cbp decoding stuff */
+     int mqdiff, mquant; /* MB quantization */
+     int ttmb = v->ttfrm; /* MB Transform type */
+     int mb_has_coeffs = 1; /* last_flag */
+     int dmv_x, dmv_y; /* Differential MV components */
+     int val; /* temp value */
+     int first_block = 1;
+     int dst_idx, off;
+     int skipped, fourmv = 0, twomv = 0;
+     int block_cbp = 0, pat, block_tt = 0;
+     int idx_mbmode = 0, mvbp;
+     int stride_y, fieldtx;
 -            s->mb_intra = v->is_intra[s->mb_x] = 1;
++    mquant = v->pq; /* Lossy initialization */
+     if (v->skip_is_raw)
+         skipped = get_bits1(gb);
+     else
+         skipped = v->s.mbskip_table[mb_pos];
+     if (!skipped) {
+         if (v->fourmvswitch)
+             idx_mbmode = get_vlc2(gb, v->mbmode_vlc->table, VC1_INTFR_4MV_MBMODE_VLC_BITS, 2); // try getting this done
+         else
+             idx_mbmode = get_vlc2(gb, v->mbmode_vlc->table, VC1_INTFR_NON4MV_MBMODE_VLC_BITS, 2); // in a single line
+         switch (ff_vc1_mbmode_intfrp[v->fourmvswitch][idx_mbmode][0]) {
+         /* store the motion vector type in a flag (useful later) */
+         case MV_PMODE_INTFR_4MV:
+             fourmv = 1;
+             v->blk_mv_type[s->block_index[0]] = 0;
+             v->blk_mv_type[s->block_index[1]] = 0;
+             v->blk_mv_type[s->block_index[2]] = 0;
+             v->blk_mv_type[s->block_index[3]] = 0;
+             break;
+         case MV_PMODE_INTFR_4MV_FIELD:
+             fourmv = 1;
+             v->blk_mv_type[s->block_index[0]] = 1;
+             v->blk_mv_type[s->block_index[1]] = 1;
+             v->blk_mv_type[s->block_index[2]] = 1;
+             v->blk_mv_type[s->block_index[3]] = 1;
+             break;
+         case MV_PMODE_INTFR_2MV_FIELD:
+             twomv = 1;
+             v->blk_mv_type[s->block_index[0]] = 1;
+             v->blk_mv_type[s->block_index[1]] = 1;
+             v->blk_mv_type[s->block_index[2]] = 1;
+             v->blk_mv_type[s->block_index[3]] = 1;
+             break;
+         case MV_PMODE_INTFR_1MV:
+             v->blk_mv_type[s->block_index[0]] = 0;
+             v->blk_mv_type[s->block_index[1]] = 0;
+             v->blk_mv_type[s->block_index[2]] = 0;
+             v->blk_mv_type[s->block_index[3]] = 0;
+             break;
+         }
+         if (ff_vc1_mbmode_intfrp[v->fourmvswitch][idx_mbmode][0] == MV_PMODE_INTFR_INTRA) { // intra MB
+             for (i = 0; i < 4; i++) {
+                 s->current_picture.motion_val[1][s->block_index[i]][0] = 0;
+                 s->current_picture.motion_val[1][s->block_index[i]][1] = 0;
+             }
+             s->current_picture.mb_type[mb_pos]                     = MB_TYPE_INTRA;
 -    int pred_flag;
++            s->mb_intra = 1;
++            v->is_intra[s->mb_x] = 0x3F;
+             for (i = 0; i < 6; i++)
+                 v->mb_type[0][s->block_index[i]] = 1;
+             fieldtx = v->fieldtx_plane[mb_pos] = get_bits1(gb);
+             mb_has_coeffs = get_bits1(gb);
+             if (mb_has_coeffs)
+                 cbp = 1 + get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_CBPCY_P_VLC_BITS, 2);
+             v->s.ac_pred = v->acpred_plane[mb_pos] = get_bits1(gb);
+             GET_MQUANT();
+             s->current_picture.qscale_table[mb_pos] = mquant;
+             /* Set DC scale - y and c use the same (not sure if necessary here) */
+             s->y_dc_scale = s->y_dc_scale_table[mquant];
+             s->c_dc_scale = s->c_dc_scale_table[mquant];
+             dst_idx = 0;
+             for (i = 0; i < 6; i++) {
+                 s->dc_val[0][s->block_index[i]] = 0;
+                 dst_idx += i >> 2;
+                 val = ((cbp >> (5 - i)) & 1);
+                 v->mb_type[0][s->block_index[i]] = s->mb_intra;
+                 v->a_avail = v->c_avail = 0;
+                 if (i == 2 || i == 3 || !s->first_slice_line)
+                     v->a_avail = v->mb_type[0][s->block_index[i] - s->block_wrap[i]];
+                 if (i == 1 || i == 3 || s->mb_x)
+                     v->c_avail = v->mb_type[0][s->block_index[i] - 1];
+                 vc1_decode_intra_block(v, s->block[i], i, val, mquant,
+                                        (i & 4) ? v->codingset2 : v->codingset);
+                 if ((i>3) && (s->flags & CODEC_FLAG_GRAY)) continue;
+                 v->vc1dsp.vc1_inv_trans_8x8(s->block[i]);
+                 if (i < 4) {
+                     stride_y = s->linesize << fieldtx;
+                     off = (fieldtx) ? ((i & 1) * 8) + ((i & 2) >> 1) * s->linesize : (i & 1) * 8 + 4 * (i & 2) * s->linesize;
+                 } else {
+                     stride_y = s->uvlinesize;
+                     off = 0;
+                 }
+                 s->idsp.put_signed_pixels_clamped(s->block[i],
+                                                   s->dest[dst_idx] + off,
+                                                   stride_y);
+                 //TODO: loop filter
+             }
+         } else { // inter MB
+             mb_has_coeffs = ff_vc1_mbmode_intfrp[v->fourmvswitch][idx_mbmode][3];
+             if (mb_has_coeffs)
+                 cbp = 1 + get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_CBPCY_P_VLC_BITS, 2);
+             if (ff_vc1_mbmode_intfrp[v->fourmvswitch][idx_mbmode][0] == MV_PMODE_INTFR_2MV_FIELD) {
+                 v->twomvbp = get_vlc2(gb, v->twomvbp_vlc->table, VC1_2MV_BLOCK_PATTERN_VLC_BITS, 1);
+             } else {
+                 if ((ff_vc1_mbmode_intfrp[v->fourmvswitch][idx_mbmode][0] == MV_PMODE_INTFR_4MV)
+                     || (ff_vc1_mbmode_intfrp[v->fourmvswitch][idx_mbmode][0] == MV_PMODE_INTFR_4MV_FIELD)) {
+                     v->fourmvbp = get_vlc2(gb, v->fourmvbp_vlc->table, VC1_4MV_BLOCK_PATTERN_VLC_BITS, 1);
+                 }
+             }
+             s->mb_intra = v->is_intra[s->mb_x] = 0;
+             for (i = 0; i < 6; i++)
+                 v->mb_type[0][s->block_index[i]] = 0;
+             fieldtx = v->fieldtx_plane[mb_pos] = ff_vc1_mbmode_intfrp[v->fourmvswitch][idx_mbmode][1];
+             /* for all motion vector read MVDATA and motion compensate each block */
+             dst_idx = 0;
+             if (fourmv) {
+                 mvbp = v->fourmvbp;
+                 for (i = 0; i < 6; i++) {
+                     if (i < 4) {
+                         dmv_x = dmv_y = 0;
+                         val   = ((mvbp >> (3 - i)) & 1);
+                         if (val) {
+                             get_mvdata_interlaced(v, &dmv_x, &dmv_y, 0);
+                         }
+                         ff_vc1_pred_mv_intfr(v, i, dmv_x, dmv_y, 0, v->range_x, v->range_y, v->mb_type[0], 0);
+                         ff_vc1_mc_4mv_luma(v, i, 0, 0);
+                     } else if (i == 4) {
+                         ff_vc1_mc_4mv_chroma4(v, 0, 0, 0);
+                     }
+                 }
+             } else if (twomv) {
+                 mvbp  = v->twomvbp;
+                 dmv_x = dmv_y = 0;
+                 if (mvbp & 2) {
+                     get_mvdata_interlaced(v, &dmv_x, &dmv_y, 0);
+                 }
+                 ff_vc1_pred_mv_intfr(v, 0, dmv_x, dmv_y, 2, v->range_x, v->range_y, v->mb_type[0], 0);
+                 ff_vc1_mc_4mv_luma(v, 0, 0, 0);
+                 ff_vc1_mc_4mv_luma(v, 1, 0, 0);
+                 dmv_x = dmv_y = 0;
+                 if (mvbp & 1) {
+                     get_mvdata_interlaced(v, &dmv_x, &dmv_y, 0);
+                 }
+                 ff_vc1_pred_mv_intfr(v, 2, dmv_x, dmv_y, 2, v->range_x, v->range_y, v->mb_type[0], 0);
+                 ff_vc1_mc_4mv_luma(v, 2, 0, 0);
+                 ff_vc1_mc_4mv_luma(v, 3, 0, 0);
+                 ff_vc1_mc_4mv_chroma4(v, 0, 0, 0);
+             } else {
+                 mvbp = ff_vc1_mbmode_intfrp[v->fourmvswitch][idx_mbmode][2];
+                 dmv_x = dmv_y = 0;
+                 if (mvbp) {
+                     get_mvdata_interlaced(v, &dmv_x, &dmv_y, 0);
+                 }
+                 ff_vc1_pred_mv_intfr(v, 0, dmv_x, dmv_y, 1, v->range_x, v->range_y, v->mb_type[0], 0);
+                 ff_vc1_mc_1mv(v, 0);
+             }
+             if (cbp)
+                 GET_MQUANT();  // p. 227
+             s->current_picture.qscale_table[mb_pos] = mquant;
+             if (!v->ttmbf && cbp)
+                 ttmb = get_vlc2(gb, ff_vc1_ttmb_vlc[v->tt_index].table, VC1_TTMB_VLC_BITS, 2);
+             for (i = 0; i < 6; i++) {
+                 s->dc_val[0][s->block_index[i]] = 0;
+                 dst_idx += i >> 2;
+                 val = ((cbp >> (5 - i)) & 1);
+                 if (!fieldtx)
+                     off = (i & 4) ? 0 : ((i & 1) * 8 + (i & 2) * 4 * s->linesize);
+                 else
+                     off = (i & 4) ? 0 : ((i & 1) * 8 + ((i > 1) * s->linesize));
+                 if (val) {
+                     pat = vc1_decode_p_block(v, s->block[i], i, mquant, ttmb,
+                                              first_block, s->dest[dst_idx] + off,
+                                              (i & 4) ? s->uvlinesize : (s->linesize << fieldtx),
+                                              (i & 4) && (s->flags & CODEC_FLAG_GRAY), &block_tt);
+                     block_cbp |= pat << (i << 2);
+                     if (!v->ttmbf && ttmb < 8)
+                         ttmb = -1;
+                     first_block = 0;
+                 }
+             }
+         }
+     } else { // skipped
+         s->mb_intra = v->is_intra[s->mb_x] = 0;
+         for (i = 0; i < 6; i++) {
+             v->mb_type[0][s->block_index[i]] = 0;
+             s->dc_val[0][s->block_index[i]] = 0;
+         }
+         s->current_picture.mb_type[mb_pos]      = MB_TYPE_SKIP;
+         s->current_picture.qscale_table[mb_pos] = 0;
+         v->blk_mv_type[s->block_index[0]] = 0;
+         v->blk_mv_type[s->block_index[1]] = 0;
+         v->blk_mv_type[s->block_index[2]] = 0;
+         v->blk_mv_type[s->block_index[3]] = 0;
+         ff_vc1_pred_mv_intfr(v, 0, 0, 0, 1, v->range_x, v->range_y, v->mb_type[0], 0);
+         ff_vc1_mc_1mv(v, 0);
+     }
+     if (s->mb_x == s->mb_width - 1)
+         memmove(v->is_intra_base, v->is_intra, sizeof(v->is_intra_base[0])*s->mb_stride);
+     return 0;
+ }
+ static int vc1_decode_p_mb_intfi(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     GetBitContext *gb = &s->gb;
+     int i;
+     int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     int cbp = 0; /* cbp decoding stuff */
+     int mqdiff, mquant; /* MB quantization */
+     int ttmb = v->ttfrm; /* MB Transform type */
+     int mb_has_coeffs = 1; /* last_flag */
+     int dmv_x, dmv_y; /* Differential MV components */
+     int val; /* temp values */
+     int first_block = 1;
+     int dst_idx, off;
 -    mquant = v->pq; /* Loosy initialization */
++    int pred_flag = 0;
+     int block_cbp = 0, pat, block_tt = 0;
+     int idx_mbmode = 0;
 -        s->mb_intra = v->is_intra[s->mb_x] = 1;
++    mquant = v->pq; /* Lossy initialization */
+     idx_mbmode = get_vlc2(gb, v->mbmode_vlc->table, VC1_IF_MBMODE_VLC_BITS, 2);
+     if (idx_mbmode <= 1) { // intra MB
 -    int idx_mbmode, interpmvp;
++        s->mb_intra = 1;
++        v->is_intra[s->mb_x] = 0x3F;
+         s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][0] = 0;
+         s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][1] = 0;
+         s->current_picture.mb_type[mb_pos + v->mb_off] = MB_TYPE_INTRA;
+         GET_MQUANT();
+         s->current_picture.qscale_table[mb_pos] = mquant;
+         /* Set DC scale - y and c use the same (not sure if necessary here) */
+         s->y_dc_scale = s->y_dc_scale_table[mquant];
+         s->c_dc_scale = s->c_dc_scale_table[mquant];
+         v->s.ac_pred  = v->acpred_plane[mb_pos] = get_bits1(gb);
+         mb_has_coeffs = idx_mbmode & 1;
+         if (mb_has_coeffs)
+             cbp = 1 + get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_ICBPCY_VLC_BITS, 2);
+         dst_idx = 0;
+         for (i = 0; i < 6; i++) {
+             s->dc_val[0][s->block_index[i]]  = 0;
+             v->mb_type[0][s->block_index[i]] = 1;
+             dst_idx += i >> 2;
+             val = ((cbp >> (5 - i)) & 1);
+             v->a_avail = v->c_avail = 0;
+             if (i == 2 || i == 3 || !s->first_slice_line)
+                 v->a_avail = v->mb_type[0][s->block_index[i] - s->block_wrap[i]];
+             if (i == 1 || i == 3 || s->mb_x)
+                 v->c_avail = v->mb_type[0][s->block_index[i] - 1];
+             vc1_decode_intra_block(v, s->block[i], i, val, mquant,
+                                    (i & 4) ? v->codingset2 : v->codingset);
+             if ((i>3) && (s->flags & CODEC_FLAG_GRAY))
+                 continue;
+             v->vc1dsp.vc1_inv_trans_8x8(s->block[i]);
+             off  = (i & 4) ? 0 : ((i & 1) * 8 + (i & 2) * 4 * s->linesize);
+             s->idsp.put_signed_pixels_clamped(s->block[i],
+                                               s->dest[dst_idx] + off,
+                                               (i & 4) ? s->uvlinesize
+                                                       : s->linesize);
+             // TODO: loop filter
+         }
+     } else {
+         s->mb_intra = v->is_intra[s->mb_x] = 0;
+         s->current_picture.mb_type[mb_pos + v->mb_off] = MB_TYPE_16x16;
+         for (i = 0; i < 6; i++) v->mb_type[0][s->block_index[i]] = 0;
+         if (idx_mbmode <= 5) { // 1-MV
+             dmv_x = dmv_y = pred_flag = 0;
+             if (idx_mbmode & 1) {
+                 get_mvdata_interlaced(v, &dmv_x, &dmv_y, &pred_flag);
+             }
+             ff_vc1_pred_mv(v, 0, dmv_x, dmv_y, 1, v->range_x, v->range_y, v->mb_type[0], pred_flag, 0);
+             ff_vc1_mc_1mv(v, 0);
+             mb_has_coeffs = !(idx_mbmode & 2);
+         } else { // 4-MV
+             v->fourmvbp = get_vlc2(gb, v->fourmvbp_vlc->table, VC1_4MV_BLOCK_PATTERN_VLC_BITS, 1);
+             for (i = 0; i < 6; i++) {
+                 if (i < 4) {
+                     dmv_x = dmv_y = pred_flag = 0;
+                     val   = ((v->fourmvbp >> (3 - i)) & 1);
+                     if (val) {
+                         get_mvdata_interlaced(v, &dmv_x, &dmv_y, &pred_flag);
+                     }
+                     ff_vc1_pred_mv(v, i, dmv_x, dmv_y, 0, v->range_x, v->range_y, v->mb_type[0], pred_flag, 0);
+                     ff_vc1_mc_4mv_luma(v, i, 0, 0);
+                 } else if (i == 4)
+                     ff_vc1_mc_4mv_chroma(v, 0);
+             }
+             mb_has_coeffs = idx_mbmode & 1;
+         }
+         if (mb_has_coeffs)
+             cbp = 1 + get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_CBPCY_P_VLC_BITS, 2);
+         if (cbp) {
+             GET_MQUANT();
+         }
+         s->current_picture.qscale_table[mb_pos] = mquant;
+         if (!v->ttmbf && cbp) {
+             ttmb = get_vlc2(gb, ff_vc1_ttmb_vlc[v->tt_index].table, VC1_TTMB_VLC_BITS, 2);
+         }
+         dst_idx = 0;
+         for (i = 0; i < 6; i++) {
+             s->dc_val[0][s->block_index[i]] = 0;
+             dst_idx += i >> 2;
+             val = ((cbp >> (5 - i)) & 1);
+             off = (i & 4) ? 0 : (i & 1) * 8 + (i & 2) * 4 * s->linesize;
+             if (val) {
+                 pat = vc1_decode_p_block(v, s->block[i], i, mquant, ttmb,
+                                          first_block, s->dest[dst_idx] + off,
+                                          (i & 4) ? s->uvlinesize : s->linesize,
+                                          (i & 4) && (s->flags & CODEC_FLAG_GRAY),
+                                          &block_tt);
+                 block_cbp |= pat << (i << 2);
+                 if (!v->ttmbf && ttmb < 8) ttmb = -1;
+                 first_block = 0;
+             }
+         }
+     }
+     if (s->mb_x == s->mb_width - 1)
+         memmove(v->is_intra_base, v->is_intra, sizeof(v->is_intra_base[0]) * s->mb_stride);
+     return 0;
+ }
+ /** Decode one B-frame MB (in Main profile)
+  */
+ static void vc1_decode_b_mb(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     GetBitContext *gb = &s->gb;
+     int i, j;
+     int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     int cbp = 0; /* cbp decoding stuff */
+     int mqdiff, mquant; /* MB quantization */
+     int ttmb = v->ttfrm; /* MB Transform type */
+     int mb_has_coeffs = 0; /* last_flag */
+     int index, index1; /* LUT indexes */
+     int val, sign; /* temp values */
+     int first_block = 1;
+     int dst_idx, off;
+     int skipped, direct;
+     int dmv_x[2], dmv_y[2];
+     int bmvtype = BMV_TYPE_BACKWARD;
+     mquant      = v->pq; /* lossy initialization */
+     s->mb_intra = 0;
+     if (v->dmb_is_raw)
+         direct = get_bits1(gb);
+     else
+         direct = v->direct_mb_plane[mb_pos];
+     if (v->skip_is_raw)
+         skipped = get_bits1(gb);
+     else
+         skipped = v->s.mbskip_table[mb_pos];
+     dmv_x[0] = dmv_x[1] = dmv_y[0] = dmv_y[1] = 0;
+     for (i = 0; i < 6; i++) {
+         v->mb_type[0][s->block_index[i]] = 0;
+         s->dc_val[0][s->block_index[i]]  = 0;
+     }
+     s->current_picture.qscale_table[mb_pos] = 0;
+     if (!direct) {
+         if (!skipped) {
+             GET_MVDATA(dmv_x[0], dmv_y[0]);
+             dmv_x[1] = dmv_x[0];
+             dmv_y[1] = dmv_y[0];
+         }
+         if (skipped || !s->mb_intra) {
+             bmvtype = decode012(gb);
+             switch (bmvtype) {
+             case 0:
+                 bmvtype = (v->bfraction >= (B_FRACTION_DEN/2)) ? BMV_TYPE_BACKWARD : BMV_TYPE_FORWARD;
+                 break;
+             case 1:
+                 bmvtype = (v->bfraction >= (B_FRACTION_DEN/2)) ? BMV_TYPE_FORWARD : BMV_TYPE_BACKWARD;
+                 break;
+             case 2:
+                 bmvtype  = BMV_TYPE_INTERPOLATED;
+                 dmv_x[0] = dmv_y[0] = 0;
+             }
+         }
+     }
+     for (i = 0; i < 6; i++)
+         v->mb_type[0][s->block_index[i]] = s->mb_intra;
+     if (skipped) {
+         if (direct)
+             bmvtype = BMV_TYPE_INTERPOLATED;
+         ff_vc1_pred_b_mv(v, dmv_x, dmv_y, direct, bmvtype);
+         vc1_b_mc(v, dmv_x, dmv_y, direct, bmvtype);
+         return;
+     }
+     if (direct) {
+         cbp = get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_CBPCY_P_VLC_BITS, 2);
+         GET_MQUANT();
+         s->mb_intra = 0;
+         s->current_picture.qscale_table[mb_pos] = mquant;
+         if (!v->ttmbf)
+             ttmb = get_vlc2(gb, ff_vc1_ttmb_vlc[v->tt_index].table, VC1_TTMB_VLC_BITS, 2);
+         dmv_x[0] = dmv_y[0] = dmv_x[1] = dmv_y[1] = 0;
+         ff_vc1_pred_b_mv(v, dmv_x, dmv_y, direct, bmvtype);
+         vc1_b_mc(v, dmv_x, dmv_y, direct, bmvtype);
+     } else {
+         if (!mb_has_coeffs && !s->mb_intra) {
+             /* no coded blocks - effectively skipped */
+             ff_vc1_pred_b_mv(v, dmv_x, dmv_y, direct, bmvtype);
+             vc1_b_mc(v, dmv_x, dmv_y, direct, bmvtype);
+             return;
+         }
+         if (s->mb_intra && !mb_has_coeffs) {
+             GET_MQUANT();
+             s->current_picture.qscale_table[mb_pos] = mquant;
+             s->ac_pred = get_bits1(gb);
+             cbp = 0;
+             ff_vc1_pred_b_mv(v, dmv_x, dmv_y, direct, bmvtype);
+         } else {
+             if (bmvtype == BMV_TYPE_INTERPOLATED) {
+                 GET_MVDATA(dmv_x[0], dmv_y[0]);
+                 if (!mb_has_coeffs) {
+                     /* interpolated skipped block */
+                     ff_vc1_pred_b_mv(v, dmv_x, dmv_y, direct, bmvtype);
+                     vc1_b_mc(v, dmv_x, dmv_y, direct, bmvtype);
+                     return;
+                 }
+             }
+             ff_vc1_pred_b_mv(v, dmv_x, dmv_y, direct, bmvtype);
+             if (!s->mb_intra) {
+                 vc1_b_mc(v, dmv_x, dmv_y, direct, bmvtype);
+             }
+             if (s->mb_intra)
+                 s->ac_pred = get_bits1(gb);
+             cbp = get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_CBPCY_P_VLC_BITS, 2);
+             GET_MQUANT();
+             s->current_picture.qscale_table[mb_pos] = mquant;
+             if (!v->ttmbf && !s->mb_intra && mb_has_coeffs)
+                 ttmb = get_vlc2(gb, ff_vc1_ttmb_vlc[v->tt_index].table, VC1_TTMB_VLC_BITS, 2);
+         }
+     }
+     dst_idx = 0;
+     for (i = 0; i < 6; i++) {
+         s->dc_val[0][s->block_index[i]] = 0;
+         dst_idx += i >> 2;
+         val = ((cbp >> (5 - i)) & 1);
+         off = (i & 4) ? 0 : ((i & 1) * 8 + (i & 2) * 4 * s->linesize);
+         v->mb_type[0][s->block_index[i]] = s->mb_intra;
+         if (s->mb_intra) {
+             /* check if prediction blocks A and C are available */
+             v->a_avail = v->c_avail = 0;
+             if (i == 2 || i == 3 || !s->first_slice_line)
+                 v->a_avail = v->mb_type[0][s->block_index[i] - s->block_wrap[i]];
+             if (i == 1 || i == 3 || s->mb_x)
+                 v->c_avail = v->mb_type[0][s->block_index[i] - 1];
+             vc1_decode_intra_block(v, s->block[i], i, val, mquant,
+                                    (i & 4) ? v->codingset2 : v->codingset);
+             if ((i>3) && (s->flags & CODEC_FLAG_GRAY))
+                 continue;
+             v->vc1dsp.vc1_inv_trans_8x8(s->block[i]);
+             if (v->rangeredfrm)
+                 for (j = 0; j < 64; j++)
+                     s->block[i][j] <<= 1;
+             s->idsp.put_signed_pixels_clamped(s->block[i],
+                                               s->dest[dst_idx] + off,
+                                               i & 4 ? s->uvlinesize
+                                                     : s->linesize);
+         } else if (val) {
+             vc1_decode_p_block(v, s->block[i], i, mquant, ttmb,
+                                first_block, s->dest[dst_idx] + off,
+                                (i & 4) ? s->uvlinesize : s->linesize,
+                                (i & 4) && (s->flags & CODEC_FLAG_GRAY), NULL);
+             if (!v->ttmbf && ttmb < 8)
+                 ttmb = -1;
+             first_block = 0;
+         }
+     }
+ }
+ /** Decode one B-frame MB (in interlaced field B picture)
+  */
+ static void vc1_decode_b_mb_intfi(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     GetBitContext *gb = &s->gb;
+     int i, j;
+     int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     int cbp = 0; /* cbp decoding stuff */
+     int mqdiff, mquant; /* MB quantization */
+     int ttmb = v->ttfrm; /* MB Transform type */
+     int mb_has_coeffs = 0; /* last_flag */
+     int val; /* temp value */
+     int first_block = 1;
+     int dst_idx, off;
+     int fwd;
+     int dmv_x[2], dmv_y[2], pred_flag[2];
+     int bmvtype = BMV_TYPE_BACKWARD;
 -    mquant      = v->pq; /* Loosy initialization */
++    int idx_mbmode;
 -        s->mb_intra = v->is_intra[s->mb_x] = 1;
++    mquant      = v->pq; /* Lossy initialization */
+     s->mb_intra = 0;
+     idx_mbmode = get_vlc2(gb, v->mbmode_vlc->table, VC1_IF_MBMODE_VLC_BITS, 2);
+     if (idx_mbmode <= 1) { // intra MB
 -            if (bmvtype == BMV_TYPE_INTERPOLATED && interpmvp) {
++        s->mb_intra = 1;
++        v->is_intra[s->mb_x] = 0x3F;
+         s->current_picture.motion_val[1][s->block_index[0]][0] = 0;
+         s->current_picture.motion_val[1][s->block_index[0]][1] = 0;
+         s->current_picture.mb_type[mb_pos + v->mb_off]         = MB_TYPE_INTRA;
+         GET_MQUANT();
+         s->current_picture.qscale_table[mb_pos] = mquant;
+         /* Set DC scale - y and c use the same (not sure if necessary here) */
+         s->y_dc_scale = s->y_dc_scale_table[mquant];
+         s->c_dc_scale = s->c_dc_scale_table[mquant];
+         v->s.ac_pred  = v->acpred_plane[mb_pos] = get_bits1(gb);
+         mb_has_coeffs = idx_mbmode & 1;
+         if (mb_has_coeffs)
+             cbp = 1 + get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_ICBPCY_VLC_BITS, 2);
+         dst_idx = 0;
+         for (i = 0; i < 6; i++) {
+             s->dc_val[0][s->block_index[i]] = 0;
+             dst_idx += i >> 2;
+             val = ((cbp >> (5 - i)) & 1);
+             v->mb_type[0][s->block_index[i]] = s->mb_intra;
+             v->a_avail                       = v->c_avail = 0;
+             if (i == 2 || i == 3 || !s->first_slice_line)
+                 v->a_avail = v->mb_type[0][s->block_index[i] - s->block_wrap[i]];
+             if (i == 1 || i == 3 || s->mb_x)
+                 v->c_avail = v->mb_type[0][s->block_index[i] - 1];
+             vc1_decode_intra_block(v, s->block[i], i, val, mquant,
+                                    (i & 4) ? v->codingset2 : v->codingset);
+             if ((i>3) && (s->flags & CODEC_FLAG_GRAY))
+                 continue;
+             v->vc1dsp.vc1_inv_trans_8x8(s->block[i]);
+             if (v->rangeredfrm)
+                 for (j = 0; j < 64; j++)
+                     s->block[i][j] <<= 1;
+             off  = (i & 4) ? 0 : ((i & 1) * 8 + (i & 2) * 4 * s->linesize);
+             s->idsp.put_signed_pixels_clamped(s->block[i],
+                                               s->dest[dst_idx] + off,
+                                               (i & 4) ? s->uvlinesize
+                                                       : s->linesize);
+             // TODO: yet to perform loop filter
+         }
+     } else {
+         s->mb_intra = v->is_intra[s->mb_x] = 0;
+         s->current_picture.mb_type[mb_pos + v->mb_off] = MB_TYPE_16x16;
+         for (i = 0; i < 6; i++) v->mb_type[0][s->block_index[i]] = 0;
+         if (v->fmb_is_raw)
+             fwd = v->forward_mb_plane[mb_pos] = get_bits1(gb);
+         else
+             fwd = v->forward_mb_plane[mb_pos];
+         if (idx_mbmode <= 5) { // 1-MV
++            int interpmvp = 0;
+             dmv_x[0]     = dmv_x[1] = dmv_y[0] = dmv_y[1] = 0;
+             pred_flag[0] = pred_flag[1] = 0;
+             if (fwd)
+                 bmvtype = BMV_TYPE_FORWARD;
+             else {
+                 bmvtype = decode012(gb);
+                 switch (bmvtype) {
+                 case 0:
+                     bmvtype = BMV_TYPE_BACKWARD;
+                     break;
+                 case 1:
+                     bmvtype = BMV_TYPE_DIRECT;
+                     break;
+                 case 2:
+                     bmvtype   = BMV_TYPE_INTERPOLATED;
+                     interpmvp = get_bits1(gb);
+                 }
+             }
+             v->bmvtype = bmvtype;
+             if (bmvtype != BMV_TYPE_DIRECT && idx_mbmode & 1) {
+                 get_mvdata_interlaced(v, &dmv_x[bmvtype == BMV_TYPE_BACKWARD], &dmv_y[bmvtype == BMV_TYPE_BACKWARD], &pred_flag[bmvtype == BMV_TYPE_BACKWARD]);
+             }
 -        s->mb_intra = v->is_intra[s->mb_x] = 1;
++            if (interpmvp) {
+                 get_mvdata_interlaced(v, &dmv_x[1], &dmv_y[1], &pred_flag[1]);
+             }
+             if (bmvtype == BMV_TYPE_DIRECT) {
+                 dmv_x[0] = dmv_y[0] = pred_flag[0] = 0;
+                 dmv_x[1] = dmv_y[1] = pred_flag[0] = 0;
++                if (!s->next_picture_ptr->field_picture) {
++                    av_log(s->avctx, AV_LOG_ERROR, "Mixed field/frame direct mode not supported\n");
++                    return;
++                }
+             }
+             ff_vc1_pred_b_mv_intfi(v, 0, dmv_x, dmv_y, 1, pred_flag);
+             vc1_b_mc(v, dmv_x, dmv_y, (bmvtype == BMV_TYPE_DIRECT), bmvtype);
+             mb_has_coeffs = !(idx_mbmode & 2);
+         } else { // 4-MV
+             if (fwd)
+                 bmvtype = BMV_TYPE_FORWARD;
+             v->bmvtype  = bmvtype;
+             v->fourmvbp = get_vlc2(gb, v->fourmvbp_vlc->table, VC1_4MV_BLOCK_PATTERN_VLC_BITS, 1);
+             for (i = 0; i < 6; i++) {
+                 if (i < 4) {
+                     dmv_x[0] = dmv_y[0] = pred_flag[0] = 0;
+                     dmv_x[1] = dmv_y[1] = pred_flag[1] = 0;
+                     val = ((v->fourmvbp >> (3 - i)) & 1);
+                     if (val) {
+                         get_mvdata_interlaced(v, &dmv_x[bmvtype == BMV_TYPE_BACKWARD],
+                                                  &dmv_y[bmvtype == BMV_TYPE_BACKWARD],
+                                              &pred_flag[bmvtype == BMV_TYPE_BACKWARD]);
+                     }
+                     ff_vc1_pred_b_mv_intfi(v, i, dmv_x, dmv_y, 0, pred_flag);
+                     ff_vc1_mc_4mv_luma(v, i, bmvtype == BMV_TYPE_BACKWARD, 0);
+                 } else if (i == 4)
+                     ff_vc1_mc_4mv_chroma(v, bmvtype == BMV_TYPE_BACKWARD);
+             }
+             mb_has_coeffs = idx_mbmode & 1;
+         }
+         if (mb_has_coeffs)
+             cbp = 1 + get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_CBPCY_P_VLC_BITS, 2);
+         if (cbp) {
+             GET_MQUANT();
+         }
+         s->current_picture.qscale_table[mb_pos] = mquant;
+         if (!v->ttmbf && cbp) {
+             ttmb = get_vlc2(gb, ff_vc1_ttmb_vlc[v->tt_index].table, VC1_TTMB_VLC_BITS, 2);
+         }
+         dst_idx = 0;
+         for (i = 0; i < 6; i++) {
+             s->dc_val[0][s->block_index[i]] = 0;
+             dst_idx += i >> 2;
+             val = ((cbp >> (5 - i)) & 1);
+             off = (i & 4) ? 0 : (i & 1) * 8 + (i & 2) * 4 * s->linesize;
+             if (val) {
+                 vc1_decode_p_block(v, s->block[i], i, mquant, ttmb,
+                                    first_block, s->dest[dst_idx] + off,
+                                    (i & 4) ? s->uvlinesize : s->linesize,
+                                    (i & 4) && (s->flags & CODEC_FLAG_GRAY), NULL);
+                 if (!v->ttmbf && ttmb < 8)
+                     ttmb = -1;
+                 first_block = 0;
+             }
+         }
+     }
+ }
+ /** Decode one B-frame MB (in interlaced frame B picture)
+  */
+ static int vc1_decode_b_mb_intfr(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     GetBitContext *gb = &s->gb;
+     int i, j;
+     int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     int cbp = 0; /* cbp decoding stuff */
+     int mqdiff, mquant; /* MB quantization */
+     int ttmb = v->ttfrm; /* MB Transform type */
+     int mvsw = 0; /* motion vector switch */
+     int mb_has_coeffs = 1; /* last_flag */
+     int dmv_x, dmv_y; /* Differential MV components */
+     int val; /* temp value */
+     int first_block = 1;
+     int dst_idx, off;
+     int skipped, direct, twomv = 0;
+     int block_cbp = 0, pat, block_tt = 0;
+     int idx_mbmode = 0, mvbp;
+     int stride_y, fieldtx;
+     int bmvtype = BMV_TYPE_BACKWARD;
+     int dir, dir2;
+     mquant = v->pq; /* Lossy initialization */
+     s->mb_intra = 0;
+     if (v->skip_is_raw)
+         skipped = get_bits1(gb);
+     else
+         skipped = v->s.mbskip_table[mb_pos];
+     if (!skipped) {
+         idx_mbmode = get_vlc2(gb, v->mbmode_vlc->table, VC1_INTFR_NON4MV_MBMODE_VLC_BITS, 2);
+         if (ff_vc1_mbmode_intfrp[0][idx_mbmode][0] == MV_PMODE_INTFR_2MV_FIELD) {
+             twomv = 1;
+             v->blk_mv_type[s->block_index[0]] = 1;
+             v->blk_mv_type[s->block_index[1]] = 1;
+             v->blk_mv_type[s->block_index[2]] = 1;
+             v->blk_mv_type[s->block_index[3]] = 1;
+         } else {
+             v->blk_mv_type[s->block_index[0]] = 0;
+             v->blk_mv_type[s->block_index[1]] = 0;
+             v->blk_mv_type[s->block_index[2]] = 0;
+             v->blk_mv_type[s->block_index[3]] = 0;
+         }
+     }
+     if (v->dmb_is_raw)
+         direct = get_bits1(gb);
+     else
+         direct = v->direct_mb_plane[mb_pos];
+     if (direct) {
++        if (s->next_picture_ptr->field_picture)
++            av_log(s->avctx, AV_LOG_WARNING, "Mixed frame/field direct mode not supported\n");
+         s->mv[0][0][0] = s->current_picture.motion_val[0][s->block_index[0]][0] = scale_mv(s->next_picture.motion_val[1][s->block_index[0]][0], v->bfraction, 0, s->quarter_sample);
+         s->mv[0][0][1] = s->current_picture.motion_val[0][s->block_index[0]][1] = scale_mv(s->next_picture.motion_val[1][s->block_index[0]][1], v->bfraction, 0, s->quarter_sample);
+         s->mv[1][0][0] = s->current_picture.motion_val[1][s->block_index[0]][0] = scale_mv(s->next_picture.motion_val[1][s->block_index[0]][0], v->bfraction, 1, s->quarter_sample);
+         s->mv[1][0][1] = s->current_picture.motion_val[1][s->block_index[0]][1] = scale_mv(s->next_picture.motion_val[1][s->block_index[0]][1], v->bfraction, 1, s->quarter_sample);
+         if (twomv) {
+             s->mv[0][2][0] = s->current_picture.motion_val[0][s->block_index[2]][0] = scale_mv(s->next_picture.motion_val[1][s->block_index[2]][0], v->bfraction, 0, s->quarter_sample);
+             s->mv[0][2][1] = s->current_picture.motion_val[0][s->block_index[2]][1] = scale_mv(s->next_picture.motion_val[1][s->block_index[2]][1], v->bfraction, 0, s->quarter_sample);
+             s->mv[1][2][0] = s->current_picture.motion_val[1][s->block_index[2]][0] = scale_mv(s->next_picture.motion_val[1][s->block_index[2]][0], v->bfraction, 1, s->quarter_sample);
+             s->mv[1][2][1] = s->current_picture.motion_val[1][s->block_index[2]][1] = scale_mv(s->next_picture.motion_val[1][s->block_index[2]][1], v->bfraction, 1, s->quarter_sample);
+             for (i = 1; i < 4; i += 2) {
+                 s->mv[0][i][0] = s->current_picture.motion_val[0][s->block_index[i]][0] = s->mv[0][i-1][0];
+                 s->mv[0][i][1] = s->current_picture.motion_val[0][s->block_index[i]][1] = s->mv[0][i-1][1];
+                 s->mv[1][i][0] = s->current_picture.motion_val[1][s->block_index[i]][0] = s->mv[1][i-1][0];
+                 s->mv[1][i][1] = s->current_picture.motion_val[1][s->block_index[i]][1] = s->mv[1][i-1][1];
+             }
+         } else {
+             for (i = 1; i < 4; i++) {
+                 s->mv[0][i][0] = s->current_picture.motion_val[0][s->block_index[i]][0] = s->mv[0][0][0];
+                 s->mv[0][i][1] = s->current_picture.motion_val[0][s->block_index[i]][1] = s->mv[0][0][1];
+                 s->mv[1][i][0] = s->current_picture.motion_val[1][s->block_index[i]][0] = s->mv[1][0][0];
+                 s->mv[1][i][1] = s->current_picture.motion_val[1][s->block_index[i]][1] = s->mv[1][0][1];
+             }
+         }
+     }
+     if (ff_vc1_mbmode_intfrp[0][idx_mbmode][0] == MV_PMODE_INTFR_INTRA) { // intra MB
+         for (i = 0; i < 4; i++) {
+             s->mv[0][i][0] = s->current_picture.motion_val[0][s->block_index[i]][0] = 0;
+             s->mv[0][i][1] = s->current_picture.motion_val[0][s->block_index[i]][1] = 0;
+             s->mv[1][i][0] = s->current_picture.motion_val[1][s->block_index[i]][0] = 0;
+             s->mv[1][i][1] = s->current_picture.motion_val[1][s->block_index[i]][1] = 0;
+         }
+         s->current_picture.mb_type[mb_pos] = MB_TYPE_INTRA;
++        s->mb_intra = 1;
++        v->is_intra[s->mb_x] = 0x3F;
+         for (i = 0; i < 6; i++)
+             v->mb_type[0][s->block_index[i]] = 1;
+         fieldtx = v->fieldtx_plane[mb_pos] = get_bits1(gb);
+         mb_has_coeffs = get_bits1(gb);
+         if (mb_has_coeffs)
+             cbp = 1 + get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_CBPCY_P_VLC_BITS, 2);
+         v->s.ac_pred = v->acpred_plane[mb_pos] = get_bits1(gb);
+         GET_MQUANT();
+         s->current_picture.qscale_table[mb_pos] = mquant;
+         /* Set DC scale - y and c use the same (not sure if necessary here) */
+         s->y_dc_scale = s->y_dc_scale_table[mquant];
+         s->c_dc_scale = s->c_dc_scale_table[mquant];
+         dst_idx = 0;
+         for (i = 0; i < 6; i++) {
+             s->dc_val[0][s->block_index[i]] = 0;
+             dst_idx += i >> 2;
+             val = ((cbp >> (5 - i)) & 1);
+             v->mb_type[0][s->block_index[i]] = s->mb_intra;
+             v->a_avail = v->c_avail = 0;
+             if (i == 2 || i == 3 || !s->first_slice_line)
+                 v->a_avail = v->mb_type[0][s->block_index[i] - s->block_wrap[i]];
+             if (i == 1 || i == 3 || s->mb_x)
+                 v->c_avail = v->mb_type[0][s->block_index[i] - 1];
+             vc1_decode_intra_block(v, s->block[i], i, val, mquant,
+                                    (i & 4) ? v->codingset2 : v->codingset);
+             if (i > 3 && (s->flags & CODEC_FLAG_GRAY))
+                 continue;
+             v->vc1dsp.vc1_inv_trans_8x8(s->block[i]);
+             if (i < 4) {
+                 stride_y = s->linesize << fieldtx;
+                 off = (fieldtx) ? ((i & 1) * 8) + ((i & 2) >> 1) * s->linesize : (i & 1) * 8 + 4 * (i & 2) * s->linesize;
+             } else {
+                 stride_y = s->uvlinesize;
+                 off = 0;
+             }
+             s->idsp.put_signed_pixels_clamped(s->block[i],
+                                               s->dest[dst_idx] + off,
+                                               stride_y);
+         }
+     } else {
+         s->mb_intra = v->is_intra[s->mb_x] = 0;
+         if (!direct) {
+             if (skipped || !s->mb_intra) {
+                 bmvtype = decode012(gb);
+                 switch (bmvtype) {
+                 case 0:
+                     bmvtype = (v->bfraction >= (B_FRACTION_DEN/2)) ? BMV_TYPE_BACKWARD : BMV_TYPE_FORWARD;
+                     break;
+                 case 1:
+                     bmvtype = (v->bfraction >= (B_FRACTION_DEN/2)) ? BMV_TYPE_FORWARD : BMV_TYPE_BACKWARD;
+                     break;
+                 case 2:
+                     bmvtype  = BMV_TYPE_INTERPOLATED;
+                 }
+             }
+             if (twomv && bmvtype != BMV_TYPE_INTERPOLATED)
+                 mvsw = get_bits1(gb);
+         }
+         if (!skipped) { // inter MB
+             mb_has_coeffs = ff_vc1_mbmode_intfrp[0][idx_mbmode][3];
+             if (mb_has_coeffs)
+                 cbp = 1 + get_vlc2(&v->s.gb, v->cbpcy_vlc->table, VC1_CBPCY_P_VLC_BITS, 2);
+             if (!direct) {
+                 if (bmvtype == BMV_TYPE_INTERPOLATED && twomv) {
+                     v->fourmvbp = get_vlc2(gb, v->fourmvbp_vlc->table, VC1_4MV_BLOCK_PATTERN_VLC_BITS, 1);
+                 } else if (bmvtype == BMV_TYPE_INTERPOLATED || twomv) {
+                     v->twomvbp = get_vlc2(gb, v->twomvbp_vlc->table, VC1_2MV_BLOCK_PATTERN_VLC_BITS, 1);
+                 }
+             }
+             for (i = 0; i < 6; i++)
+                 v->mb_type[0][s->block_index[i]] = 0;
+             fieldtx = v->fieldtx_plane[mb_pos] = ff_vc1_mbmode_intfrp[0][idx_mbmode][1];
+             /* for all motion vector read MVDATA and motion compensate each block */
+             dst_idx = 0;
+             if (direct) {
+                 if (twomv) {
+                     for (i = 0; i < 4; i++) {
+                         ff_vc1_mc_4mv_luma(v, i, 0, 0);
+                         ff_vc1_mc_4mv_luma(v, i, 1, 1);
+                     }
+                     ff_vc1_mc_4mv_chroma4(v, 0, 0, 0);
+                     ff_vc1_mc_4mv_chroma4(v, 1, 1, 1);
+                 } else {
+                     ff_vc1_mc_1mv(v, 0);
+                     ff_vc1_interp_mc(v);
+                 }
+             } else if (twomv && bmvtype == BMV_TYPE_INTERPOLATED) {
+                 mvbp = v->fourmvbp;
+                 for (i = 0; i < 4; i++) {
+                     dir = i==1 || i==3;
+                     dmv_x = dmv_y = 0;
+                     val = ((mvbp >> (3 - i)) & 1);
+                     if (val)
+                         get_mvdata_interlaced(v, &dmv_x, &dmv_y, 0);
+                     j = i > 1 ? 2 : 0;
+                     ff_vc1_pred_mv_intfr(v, j, dmv_x, dmv_y, 2, v->range_x, v->range_y, v->mb_type[0], dir);
+                     ff_vc1_mc_4mv_luma(v, j, dir, dir);
+                     ff_vc1_mc_4mv_luma(v, j+1, dir, dir);
+                 }
+                 ff_vc1_mc_4mv_chroma4(v, 0, 0, 0);
+                 ff_vc1_mc_4mv_chroma4(v, 1, 1, 1);
+             } else if (bmvtype == BMV_TYPE_INTERPOLATED) {
+                 mvbp = v->twomvbp;
+                 dmv_x = dmv_y = 0;
+                 if (mvbp & 2)
+                     get_mvdata_interlaced(v, &dmv_x, &dmv_y, 0);
+                 ff_vc1_pred_mv_intfr(v, 0, dmv_x, dmv_y, 1, v->range_x, v->range_y, v->mb_type[0], 0);
+                 ff_vc1_mc_1mv(v, 0);
+                 dmv_x = dmv_y = 0;
+                 if (mvbp & 1)
+                     get_mvdata_interlaced(v, &dmv_x, &dmv_y, 0);
+                 ff_vc1_pred_mv_intfr(v, 0, dmv_x, dmv_y, 1, v->range_x, v->range_y, v->mb_type[0], 1);
+                 ff_vc1_interp_mc(v);
+             } else if (twomv) {
+                 dir = bmvtype == BMV_TYPE_BACKWARD;
+                 dir2 = dir;
+                 if (mvsw)
+                     dir2 = !dir;
+                 mvbp = v->twomvbp;
+                 dmv_x = dmv_y = 0;
+                 if (mvbp & 2)
+                     get_mvdata_interlaced(v, &dmv_x, &dmv_y, 0);
+                 ff_vc1_pred_mv_intfr(v, 0, dmv_x, dmv_y, 2, v->range_x, v->range_y, v->mb_type[0], dir);
+                 dmv_x = dmv_y = 0;
+                 if (mvbp & 1)
+                     get_mvdata_interlaced(v, &dmv_x, &dmv_y, 0);
+                 ff_vc1_pred_mv_intfr(v, 2, dmv_x, dmv_y, 2, v->range_x, v->range_y, v->mb_type[0], dir2);
+                 if (mvsw) {
+                     for (i = 0; i < 2; i++) {
+                         s->mv[dir][i+2][0] = s->mv[dir][i][0] = s->current_picture.motion_val[dir][s->block_index[i+2]][0] = s->current_picture.motion_val[dir][s->block_index[i]][0];
+                         s->mv[dir][i+2][1] = s->mv[dir][i][1] = s->current_picture.motion_val[dir][s->block_index[i+2]][1] = s->current_picture.motion_val[dir][s->block_index[i]][1];
+                         s->mv[dir2][i+2][0] = s->mv[dir2][i][0] = s->current_picture.motion_val[dir2][s->block_index[i]][0] = s->current_picture.motion_val[dir2][s->block_index[i+2]][0];
+                         s->mv[dir2][i+2][1] = s->mv[dir2][i][1] = s->current_picture.motion_val[dir2][s->block_index[i]][1] = s->current_picture.motion_val[dir2][s->block_index[i+2]][1];
+                     }
+                 } else {
+                     ff_vc1_pred_mv_intfr(v, 0, 0, 0, 2, v->range_x, v->range_y, v->mb_type[0], !dir);
+                     ff_vc1_pred_mv_intfr(v, 2, 0, 0, 2, v->range_x, v->range_y, v->mb_type[0], !dir);
+                 }
+                 ff_vc1_mc_4mv_luma(v, 0, dir, 0);
+                 ff_vc1_mc_4mv_luma(v, 1, dir, 0);
+                 ff_vc1_mc_4mv_luma(v, 2, dir2, 0);
+                 ff_vc1_mc_4mv_luma(v, 3, dir2, 0);
+                 ff_vc1_mc_4mv_chroma4(v, dir, dir2, 0);
+             } else {
+                 dir = bmvtype == BMV_TYPE_BACKWARD;
+                 mvbp = ff_vc1_mbmode_intfrp[0][idx_mbmode][2];
+                 dmv_x = dmv_y = 0;
+                 if (mvbp)
+                     get_mvdata_interlaced(v, &dmv_x, &dmv_y, 0);
+                 ff_vc1_pred_mv_intfr(v, 0, dmv_x, dmv_y, 1, v->range_x, v->range_y, v->mb_type[0], dir);
+                 v->blk_mv_type[s->block_index[0]] = 1;
+                 v->blk_mv_type[s->block_index[1]] = 1;
+                 v->blk_mv_type[s->block_index[2]] = 1;
+                 v->blk_mv_type[s->block_index[3]] = 1;
+                 ff_vc1_pred_mv_intfr(v, 0, 0, 0, 2, v->range_x, v->range_y, 0, !dir);
+                 for (i = 0; i < 2; i++) {
+                     s->mv[!dir][i+2][0] = s->mv[!dir][i][0] = s->current_picture.motion_val[!dir][s->block_index[i+2]][0] = s->current_picture.motion_val[!dir][s->block_index[i]][0];
+                     s->mv[!dir][i+2][1] = s->mv[!dir][i][1] = s->current_picture.motion_val[!dir][s->block_index[i+2]][1] = s->current_picture.motion_val[!dir][s->block_index[i]][1];
+                 }
+                 ff_vc1_mc_1mv(v, dir);
+             }
+             if (cbp)
+                 GET_MQUANT();  // p. 227
+             s->current_picture.qscale_table[mb_pos] = mquant;
+             if (!v->ttmbf && cbp)
+                 ttmb = get_vlc2(gb, ff_vc1_ttmb_vlc[v->tt_index].table, VC1_TTMB_VLC_BITS, 2);
+             for (i = 0; i < 6; i++) {
+                 s->dc_val[0][s->block_index[i]] = 0;
+                 dst_idx += i >> 2;
+                 val = ((cbp >> (5 - i)) & 1);
+                 if (!fieldtx)
+                     off = (i & 4) ? 0 : ((i & 1) * 8 + (i & 2) * 4 * s->linesize);
+                 else
+                     off = (i & 4) ? 0 : ((i & 1) * 8 + ((i > 1) * s->linesize));
+                 if (val) {
+                     pat = vc1_decode_p_block(v, s->block[i], i, mquant, ttmb,
+                                              first_block, s->dest[dst_idx] + off,
+                                              (i & 4) ? s->uvlinesize : (s->linesize << fieldtx),
+                                              (i & 4) && (s->flags & CODEC_FLAG_GRAY), &block_tt);
+                     block_cbp |= pat << (i << 2);
+                     if (!v->ttmbf && ttmb < 8)
+                         ttmb = -1;
+                     first_block = 0;
+                 }
+             }
+         } else { // skipped
+             dir = 0;
+             for (i = 0; i < 6; i++) {
+                 v->mb_type[0][s->block_index[i]] = 0;
+                 s->dc_val[0][s->block_index[i]] = 0;
+             }
+             s->current_picture.mb_type[mb_pos]      = MB_TYPE_SKIP;
+             s->current_picture.qscale_table[mb_pos] = 0;
+             v->blk_mv_type[s->block_index[0]] = 0;
+             v->blk_mv_type[s->block_index[1]] = 0;
+             v->blk_mv_type[s->block_index[2]] = 0;
+             v->blk_mv_type[s->block_index[3]] = 0;
+             if (!direct) {
+                 if (bmvtype == BMV_TYPE_INTERPOLATED) {
+                     ff_vc1_pred_mv_intfr(v, 0, 0, 0, 1, v->range_x, v->range_y, v->mb_type[0], 0);
+                     ff_vc1_pred_mv_intfr(v, 0, 0, 0, 1, v->range_x, v->range_y, v->mb_type[0], 1);
+                 } else {
+                     dir = bmvtype == BMV_TYPE_BACKWARD;
+                     ff_vc1_pred_mv_intfr(v, 0, 0, 0, 1, v->range_x, v->range_y, v->mb_type[0], dir);
+                     if (mvsw) {
+                         int dir2 = dir;
+                         if (mvsw)
+                             dir2 = !dir;
+                         for (i = 0; i < 2; i++) {
+                             s->mv[dir][i+2][0] = s->mv[dir][i][0] = s->current_picture.motion_val[dir][s->block_index[i+2]][0] = s->current_picture.motion_val[dir][s->block_index[i]][0];
+                             s->mv[dir][i+2][1] = s->mv[dir][i][1] = s->current_picture.motion_val[dir][s->block_index[i+2]][1] = s->current_picture.motion_val[dir][s->block_index[i]][1];
+                             s->mv[dir2][i+2][0] = s->mv[dir2][i][0] = s->current_picture.motion_val[dir2][s->block_index[i]][0] = s->current_picture.motion_val[dir2][s->block_index[i+2]][0];
+                             s->mv[dir2][i+2][1] = s->mv[dir2][i][1] = s->current_picture.motion_val[dir2][s->block_index[i]][1] = s->current_picture.motion_val[dir2][s->block_index[i+2]][1];
+                         }
+                     } else {
+                         v->blk_mv_type[s->block_index[0]] = 1;
+                         v->blk_mv_type[s->block_index[1]] = 1;
+                         v->blk_mv_type[s->block_index[2]] = 1;
+                         v->blk_mv_type[s->block_index[3]] = 1;
+                         ff_vc1_pred_mv_intfr(v, 0, 0, 0, 2, v->range_x, v->range_y, 0, !dir);
+                         for (i = 0; i < 2; i++) {
+                             s->mv[!dir][i+2][0] = s->mv[!dir][i][0] = s->current_picture.motion_val[!dir][s->block_index[i+2]][0] = s->current_picture.motion_val[!dir][s->block_index[i]][0];
+                             s->mv[!dir][i+2][1] = s->mv[!dir][i][1] = s->current_picture.motion_val[!dir][s->block_index[i+2]][1] = s->current_picture.motion_val[!dir][s->block_index[i]][1];
+                         }
+                     }
+                 }
+             }
+             ff_vc1_mc_1mv(v, dir);
+             if (direct || bmvtype == BMV_TYPE_INTERPOLATED) {
+                 ff_vc1_interp_mc(v);
+             }
+         }
+     }
+     if (s->mb_x == s->mb_width - 1)
+         memmove(v->is_intra_base, v->is_intra, sizeof(v->is_intra_base[0]) * s->mb_stride);
+     v->cbp[s->mb_x]      = block_cbp;
+     v->ttblk[s->mb_x]    = block_tt;
+     return 0;
+ }
+ /** Decode blocks of I-frame
+  */
+ static void vc1_decode_i_blocks(VC1Context *v)
+ {
+     int k, j;
+     MpegEncContext *s = &v->s;
+     int cbp, val;
+     uint8_t *coded_val;
+     int mb_pos;
+     /* select codingmode used for VLC tables selection */
+     switch (v->y_ac_table_index) {
+     case 0:
+         v->codingset = (v->pqindex <= 8) ? CS_HIGH_RATE_INTRA : CS_LOW_MOT_INTRA;
+         break;
+     case 1:
+         v->codingset = CS_HIGH_MOT_INTRA;
+         break;
+     case 2:
+         v->codingset = CS_MID_RATE_INTRA;
+         break;
+     }
+     switch (v->c_ac_table_index) {
+     case 0:
+         v->codingset2 = (v->pqindex <= 8) ? CS_HIGH_RATE_INTER : CS_LOW_MOT_INTER;
+         break;
+     case 1:
+         v->codingset2 = CS_HIGH_MOT_INTER;
+         break;
+     case 2:
+         v->codingset2 = CS_MID_RATE_INTER;
+         break;
+     }
+     /* Set DC scale - y and c use the same */
+     s->y_dc_scale = s->y_dc_scale_table[v->pq];
+     s->c_dc_scale = s->c_dc_scale_table[v->pq];
+     //do frame decode
+     s->mb_x = s->mb_y = 0;
+     s->mb_intra         = 1;
+     s->first_slice_line = 1;
+     for (s->mb_y = 0; s->mb_y < s->end_mb_y; s->mb_y++) {
+         s->mb_x = 0;
+         init_block_index(v);
+         for (; s->mb_x < v->end_mb_x; s->mb_x++) {
+             uint8_t *dst[6];
+             ff_update_block_index(s);
+             dst[0] = s->dest[0];
+             dst[1] = dst[0] + 8;
+             dst[2] = s->dest[0] + s->linesize * 8;
+             dst[3] = dst[2] + 8;
+             dst[4] = s->dest[1];
+             dst[5] = s->dest[2];
+             s->bdsp.clear_blocks(s->block[0]);
+             mb_pos = s->mb_x + s->mb_y * s->mb_width;
+             s->current_picture.mb_type[mb_pos]                     = MB_TYPE_INTRA;
+             s->current_picture.qscale_table[mb_pos]                = v->pq;
+             s->current_picture.motion_val[1][s->block_index[0]][0] = 0;
+             s->current_picture.motion_val[1][s->block_index[0]][1] = 0;
+             // do actual MB decoding and displaying
+             cbp = get_vlc2(&v->s.gb, ff_msmp4_mb_i_vlc.table, MB_INTRA_VLC_BITS, 2);
+             v->s.ac_pred = get_bits1(&v->s.gb);
+             for (k = 0; k < 6; k++) {
+                 val = ((cbp >> (5 - k)) & 1);
+                 if (k < 4) {
+                     int pred   = vc1_coded_block_pred(&v->s, k, &coded_val);
+                     val        = val ^ pred;
+                     *coded_val = val;
+                 }
+                 cbp |= val << (5 - k);
+                 vc1_decode_i_block(v, s->block[k], k, val, (k < 4) ? v->codingset : v->codingset2);
+                 if (k > 3 && (s->flags & CODEC_FLAG_GRAY))
+                     continue;
+                 v->vc1dsp.vc1_inv_trans_8x8(s->block[k]);
+                 if (v->pq >= 9 && v->overlap) {
+                     if (v->rangeredfrm)
+                         for (j = 0; j < 64; j++)
+                             s->block[k][j] <<= 1;
+                     s->idsp.put_signed_pixels_clamped(s->block[k], dst[k],
+                                                       k & 4 ? s->uvlinesize
+                                                             : s->linesize);
+                 } else {
+                     if (v->rangeredfrm)
+                         for (j = 0; j < 64; j++)
+                             s->block[k][j] = (s->block[k][j] - 64) << 1;
+                     s->idsp.put_pixels_clamped(s->block[k], dst[k],
+                                                k & 4 ? s->uvlinesize
+                                                      : s->linesize);
+                 }
+             }
+             if (v->pq >= 9 && v->overlap) {
+                 if (s->mb_x) {
+                     v->vc1dsp.vc1_h_overlap(s->dest[0], s->linesize);
+                     v->vc1dsp.vc1_h_overlap(s->dest[0] + 8 * s->linesize, s->linesize);
+                     if (!(s->flags & CODEC_FLAG_GRAY)) {
+                         v->vc1dsp.vc1_h_overlap(s->dest[1], s->uvlinesize);
+                         v->vc1dsp.vc1_h_overlap(s->dest[2], s->uvlinesize);
+                     }
+                 }
+                 v->vc1dsp.vc1_h_overlap(s->dest[0] + 8, s->linesize);
+                 v->vc1dsp.vc1_h_overlap(s->dest[0] + 8 * s->linesize + 8, s->linesize);
+                 if (!s->first_slice_line) {
+                     v->vc1dsp.vc1_v_overlap(s->dest[0], s->linesize);
+                     v->vc1dsp.vc1_v_overlap(s->dest[0] + 8, s->linesize);
+                     if (!(s->flags & CODEC_FLAG_GRAY)) {
+                         v->vc1dsp.vc1_v_overlap(s->dest[1], s->uvlinesize);
+                         v->vc1dsp.vc1_v_overlap(s->dest[2], s->uvlinesize);
+                     }
+                 }
+                 v->vc1dsp.vc1_v_overlap(s->dest[0] + 8 * s->linesize, s->linesize);
+                 v->vc1dsp.vc1_v_overlap(s->dest[0] + 8 * s->linesize + 8, s->linesize);
+             }
+             if (v->s.loop_filter)
+                 ff_vc1_loop_filter_iblk(v, v->pq);
+             if (get_bits_count(&s->gb) > v->bits) {
+                 ff_er_add_slice(&s->er, 0, 0, s->mb_x, s->mb_y, ER_MB_ERROR);
+                 av_log(s->avctx, AV_LOG_ERROR, "Bits overconsumption: %i > %i\n",
+                        get_bits_count(&s->gb), v->bits);
+                 return;
+             }
+         }
+         if (!v->s.loop_filter)
+             ff_mpeg_draw_horiz_band(s, s->mb_y * 16, 16);
+         else if (s->mb_y)
+             ff_mpeg_draw_horiz_band(s, (s->mb_y - 1) * 16, 16);
+         s->first_slice_line = 0;
+     }
+     if (v->s.loop_filter)
+         ff_mpeg_draw_horiz_band(s, (s->end_mb_y - 1) * 16, 16);
+     /* This is intentionally mb_height and not end_mb_y - unlike in advanced
+      * profile, these only differ are when decoding MSS2 rectangles. */
+     ff_er_add_slice(&s->er, 0, 0, s->mb_width - 1, s->mb_height - 1, ER_MB_END);
+ }
+ /** Decode blocks of I-frame for advanced profile
+  */
+ static void vc1_decode_i_blocks_adv(VC1Context *v)
+ {
+     int k;
+     MpegEncContext *s = &v->s;
+     int cbp, val;
+     uint8_t *coded_val;
+     int mb_pos;
+     int mquant = v->pq;
+     int mqdiff;
+     GetBitContext *gb = &s->gb;
+     /* select codingmode used for VLC tables selection */
+     switch (v->y_ac_table_index) {
+     case 0:
+         v->codingset = (v->pqindex <= 8) ? CS_HIGH_RATE_INTRA : CS_LOW_MOT_INTRA;
+         break;
+     case 1:
+         v->codingset = CS_HIGH_MOT_INTRA;
+         break;
+     case 2:
+         v->codingset = CS_MID_RATE_INTRA;
+         break;
+     }
+     switch (v->c_ac_table_index) {
+     case 0:
+         v->codingset2 = (v->pqindex <= 8) ? CS_HIGH_RATE_INTER : CS_LOW_MOT_INTER;
+         break;
+     case 1:
+         v->codingset2 = CS_HIGH_MOT_INTER;
+         break;
+     case 2:
+         v->codingset2 = CS_MID_RATE_INTER;
+         break;
+     }
+     // do frame decode
+     s->mb_x             = s->mb_y = 0;
+     s->mb_intra         = 1;
+     s->first_slice_line = 1;
+     s->mb_y             = s->start_mb_y;
+     if (s->start_mb_y) {
+         s->mb_x = 0;
+         init_block_index(v);
+         memset(&s->coded_block[s->block_index[0] - s->b8_stride], 0,
+                (1 + s->b8_stride) * sizeof(*s->coded_block));
+     }
+     for (; s->mb_y < s->end_mb_y; s->mb_y++) {
+         s->mb_x = 0;
+         init_block_index(v);
+         for (;s->mb_x < s->mb_width; s->mb_x++) {
+             int16_t (*block)[64] = v->block[v->cur_blk_idx];
+             ff_update_block_index(s);
+             s->bdsp.clear_blocks(block[0]);
+             mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+             s->current_picture.mb_type[mb_pos + v->mb_off]                         = MB_TYPE_INTRA;
+             s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][0] = 0;
+             s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][1] = 0;
+             // do actual MB decoding and displaying
+             if (v->fieldtx_is_raw)
+                 v->fieldtx_plane[mb_pos] = get_bits1(&v->s.gb);
+             cbp = get_vlc2(&v->s.gb, ff_msmp4_mb_i_vlc.table, MB_INTRA_VLC_BITS, 2);
+             if ( v->acpred_is_raw)
+                 v->s.ac_pred = get_bits1(&v->s.gb);
+             else
+                 v->s.ac_pred = v->acpred_plane[mb_pos];
+             if (v->condover == CONDOVER_SELECT && v->overflg_is_raw)
+                 v->over_flags_plane[mb_pos] = get_bits1(&v->s.gb);
+             GET_MQUANT();
+             s->current_picture.qscale_table[mb_pos] = mquant;
+             /* Set DC scale - y and c use the same */
+             s->y_dc_scale = s->y_dc_scale_table[mquant];
+             s->c_dc_scale = s->c_dc_scale_table[mquant];
+             for (k = 0; k < 6; k++) {
+                 val = ((cbp >> (5 - k)) & 1);
+                 if (k < 4) {
+                     int pred   = vc1_coded_block_pred(&v->s, k, &coded_val);
+                     val        = val ^ pred;
+                     *coded_val = val;
+                 }
+                 cbp |= val << (5 - k);
+                 v->a_avail = !s->first_slice_line || (k == 2 || k == 3);
+                 v->c_avail = !!s->mb_x || (k == 1 || k == 3);
+                 vc1_decode_i_block_adv(v, block[k], k, val,
+                                        (k < 4) ? v->codingset : v->codingset2, mquant);
+                 if (k > 3 && (s->flags & CODEC_FLAG_GRAY))
+                     continue;
+                 v->vc1dsp.vc1_inv_trans_8x8(block[k]);
+             }
+             ff_vc1_smooth_overlap_filter_iblk(v);
+             vc1_put_signed_blocks_clamped(v);
+             if (v->s.loop_filter)
+                 ff_vc1_loop_filter_iblk_delayed(v, v->pq);
+             if (get_bits_count(&s->gb) > v->bits) {
+                 // TODO: may need modification to handle slice coding
+                 ff_er_add_slice(&s->er, 0, s->start_mb_y, s->mb_x, s->mb_y, ER_MB_ERROR);
+                 av_log(s->avctx, AV_LOG_ERROR, "Bits overconsumption: %i > %i\n",
+                        get_bits_count(&s->gb), v->bits);
+                 return;
+             }
+         }
+         if (!v->s.loop_filter)
+             ff_mpeg_draw_horiz_band(s, s->mb_y * 16, 16);
+         else if (s->mb_y)
+             ff_mpeg_draw_horiz_band(s, (s->mb_y-1) * 16, 16);
+         s->first_slice_line = 0;
+     }
+     /* raw bottom MB row */
+     s->mb_x = 0;
+     init_block_index(v);
+     for (;s->mb_x < s->mb_width; s->mb_x++) {
+         ff_update_block_index(s);
+         vc1_put_signed_blocks_clamped(v);
+         if (v->s.loop_filter)
+             ff_vc1_loop_filter_iblk_delayed(v, v->pq);
+     }
+     if (v->s.loop_filter)
+         ff_mpeg_draw_horiz_band(s, (s->end_mb_y-1)*16, 16);
+     ff_er_add_slice(&s->er, 0, s->start_mb_y << v->field_mode, s->mb_width - 1,
+                     (s->end_mb_y << v->field_mode) - 1, ER_MB_END);
+ }
+ static void vc1_decode_p_blocks(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     int apply_loop_filter;
+     /* select codingmode used for VLC tables selection */
+     switch (v->c_ac_table_index) {
+     case 0:
+         v->codingset = (v->pqindex <= 8) ? CS_HIGH_RATE_INTRA : CS_LOW_MOT_INTRA;
+         break;
+     case 1:
+         v->codingset = CS_HIGH_MOT_INTRA;
+         break;
+     case 2:
+         v->codingset = CS_MID_RATE_INTRA;
+         break;
+     }
+     switch (v->c_ac_table_index) {
+     case 0:
+         v->codingset2 = (v->pqindex <= 8) ? CS_HIGH_RATE_INTER : CS_LOW_MOT_INTER;
+         break;
+     case 1:
+         v->codingset2 = CS_HIGH_MOT_INTER;
+         break;
+     case 2:
+         v->codingset2 = CS_MID_RATE_INTER;
+         break;
+     }
+     apply_loop_filter   = s->loop_filter && !(s->avctx->skip_loop_filter >= AVDISCARD_NONKEY) &&
+                           v->fcm == PROGRESSIVE;
+     s->first_slice_line = 1;
+     memset(v->cbp_base, 0, sizeof(v->cbp_base[0])*2*s->mb_stride);
+     for (s->mb_y = s->start_mb_y; s->mb_y < s->end_mb_y; s->mb_y++) {
+         s->mb_x = 0;
+         init_block_index(v);
+         for (; s->mb_x < s->mb_width; s->mb_x++) {
+             ff_update_block_index(s);
+             if (v->fcm == ILACE_FIELD)
+                 vc1_decode_p_mb_intfi(v);
+             else if (v->fcm == ILACE_FRAME)
+                 vc1_decode_p_mb_intfr(v);
+             else vc1_decode_p_mb(v);
+             if (s->mb_y != s->start_mb_y && apply_loop_filter)
+                 ff_vc1_apply_p_loop_filter(v);
+             if (get_bits_count(&s->gb) > v->bits || get_bits_count(&s->gb) < 0) {
+                 // TODO: may need modification to handle slice coding
+                 ff_er_add_slice(&s->er, 0, s->start_mb_y, s->mb_x, s->mb_y, ER_MB_ERROR);
+                 av_log(s->avctx, AV_LOG_ERROR, "Bits overconsumption: %i > %i at %ix%i\n",
+                        get_bits_count(&s->gb), v->bits, s->mb_x, s->mb_y);
+                 return;
+             }
+         }
+         memmove(v->cbp_base,      v->cbp,      sizeof(v->cbp_base[0])      * s->mb_stride);
+         memmove(v->ttblk_base,    v->ttblk,    sizeof(v->ttblk_base[0])    * s->mb_stride);
+         memmove(v->is_intra_base, v->is_intra, sizeof(v->is_intra_base[0]) * s->mb_stride);
+         memmove(v->luma_mv_base,  v->luma_mv,  sizeof(v->luma_mv_base[0])  * s->mb_stride);
+         if (s->mb_y != s->start_mb_y) ff_mpeg_draw_horiz_band(s, (s->mb_y - 1) * 16, 16);
+         s->first_slice_line = 0;
+     }
+     if (apply_loop_filter) {
+         s->mb_x = 0;
+         init_block_index(v);
+         for (; s->mb_x < s->mb_width; s->mb_x++) {
+             ff_update_block_index(s);
+             ff_vc1_apply_p_loop_filter(v);
+         }
+     }
+     if (s->end_mb_y >= s->start_mb_y)
+         ff_mpeg_draw_horiz_band(s, (s->end_mb_y - 1) * 16, 16);
+     ff_er_add_slice(&s->er, 0, s->start_mb_y << v->field_mode, s->mb_width - 1,
+                     (s->end_mb_y << v->field_mode) - 1, ER_MB_END);
+ }
+ static void vc1_decode_b_blocks(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     /* select codingmode used for VLC tables selection */
+     switch (v->c_ac_table_index) {
+     case 0:
+         v->codingset = (v->pqindex <= 8) ? CS_HIGH_RATE_INTRA : CS_LOW_MOT_INTRA;
+         break;
+     case 1:
+         v->codingset = CS_HIGH_MOT_INTRA;
+         break;
+     case 2:
+         v->codingset = CS_MID_RATE_INTRA;
+         break;
+     }
+     switch (v->c_ac_table_index) {
+     case 0:
+         v->codingset2 = (v->pqindex <= 8) ? CS_HIGH_RATE_INTER : CS_LOW_MOT_INTER;
+         break;
+     case 1:
+         v->codingset2 = CS_HIGH_MOT_INTER;
+         break;
+     case 2:
+         v->codingset2 = CS_MID_RATE_INTER;
+         break;
+     }
+     s->first_slice_line = 1;
+     for (s->mb_y = s->start_mb_y; s->mb_y < s->end_mb_y; s->mb_y++) {
+         s->mb_x = 0;
+         init_block_index(v);
+         for (; s->mb_x < s->mb_width; s->mb_x++) {
+             ff_update_block_index(s);
+             if (v->fcm == ILACE_FIELD)
+                 vc1_decode_b_mb_intfi(v);
+             else if (v->fcm == ILACE_FRAME)
+                 vc1_decode_b_mb_intfr(v);
+             else
+                 vc1_decode_b_mb(v);
+             if (get_bits_count(&s->gb) > v->bits || get_bits_count(&s->gb) < 0) {
+                 // TODO: may need modification to handle slice coding
+                 ff_er_add_slice(&s->er, 0, s->start_mb_y, s->mb_x, s->mb_y, ER_MB_ERROR);
+                 av_log(s->avctx, AV_LOG_ERROR, "Bits overconsumption: %i > %i at %ix%i\n",
+                        get_bits_count(&s->gb), v->bits, s->mb_x, s->mb_y);
+                 return;
+             }
+             if (v->s.loop_filter)
+                 ff_vc1_loop_filter_iblk(v, v->pq);
+         }
+         if (!v->s.loop_filter)
+             ff_mpeg_draw_horiz_band(s, s->mb_y * 16, 16);
+         else if (s->mb_y)
+             ff_mpeg_draw_horiz_band(s, (s->mb_y - 1) * 16, 16);
+         s->first_slice_line = 0;
+     }
+     if (v->s.loop_filter)
+         ff_mpeg_draw_horiz_band(s, (s->end_mb_y - 1) * 16, 16);
+     ff_er_add_slice(&s->er, 0, s->start_mb_y << v->field_mode, s->mb_width - 1,
+                     (s->end_mb_y << v->field_mode) - 1, ER_MB_END);
+ }
+ static void vc1_decode_skip_blocks(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     if (!v->s.last_picture.f->data[0])
+         return;
+     ff_er_add_slice(&s->er, 0, s->start_mb_y, s->mb_width - 1, s->end_mb_y - 1, ER_MB_END);
+     s->first_slice_line = 1;
+     for (s->mb_y = s->start_mb_y; s->mb_y < s->end_mb_y; s->mb_y++) {
+         s->mb_x = 0;
+         init_block_index(v);
+         ff_update_block_index(s);
+         memcpy(s->dest[0], s->last_picture.f->data[0] + s->mb_y * 16 * s->linesize,   s->linesize   * 16);
+         memcpy(s->dest[1], s->last_picture.f->data[1] + s->mb_y *  8 * s->uvlinesize, s->uvlinesize *  8);
+         memcpy(s->dest[2], s->last_picture.f->data[2] + s->mb_y *  8 * s->uvlinesize, s->uvlinesize *  8);
+         ff_mpeg_draw_horiz_band(s, s->mb_y * 16, 16);
+         s->first_slice_line = 0;
+     }
+     s->pict_type = AV_PICTURE_TYPE_P;
+ }
+ void ff_vc1_decode_blocks(VC1Context *v)
+ {
+     v->s.esc3_level_length = 0;
+     if (v->x8_type) {
+         ff_intrax8_decode_picture(&v->x8, 2*v->pq + v->halfpq, v->pq * !v->pquantizer);
+     } else {
+         v->cur_blk_idx     =  0;
+         v->left_blk_idx    = -1;
+         v->topleft_blk_idx =  1;
+         v->top_blk_idx     =  2;
+         switch (v->s.pict_type) {
+         case AV_PICTURE_TYPE_I:
+             if (v->profile == PROFILE_ADVANCED)
+                 vc1_decode_i_blocks_adv(v);
+             else
+                 vc1_decode_i_blocks(v);
+             break;
+         case AV_PICTURE_TYPE_P:
+             if (v->p_frame_skipped)
+                 vc1_decode_skip_blocks(v);
+             else
+                 vc1_decode_p_blocks(v);
+             break;
+         case AV_PICTURE_TYPE_B:
+             if (v->bi_type) {
+                 if (v->profile == PROFILE_ADVANCED)
+                     vc1_decode_i_blocks_adv(v);
+                 else
+                     vc1_decode_i_blocks(v);
+             } else
+                 vc1_decode_b_blocks(v);
+             break;
+         }
+     }
+ }
index 0000000,1b76d83..95c4519
mode 000000,100644..100644
--- /dev/null
@@@ -1,0 -1,353 +1,353 @@@
 - * This file is part of Libav.
+ /*
+  * VC-1 and WMV3 decoder
+  * Copyright (c) 2011 Mashiat Sarker Shakkhar
+  * Copyright (c) 2006-2007 Konstantin Shishkov
+  * Partly based on vc9.c (c) 2005 Anonymous, Alex Beregszaszi, Michael Niedermayer
+  *
 - * Libav is free software; you can redistribute it and/or
++ * This file is part of FFmpeg.
+  *
 - * Libav is distributed in the hope that it will be useful,
++ * FFmpeg is free software; you can redistribute it and/or
+  * modify it under the terms of the GNU Lesser General Public
+  * License as published by the Free Software Foundation; either
+  * version 2.1 of the License, or (at your option) any later version.
+  *
 - * License along with Libav; if not, write to the Free Software
++ * FFmpeg is distributed in the hope that it will be useful,
+  * but WITHOUT ANY WARRANTY; without even the implied warranty of
+  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+  * Lesser General Public License for more details.
+  *
+  * You should have received a copy of the GNU Lesser General Public
 -        block_is_intra = mb_is_intra >> (block_num * 4), bottom_is_intra;
++ * License along with FFmpeg; if not, write to the Free Software
+  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+  */
+ /**
+  * @file
+  * VC-1 and WMV3 loopfilter
+  */
+ #include "avcodec.h"
+ #include "mpegvideo.h"
+ #include "vc1.h"
+ #include "vc1dsp.h"
+ void ff_vc1_loop_filter_iblk(VC1Context *v, int pq)
+ {
+     MpegEncContext *s = &v->s;
+     int j;
+     if (!s->first_slice_line) {
+         v->vc1dsp.vc1_v_loop_filter16(s->dest[0], s->linesize, pq);
+         if (s->mb_x)
+             v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize, s->linesize, pq);
+         v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize + 8, s->linesize, pq);
+         for (j = 0; j < 2; j++) {
+             v->vc1dsp.vc1_v_loop_filter8(s->dest[j + 1], s->uvlinesize, pq);
+             if (s->mb_x)
+                 v->vc1dsp.vc1_h_loop_filter8(s->dest[j + 1] - 8 * s->uvlinesize, s->uvlinesize, pq);
+         }
+     }
+     v->vc1dsp.vc1_v_loop_filter16(s->dest[0] + 8 * s->linesize, s->linesize, pq);
+     if (s->mb_y == s->end_mb_y - 1) {
+         if (s->mb_x) {
+             v->vc1dsp.vc1_h_loop_filter16(s->dest[0], s->linesize, pq);
+             v->vc1dsp.vc1_h_loop_filter8(s->dest[1], s->uvlinesize, pq);
+             v->vc1dsp.vc1_h_loop_filter8(s->dest[2], s->uvlinesize, pq);
+         }
+         v->vc1dsp.vc1_h_loop_filter16(s->dest[0] + 8, s->linesize, pq);
+     }
+ }
+ void ff_vc1_loop_filter_iblk_delayed(VC1Context *v, int pq)
+ {
+     MpegEncContext *s = &v->s;
+     int j;
+     /* The loopfilter runs 1 row and 1 column behind the overlap filter, which
+      * means it runs two rows/cols behind the decoding loop. */
+     if (!s->first_slice_line) {
+         if (s->mb_x) {
+             if (s->mb_y >= s->start_mb_y + 2) {
+                 v->vc1dsp.vc1_v_loop_filter16(s->dest[0] - 16 * s->linesize - 16, s->linesize, pq);
+                 if (s->mb_x >= 2)
+                     v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 32 * s->linesize - 16, s->linesize, pq);
+                 v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 32 * s->linesize - 8, s->linesize, pq);
+                 for (j = 0; j < 2; j++) {
+                     v->vc1dsp.vc1_v_loop_filter8(s->dest[j + 1] - 8 * s->uvlinesize - 8, s->uvlinesize, pq);
+                     if (s->mb_x >= 2) {
+                         v->vc1dsp.vc1_h_loop_filter8(s->dest[j + 1] - 16 * s->uvlinesize - 8, s->uvlinesize, pq);
+                     }
+                 }
+             }
+             v->vc1dsp.vc1_v_loop_filter16(s->dest[0] - 8 * s->linesize - 16, s->linesize, pq);
+         }
+         if (s->mb_x == s->mb_width - 1) {
+             if (s->mb_y >= s->start_mb_y + 2) {
+                 v->vc1dsp.vc1_v_loop_filter16(s->dest[0] - 16 * s->linesize, s->linesize, pq);
+                 if (s->mb_x)
+                     v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 32 * s->linesize, s->linesize, pq);
+                 v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 32 * s->linesize + 8, s->linesize, pq);
+                 for (j = 0; j < 2; j++) {
+                     v->vc1dsp.vc1_v_loop_filter8(s->dest[j + 1] - 8 * s->uvlinesize, s->uvlinesize, pq);
+                     if (s->mb_x >= 2) {
+                         v->vc1dsp.vc1_h_loop_filter8(s->dest[j + 1] - 16 * s->uvlinesize, s->uvlinesize, pq);
+                     }
+                 }
+             }
+             v->vc1dsp.vc1_v_loop_filter16(s->dest[0] - 8 * s->linesize, s->linesize, pq);
+         }
+         if (s->mb_y == s->end_mb_y) {
+             if (s->mb_x) {
+                 if (s->mb_x >= 2)
+                     v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize - 16, s->linesize, pq);
+                 v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize - 8, s->linesize, pq);
+                 if (s->mb_x >= 2) {
+                     for (j = 0; j < 2; j++) {
+                         v->vc1dsp.vc1_h_loop_filter8(s->dest[j + 1] - 8 * s->uvlinesize - 8, s->uvlinesize, pq);
+                     }
+                 }
+             }
+             if (s->mb_x == s->mb_width - 1) {
+                 if (s->mb_x)
+                     v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize, s->linesize, pq);
+                 v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize + 8, s->linesize, pq);
+                 if (s->mb_x) {
+                     for (j = 0; j < 2; j++) {
+                         v->vc1dsp.vc1_h_loop_filter8(s->dest[j + 1] - 8 * s->uvlinesize, s->uvlinesize, pq);
+                     }
+                 }
+             }
+         }
+     }
+ }
+ void ff_vc1_smooth_overlap_filter_iblk(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     int mb_pos;
+     if (v->condover == CONDOVER_NONE)
+         return;
+     mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     /* Within a MB, the horizontal overlap always runs before the vertical.
+      * To accomplish that, we run the H on left and internal borders of the
+      * currently decoded MB. Then, we wait for the next overlap iteration
+      * to do H overlap on the right edge of this MB, before moving over and
+      * running the V overlap. Therefore, the V overlap makes us trail by one
+      * MB col and the H overlap filter makes us trail by one MB row. This
+      * is reflected in the time at which we run the put_pixels loop. */
+     if (v->condover == CONDOVER_ALL || v->pq >= 9 || v->over_flags_plane[mb_pos]) {
+         if (s->mb_x && (v->condover == CONDOVER_ALL || v->pq >= 9 ||
+                         v->over_flags_plane[mb_pos - 1])) {
+             v->vc1dsp.vc1_h_s_overlap(v->block[v->left_blk_idx][1],
+                                       v->block[v->cur_blk_idx][0]);
+             v->vc1dsp.vc1_h_s_overlap(v->block[v->left_blk_idx][3],
+                                       v->block[v->cur_blk_idx][2]);
+             if (!(s->flags & CODEC_FLAG_GRAY)) {
+                 v->vc1dsp.vc1_h_s_overlap(v->block[v->left_blk_idx][4],
+                                           v->block[v->cur_blk_idx][4]);
+                 v->vc1dsp.vc1_h_s_overlap(v->block[v->left_blk_idx][5],
+                                           v->block[v->cur_blk_idx][5]);
+             }
+         }
+         v->vc1dsp.vc1_h_s_overlap(v->block[v->cur_blk_idx][0],
+                                   v->block[v->cur_blk_idx][1]);
+         v->vc1dsp.vc1_h_s_overlap(v->block[v->cur_blk_idx][2],
+                                   v->block[v->cur_blk_idx][3]);
+         if (s->mb_x == s->mb_width - 1) {
+             if (!s->first_slice_line && (v->condover == CONDOVER_ALL || v->pq >= 9 ||
+                                          v->over_flags_plane[mb_pos - s->mb_stride])) {
+                 v->vc1dsp.vc1_v_s_overlap(v->block[v->top_blk_idx][2],
+                                           v->block[v->cur_blk_idx][0]);
+                 v->vc1dsp.vc1_v_s_overlap(v->block[v->top_blk_idx][3],
+                                           v->block[v->cur_blk_idx][1]);
+                 if (!(s->flags & CODEC_FLAG_GRAY)) {
+                     v->vc1dsp.vc1_v_s_overlap(v->block[v->top_blk_idx][4],
+                                               v->block[v->cur_blk_idx][4]);
+                     v->vc1dsp.vc1_v_s_overlap(v->block[v->top_blk_idx][5],
+                                               v->block[v->cur_blk_idx][5]);
+                 }
+             }
+             v->vc1dsp.vc1_v_s_overlap(v->block[v->cur_blk_idx][0],
+                                       v->block[v->cur_blk_idx][2]);
+             v->vc1dsp.vc1_v_s_overlap(v->block[v->cur_blk_idx][1],
+                                       v->block[v->cur_blk_idx][3]);
+         }
+     }
+     if (s->mb_x && (v->condover == CONDOVER_ALL || v->over_flags_plane[mb_pos - 1])) {
+         if (!s->first_slice_line && (v->condover == CONDOVER_ALL || v->pq >= 9 ||
+                                      v->over_flags_plane[mb_pos - s->mb_stride - 1])) {
+             v->vc1dsp.vc1_v_s_overlap(v->block[v->topleft_blk_idx][2],
+                                       v->block[v->left_blk_idx][0]);
+             v->vc1dsp.vc1_v_s_overlap(v->block[v->topleft_blk_idx][3],
+                                       v->block[v->left_blk_idx][1]);
+             if (!(s->flags & CODEC_FLAG_GRAY)) {
+                 v->vc1dsp.vc1_v_s_overlap(v->block[v->topleft_blk_idx][4],
+                                           v->block[v->left_blk_idx][4]);
+                 v->vc1dsp.vc1_v_s_overlap(v->block[v->topleft_blk_idx][5],
+                                           v->block[v->left_blk_idx][5]);
+             }
+         }
+         v->vc1dsp.vc1_v_s_overlap(v->block[v->left_blk_idx][0],
+                                   v->block[v->left_blk_idx][2]);
+         v->vc1dsp.vc1_v_s_overlap(v->block[v->left_blk_idx][1],
+                                   v->block[v->left_blk_idx][3]);
+     }
+ }
+ static av_always_inline void vc1_apply_p_v_loop_filter(VC1Context *v, int block_num)
+ {
+     MpegEncContext *s  = &v->s;
+     int mb_cbp         = v->cbp[s->mb_x - s->mb_stride],
+         block_cbp      = mb_cbp      >> (block_num * 4), bottom_cbp,
+         mb_is_intra    = v->is_intra[s->mb_x - s->mb_stride],
 -            bottom_is_intra = v->is_intra[s->mb_x] >> (block_num * 4);
++        block_is_intra = mb_is_intra >> block_num, bottom_is_intra;
+     int idx, linesize  = block_num > 3 ? s->uvlinesize : s->linesize, ttblk;
+     uint8_t *dst;
+     if (block_num > 3) {
+         dst      = s->dest[block_num - 3];
+     } else {
+         dst      = s->dest[0] + (block_num & 1) * 8 + ((block_num & 2) * 4 - 8) * linesize;
+     }
+     if (s->mb_y != s->end_mb_y || block_num < 2) {
+         int16_t (*mv)[2];
+         int mv_stride;
+         if (block_num > 3) {
+             bottom_cbp      = v->cbp[s->mb_x]      >> (block_num * 4);
 -            bottom_is_intra = (block_num < 2) ? (mb_is_intra          >> ((block_num + 2) * 4))
 -                                              : (v->is_intra[s->mb_x] >> ((block_num - 2) * 4));
++            bottom_is_intra = v->is_intra[s->mb_x] >> block_num;
+             mv              = &v->luma_mv[s->mb_x - s->mb_stride];
+             mv_stride       = s->mb_stride;
+         } else {
+             bottom_cbp      = (block_num < 2) ? (mb_cbp               >> ((block_num + 2) * 4))
+                                               : (v->cbp[s->mb_x]      >> ((block_num - 2) * 4));
 -        block_is_intra = mb_is_intra >> (block_num * 4), right_is_intra;
++            bottom_is_intra = (block_num < 2) ? (mb_is_intra          >> (block_num + 2))
++                                              : (v->is_intra[s->mb_x] >> (block_num - 2));
+             mv_stride       = s->b8_stride;
+             mv              = &s->current_picture.motion_val[0][s->block_index[block_num] - 2 * mv_stride];
+         }
+         if (bottom_is_intra & 1 || block_is_intra & 1 ||
+             mv[0][0] != mv[mv_stride][0] || mv[0][1] != mv[mv_stride][1]) {
+             v->vc1dsp.vc1_v_loop_filter8(dst, linesize, v->pq);
+         } else {
+             idx = ((bottom_cbp >> 2) | block_cbp) & 3;
+             if (idx == 3) {
+                 v->vc1dsp.vc1_v_loop_filter8(dst, linesize, v->pq);
+             } else if (idx) {
+                 if (idx == 1)
+                     v->vc1dsp.vc1_v_loop_filter4(dst + 4, linesize, v->pq);
+                 else
+                     v->vc1dsp.vc1_v_loop_filter4(dst,     linesize, v->pq);
+             }
+         }
+     }
+     dst -= 4 * linesize;
+     ttblk = (v->ttblk[s->mb_x - s->mb_stride] >> (block_num * 4)) & 0xF;
+     if (ttblk == TT_4X4 || ttblk == TT_8X4) {
+         idx = (block_cbp | (block_cbp >> 2)) & 3;
+         if (idx == 3) {
+             v->vc1dsp.vc1_v_loop_filter8(dst, linesize, v->pq);
+         } else if (idx) {
+             if (idx == 1)
+                 v->vc1dsp.vc1_v_loop_filter4(dst + 4, linesize, v->pq);
+             else
+                 v->vc1dsp.vc1_v_loop_filter4(dst,     linesize, v->pq);
+         }
+     }
+ }
+ static av_always_inline void vc1_apply_p_h_loop_filter(VC1Context *v, int block_num)
+ {
+     MpegEncContext *s  = &v->s;
+     int mb_cbp         = v->cbp[s->mb_x - 1 - s->mb_stride],
+         block_cbp      = mb_cbp      >> (block_num * 4), right_cbp,
+         mb_is_intra    = v->is_intra[s->mb_x - 1 - s->mb_stride],
 -            right_is_intra = v->is_intra[s->mb_x - s->mb_stride] >> (block_num * 4);
++        block_is_intra = mb_is_intra >> block_num, right_is_intra;
+     int idx, linesize  = block_num > 3 ? s->uvlinesize : s->linesize, ttblk;
+     uint8_t *dst;
+     if (block_num > 3) {
+         dst = s->dest[block_num - 3] - 8 * linesize;
+     } else {
+         dst = s->dest[0] + (block_num & 1) * 8 + ((block_num & 2) * 4 - 16) * linesize - 8;
+     }
+     if (s->mb_x != s->mb_width || !(block_num & 5)) {
+         int16_t (*mv)[2];
+         if (block_num > 3) {
+             right_cbp      = v->cbp[s->mb_x - s->mb_stride] >> (block_num * 4);
 -            right_is_intra = (block_num & 1) ? (v->is_intra[s->mb_x - s->mb_stride] >> ((block_num - 1) * 4))
 -                                             : (mb_is_intra                         >> ((block_num + 1) * 4));
++            right_is_intra = v->is_intra[s->mb_x - s->mb_stride] >> block_num;
+             mv             = &v->luma_mv[s->mb_x - s->mb_stride - 1];
+         } else {
+             right_cbp      = (block_num & 1) ? (v->cbp[s->mb_x - s->mb_stride]      >> ((block_num - 1) * 4))
+                                              : (mb_cbp                              >> ((block_num + 1) * 4));
++            right_is_intra = (block_num & 1) ? (v->is_intra[s->mb_x - s->mb_stride] >> (block_num - 1))
++                                             : (mb_is_intra                         >> (block_num + 1));
+             mv             = &s->current_picture.motion_val[0][s->block_index[block_num] - s->b8_stride * 2 - 2];
+         }
+         if (block_is_intra & 1 || right_is_intra & 1 || mv[0][0] != mv[1][0] || mv[0][1] != mv[1][1]) {
+             v->vc1dsp.vc1_h_loop_filter8(dst, linesize, v->pq);
+         } else {
+             idx = ((right_cbp >> 1) | block_cbp) & 5; // FIXME check
+             if (idx == 5) {
+                 v->vc1dsp.vc1_h_loop_filter8(dst, linesize, v->pq);
+             } else if (idx) {
+                 if (idx == 1)
+                     v->vc1dsp.vc1_h_loop_filter4(dst + 4 * linesize, linesize, v->pq);
+                 else
+                     v->vc1dsp.vc1_h_loop_filter4(dst,                linesize, v->pq);
+             }
+         }
+     }
+     dst -= 4;
+     ttblk = (v->ttblk[s->mb_x - s->mb_stride - 1] >> (block_num * 4)) & 0xf;
+     if (ttblk == TT_4X4 || ttblk == TT_4X8) {
+         idx = (block_cbp | (block_cbp >> 1)) & 5;
+         if (idx == 5) {
+             v->vc1dsp.vc1_h_loop_filter8(dst, linesize, v->pq);
+         } else if (idx) {
+             if (idx == 1)
+                 v->vc1dsp.vc1_h_loop_filter4(dst + linesize * 4, linesize, v->pq);
+             else
+                 v->vc1dsp.vc1_h_loop_filter4(dst,                linesize, v->pq);
+         }
+     }
+ }
+ void ff_vc1_apply_p_loop_filter(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     int i;
+     for (i = 0; i < 6; i++) {
+         vc1_apply_p_v_loop_filter(v, i);
+     }
+     /* V always precedes H, therefore we run H one MB before V;
+      * at the end of a row, we catch up to complete the row */
+     if (s->mb_x) {
+         for (i = 0; i < 6; i++) {
+             vc1_apply_p_h_loop_filter(v, i);
+         }
+         if (s->mb_x == s->mb_width - 1) {
+             s->mb_x++;
+             ff_update_block_index(s);
+             for (i = 0; i < 6; i++) {
+                 vc1_apply_p_h_loop_filter(v, i);
+             }
+         }
+     }
+ }
index 0000000,54c4e72..3379932
mode 000000,100644..100644
--- /dev/null
@@@ -1,0 -1,948 +1,954 @@@
 - * This file is part of Libav.
+ /*
+  * VC-1 and WMV3 decoder
+  * Copyright (c) 2011 Mashiat Sarker Shakkhar
+  * Copyright (c) 2006-2007 Konstantin Shishkov
+  * Partly based on vc9.c (c) 2005 Anonymous, Alex Beregszaszi, Michael Niedermayer
+  *
 - * Libav is free software; you can redistribute it and/or
++ * This file is part of FFmpeg.
+  *
 - * Libav is distributed in the hope that it will be useful,
++ * FFmpeg is free software; you can redistribute it and/or
+  * modify it under the terms of the GNU Lesser General Public
+  * License as published by the Free Software Foundation; either
+  * version 2.1 of the License, or (at your option) any later version.
+  *
 - * License along with Libav; if not, write to the Free Software
++ * FFmpeg is distributed in the hope that it will be useful,
+  * but WITHOUT ANY WARRANTY; without even the implied warranty of
+  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+  * Lesser General Public License for more details.
+  *
+  * You should have received a copy of the GNU Lesser General Public
 -            use_ic = v->curr_use_ic;
++ * License along with FFmpeg; if not, write to the Free Software
+  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+  */
+ /**
+  * @file
+  * VC-1 and WMV3 block decoding routines
+  */
+ #include "avcodec.h"
+ #include "h264chroma.h"
+ #include "mathops.h"
+ #include "mpegvideo.h"
+ #include "vc1.h"
+ /** Do motion compensation over 1 macroblock
+  * Mostly adapted hpel_motion and qpel_motion from mpegvideo.c
+  */
+ void ff_vc1_mc_1mv(VC1Context *v, int dir)
+ {
+     MpegEncContext *s = &v->s;
+     H264ChromaContext *h264chroma = &v->h264chroma;
+     uint8_t *srcY, *srcU, *srcV;
+     int dxy, mx, my, uvmx, uvmy, src_x, src_y, uvsrc_x, uvsrc_y;
+     int v_edge_pos = s->v_edge_pos >> v->field_mode;
+     int i;
+     uint8_t (*luty)[256], (*lutuv)[256];
+     int use_ic;
+     if ((!v->field_mode ||
+          (v->ref_field_type[dir] == 1 && v->cur_field_type == 1)) &&
+         !v->s.last_picture.f->data[0])
+         return;
+     mx = s->mv[dir][0][0];
+     my = s->mv[dir][0][1];
+     // store motion vectors for further use in B frames
+     if (s->pict_type == AV_PICTURE_TYPE_P) {
+         for (i = 0; i < 4; i++) {
+             s->current_picture.motion_val[1][s->block_index[i] + v->blocks_off][0] = mx;
+             s->current_picture.motion_val[1][s->block_index[i] + v->blocks_off][1] = my;
+         }
+     }
+     uvmx = (mx + ((mx & 3) == 3)) >> 1;
+     uvmy = (my + ((my & 3) == 3)) >> 1;
+     v->luma_mv[s->mb_x][0] = uvmx;
+     v->luma_mv[s->mb_x][1] = uvmy;
+     if (v->field_mode &&
+         v->cur_field_type != v->ref_field_type[dir]) {
+         my   = my   - 2 + 4 * v->cur_field_type;
+         uvmy = uvmy - 2 + 4 * v->cur_field_type;
+     }
+     // fastuvmc shall be ignored for interlaced frame picture
+     if (v->fastuvmc && (v->fcm != ILACE_FRAME)) {
+         uvmx = uvmx + ((uvmx < 0) ? (uvmx & 1) : -(uvmx & 1));
+         uvmy = uvmy + ((uvmy < 0) ? (uvmy & 1) : -(uvmy & 1));
+     }
+     if (!dir) {
+         if (v->field_mode && (v->cur_field_type != v->ref_field_type[dir]) && v->second_field) {
+             srcY = s->current_picture.f->data[0];
+             srcU = s->current_picture.f->data[1];
+             srcV = s->current_picture.f->data[2];
+             luty  = v->curr_luty;
+             lutuv = v->curr_lutuv;
 -        uint8_t *uvbuf = s->edge_emu_buffer + 19 * s->linesize;
++            use_ic = *v->curr_use_ic;
+         } else {
+             srcY = s->last_picture.f->data[0];
+             srcU = s->last_picture.f->data[1];
+             srcV = s->last_picture.f->data[2];
+             luty  = v->last_luty;
+             lutuv = v->last_lutuv;
+             use_ic = v->last_use_ic;
+         }
+     } else {
+         srcY = s->next_picture.f->data[0];
+         srcU = s->next_picture.f->data[1];
+         srcV = s->next_picture.f->data[2];
+         luty  = v->next_luty;
+         lutuv = v->next_lutuv;
+         use_ic = v->next_use_ic;
+     }
+     if (!srcY || !srcU) {
+         av_log(v->s.avctx, AV_LOG_ERROR, "Referenced frame missing.\n");
+         return;
+     }
+     src_x   = s->mb_x * 16 + (mx   >> 2);
+     src_y   = s->mb_y * 16 + (my   >> 2);
+     uvsrc_x = s->mb_x *  8 + (uvmx >> 2);
+     uvsrc_y = s->mb_y *  8 + (uvmy >> 2);
+     if (v->profile != PROFILE_ADVANCED) {
+         src_x   = av_clip(  src_x, -16, s->mb_width  * 16);
+         src_y   = av_clip(  src_y, -16, s->mb_height * 16);
+         uvsrc_x = av_clip(uvsrc_x,  -8, s->mb_width  *  8);
+         uvsrc_y = av_clip(uvsrc_y,  -8, s->mb_height *  8);
+     } else {
+         src_x   = av_clip(  src_x, -17, s->avctx->coded_width);
+         src_y   = av_clip(  src_y, -18, s->avctx->coded_height + 1);
+         uvsrc_x = av_clip(uvsrc_x,  -8, s->avctx->coded_width  >> 1);
+         uvsrc_y = av_clip(uvsrc_y,  -8, s->avctx->coded_height >> 1);
+     }
+     srcY += src_y   * s->linesize   + src_x;
+     srcU += uvsrc_y * s->uvlinesize + uvsrc_x;
+     srcV += uvsrc_y * s->uvlinesize + uvsrc_x;
+     if (v->field_mode && v->ref_field_type[dir]) {
+         srcY += s->current_picture_ptr->f->linesize[0];
+         srcU += s->current_picture_ptr->f->linesize[1];
+         srcV += s->current_picture_ptr->f->linesize[2];
+     }
+     /* for grayscale we should not try to read from unknown area */
+     if (s->flags & CODEC_FLAG_GRAY) {
+         srcU = s->edge_emu_buffer + 18 * s->linesize;
+         srcV = s->edge_emu_buffer + 18 * s->linesize;
+     }
+     if (v->rangeredfrm || use_ic
+         || s->h_edge_pos < 22 || v_edge_pos < 22
+         || (unsigned)(src_x - s->mspel) > s->h_edge_pos - (mx&3) - 16 - s->mspel * 3
+         || (unsigned)(src_y - 1)        > v_edge_pos    - (my&3) - 16 - 3) {
 -        s->vdsp.emulated_edge_mc(uvbuf, srcU,
++        uint8_t *ubuf = s->edge_emu_buffer + 19 * s->linesize;
++        uint8_t *vbuf = ubuf + 9 * s->uvlinesize;
+         srcY -= s->mspel * (1 + s->linesize);
+         s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
+                                  s->linesize, s->linesize,
+                                  17 + s->mspel * 2, 17 + s->mspel * 2,
+                                  src_x - s->mspel, src_y - s->mspel,
+                                  s->h_edge_pos, v_edge_pos);
+         srcY = s->edge_emu_buffer;
 -                                 uvsrc_x, uvsrc_y, s->h_edge_pos >> 1, v_edge_pos >> 1);
 -        s->vdsp.emulated_edge_mc(uvbuf + 16, srcV,
++        s->vdsp.emulated_edge_mc(ubuf, srcU,
+                                  s->uvlinesize, s->uvlinesize,
+                                  8 + 1, 8 + 1,
 -                                 uvsrc_x, uvsrc_y, s->h_edge_pos >> 1, v_edge_pos >> 1);
 -        srcU = uvbuf;
 -        srcV = uvbuf + 16;
++                                 uvsrc_x, uvsrc_y,
++                                 s->h_edge_pos >> 1, v_edge_pos >> 1);
++        s->vdsp.emulated_edge_mc(vbuf, srcV,
+                                  s->uvlinesize, s->uvlinesize,
+                                  8 + 1, 8 + 1,
 -        v->vc1dsp.put_vc1_mspel_pixels_tab[dxy](s->dest[0]    , srcY    , s->linesize, v->rnd);
 -        v->vc1dsp.put_vc1_mspel_pixels_tab[dxy](s->dest[0] + 8, srcY + 8, s->linesize, v->rnd);
 -        srcY += s->linesize * 8;
 -        v->vc1dsp.put_vc1_mspel_pixels_tab[dxy](s->dest[0] + 8 * s->linesize    , srcY    , s->linesize, v->rnd);
 -        v->vc1dsp.put_vc1_mspel_pixels_tab[dxy](s->dest[0] + 8 * s->linesize + 8, srcY + 8, s->linesize, v->rnd);
++                                 uvsrc_x, uvsrc_y,
++                                 s->h_edge_pos >> 1, v_edge_pos >> 1);
++        srcU = ubuf;
++        srcV = vbuf;
+         /* if we deal with range reduction we need to scale source blocks */
+         if (v->rangeredfrm) {
+             int i, j;
+             uint8_t *src, *src2;
+             src = srcY;
+             for (j = 0; j < 17 + s->mspel * 2; j++) {
+                 for (i = 0; i < 17 + s->mspel * 2; i++)
+                     src[i] = ((src[i] - 128) >> 1) + 128;
+                 src += s->linesize;
+             }
+             src  = srcU;
+             src2 = srcV;
+             for (j = 0; j < 9; j++) {
+                 for (i = 0; i < 9; i++) {
+                     src[i]  = ((src[i]  - 128) >> 1) + 128;
+                     src2[i] = ((src2[i] - 128) >> 1) + 128;
+                 }
+                 src  += s->uvlinesize;
+                 src2 += s->uvlinesize;
+             }
+         }
+         /* if we deal with intensity compensation we need to scale source blocks */
+         if (use_ic) {
+             int i, j;
+             uint8_t *src, *src2;
+             src = srcY;
+             for (j = 0; j < 17 + s->mspel * 2; j++) {
+                 int f = v->field_mode ? v->ref_field_type[dir] : ((j + src_y - s->mspel) & 1) ;
+                 for (i = 0; i < 17 + s->mspel * 2; i++)
+                     src[i] = luty[f][src[i]];
+                 src += s->linesize;
+             }
+             src  = srcU;
+             src2 = srcV;
+             for (j = 0; j < 9; j++) {
+                 int f = v->field_mode ? v->ref_field_type[dir] : ((j + uvsrc_y) & 1);
+                 for (i = 0; i < 9; i++) {
+                     src[i]  = lutuv[f][src[i]];
+                     src2[i] = lutuv[f][src2[i]];
+                 }
+                 src  += s->uvlinesize;
+                 src2 += s->uvlinesize;
+             }
+         }
+         srcY += s->mspel * (1 + s->linesize);
+     }
+     if (s->mspel) {
+         dxy = ((my & 3) << 2) | (mx & 3);
 -            use_ic = v->curr_use_ic;
++        v->vc1dsp.put_vc1_mspel_pixels_tab[0][dxy](s->dest[0]    , srcY    , s->linesize, v->rnd);
+     } else { // hpel mc - always used for luma
+         dxy = (my & 2) | ((mx & 2) >> 1);
+         if (!v->rnd)
+             s->hdsp.put_pixels_tab[0][dxy](s->dest[0], srcY, s->linesize, 16);
+         else
+             s->hdsp.put_no_rnd_pixels_tab[0][dxy](s->dest[0], srcY, s->linesize, 16);
+     }
+     if (s->flags & CODEC_FLAG_GRAY) return;
+     /* Chroma MC always uses qpel bilinear */
+     uvmx = (uvmx & 3) << 1;
+     uvmy = (uvmy & 3) << 1;
+     if (!v->rnd) {
+         h264chroma->put_h264_chroma_pixels_tab[0](s->dest[1], srcU, s->uvlinesize, 8, uvmx, uvmy);
+         h264chroma->put_h264_chroma_pixels_tab[0](s->dest[2], srcV, s->uvlinesize, 8, uvmx, uvmy);
+     } else {
+         v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[0](s->dest[1], srcU, s->uvlinesize, 8, uvmx, uvmy);
+         v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[0](s->dest[2], srcV, s->uvlinesize, 8, uvmx, uvmy);
+     }
+ }
+ static inline int median4(int a, int b, int c, int d)
+ {
+     if (a < b) {
+         if (c < d) return (FFMIN(b, d) + FFMAX(a, c)) / 2;
+         else       return (FFMIN(b, c) + FFMAX(a, d)) / 2;
+     } else {
+         if (c < d) return (FFMIN(a, d) + FFMAX(b, c)) / 2;
+         else       return (FFMIN(a, c) + FFMAX(b, d)) / 2;
+     }
+ }
+ /** Do motion compensation for 4-MV macroblock - luminance block
+  */
+ void ff_vc1_mc_4mv_luma(VC1Context *v, int n, int dir, int avg)
+ {
+     MpegEncContext *s = &v->s;
+     uint8_t *srcY;
+     int dxy, mx, my, src_x, src_y;
+     int off;
+     int fieldmv = (v->fcm == ILACE_FRAME) ? v->blk_mv_type[s->block_index[n]] : 0;
+     int v_edge_pos = s->v_edge_pos >> v->field_mode;
+     uint8_t (*luty)[256];
+     int use_ic;
+     if ((!v->field_mode ||
+          (v->ref_field_type[dir] == 1 && v->cur_field_type == 1)) &&
+         !v->s.last_picture.f->data[0])
+         return;
+     mx = s->mv[dir][n][0];
+     my = s->mv[dir][n][1];
+     if (!dir) {
+         if (v->field_mode && (v->cur_field_type != v->ref_field_type[dir]) && v->second_field) {
+             srcY = s->current_picture.f->data[0];
+             luty = v->curr_luty;
 -            v->vc1dsp.avg_vc1_mspel_pixels_tab[dxy](s->dest[0] + off, srcY, s->linesize << fieldmv, v->rnd);
++            use_ic = *v->curr_use_ic;
+         } else {
+             srcY = s->last_picture.f->data[0];
+             luty = v->last_luty;
+             use_ic = v->last_use_ic;
+         }
+     } else {
+         srcY = s->next_picture.f->data[0];
+         luty = v->next_luty;
+         use_ic = v->next_use_ic;
+     }
+     if (!srcY) {
+         av_log(v->s.avctx, AV_LOG_ERROR, "Referenced frame missing.\n");
+         return;
+     }
+     if (v->field_mode) {
+         if (v->cur_field_type != v->ref_field_type[dir])
+             my = my - 2 + 4 * v->cur_field_type;
+     }
+     if (s->pict_type == AV_PICTURE_TYPE_P && n == 3 && v->field_mode) {
+         int same_count = 0, opp_count = 0, k;
+         int chosen_mv[2][4][2], f;
+         int tx, ty;
+         for (k = 0; k < 4; k++) {
+             f = v->mv_f[0][s->block_index[k] + v->blocks_off];
+             chosen_mv[f][f ? opp_count : same_count][0] = s->mv[0][k][0];
+             chosen_mv[f][f ? opp_count : same_count][1] = s->mv[0][k][1];
+             opp_count  += f;
+             same_count += 1 - f;
+         }
+         f = opp_count > same_count;
+         switch (f ? opp_count : same_count) {
+         case 4:
+             tx = median4(chosen_mv[f][0][0], chosen_mv[f][1][0],
+                          chosen_mv[f][2][0], chosen_mv[f][3][0]);
+             ty = median4(chosen_mv[f][0][1], chosen_mv[f][1][1],
+                          chosen_mv[f][2][1], chosen_mv[f][3][1]);
+             break;
+         case 3:
+             tx = mid_pred(chosen_mv[f][0][0], chosen_mv[f][1][0], chosen_mv[f][2][0]);
+             ty = mid_pred(chosen_mv[f][0][1], chosen_mv[f][1][1], chosen_mv[f][2][1]);
+             break;
+         case 2:
+             tx = (chosen_mv[f][0][0] + chosen_mv[f][1][0]) / 2;
+             ty = (chosen_mv[f][0][1] + chosen_mv[f][1][1]) / 2;
+             break;
++        default:
++            av_assert0(0);
+         }
+         s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][0] = tx;
+         s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][1] = ty;
+         for (k = 0; k < 4; k++)
+             v->mv_f[1][s->block_index[k] + v->blocks_off] = f;
+     }
+     if (v->fcm == ILACE_FRAME) {  // not sure if needed for other types of picture
+         int qx, qy;
+         int width  = s->avctx->coded_width;
+         int height = s->avctx->coded_height >> 1;
+         if (s->pict_type == AV_PICTURE_TYPE_P) {
+             s->current_picture.motion_val[1][s->block_index[n] + v->blocks_off][0] = mx;
+             s->current_picture.motion_val[1][s->block_index[n] + v->blocks_off][1] = my;
+         }
+         qx = (s->mb_x * 16) + (mx >> 2);
+         qy = (s->mb_y *  8) + (my >> 3);
+         if (qx < -17)
+             mx -= 4 * (qx + 17);
+         else if (qx > width)
+             mx -= 4 * (qx - width);
+         if (qy < -18)
+             my -= 8 * (qy + 18);
+         else if (qy > height + 1)
+             my -= 8 * (qy - height - 1);
+     }
+     if ((v->fcm == ILACE_FRAME) && fieldmv)
+         off = ((n > 1) ? s->linesize : 0) + (n & 1) * 8;
+     else
+         off = s->linesize * 4 * (n & 2) + (n & 1) * 8;
+     src_x = s->mb_x * 16 + (n & 1) * 8 + (mx >> 2);
+     if (!fieldmv)
+         src_y = s->mb_y * 16 + (n & 2) * 4 + (my >> 2);
+     else
+         src_y = s->mb_y * 16 + ((n > 1) ? 1 : 0) + (my >> 2);
+     if (v->profile != PROFILE_ADVANCED) {
+         src_x = av_clip(src_x, -16, s->mb_width  * 16);
+         src_y = av_clip(src_y, -16, s->mb_height * 16);
+     } else {
+         src_x = av_clip(src_x, -17, s->avctx->coded_width);
+         if (v->fcm == ILACE_FRAME) {
+             if (src_y & 1)
+                 src_y = av_clip(src_y, -17, s->avctx->coded_height + 1);
+             else
+                 src_y = av_clip(src_y, -18, s->avctx->coded_height);
+         } else {
+             src_y = av_clip(src_y, -18, s->avctx->coded_height + 1);
+         }
+     }
+     srcY += src_y * s->linesize + src_x;
+     if (v->field_mode && v->ref_field_type[dir])
+         srcY += s->current_picture_ptr->f->linesize[0];
+     if (fieldmv && !(src_y & 1))
+         v_edge_pos--;
+     if (fieldmv && (src_y & 1) && src_y < 4)
+         src_y--;
+     if (v->rangeredfrm || use_ic
+         || s->h_edge_pos < 13 || v_edge_pos < 23
+         || (unsigned)(src_x - s->mspel) > s->h_edge_pos - (mx & 3) - 8 - s->mspel * 2
+         || (unsigned)(src_y - (s->mspel << fieldmv)) > v_edge_pos - (my & 3) - ((8 + s->mspel * 2) << fieldmv)) {
+         srcY -= s->mspel * (1 + (s->linesize << fieldmv));
+         /* check emulate edge stride and offset */
+         s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
+                                  s->linesize, s->linesize,
+                                  9 + s->mspel * 2, (9 + s->mspel * 2) << fieldmv,
+                                  src_x - s->mspel, src_y - (s->mspel << fieldmv),
+                                  s->h_edge_pos, v_edge_pos);
+         srcY = s->edge_emu_buffer;
+         /* if we deal with range reduction we need to scale source blocks */
+         if (v->rangeredfrm) {
+             int i, j;
+             uint8_t *src;
+             src = srcY;
+             for (j = 0; j < 9 + s->mspel * 2; j++) {
+                 for (i = 0; i < 9 + s->mspel * 2; i++)
+                     src[i] = ((src[i] - 128) >> 1) + 128;
+                 src += s->linesize << fieldmv;
+             }
+         }
+         /* if we deal with intensity compensation we need to scale source blocks */
+         if (use_ic) {
+             int i, j;
+             uint8_t *src;
+             src = srcY;
+             for (j = 0; j < 9 + s->mspel * 2; j++) {
+                 int f = v->field_mode ? v->ref_field_type[dir] : (((j<<fieldmv)+src_y - (s->mspel << fieldmv)) & 1);
+                 for (i = 0; i < 9 + s->mspel * 2; i++)
+                     src[i] = luty[f][src[i]];
+                 src += s->linesize << fieldmv;
+             }
+         }
+         srcY += s->mspel * (1 + (s->linesize << fieldmv));
+     }
+     if (s->mspel) {
+         dxy = ((my & 3) << 2) | (mx & 3);
+         if (avg)
 -            v->vc1dsp.put_vc1_mspel_pixels_tab[dxy](s->dest[0] + off, srcY, s->linesize << fieldmv, v->rnd);
++            v->vc1dsp.avg_vc1_mspel_pixels_tab[1][dxy](s->dest[0] + off, srcY, s->linesize << fieldmv, v->rnd);
+         else
 -            use_ic = v->curr_use_ic;
++            v->vc1dsp.put_vc1_mspel_pixels_tab[1][dxy](s->dest[0] + off, srcY, s->linesize << fieldmv, v->rnd);
+     } else { // hpel mc - always used for luma
+         dxy = (my & 2) | ((mx & 2) >> 1);
+         if (!v->rnd)
+             s->hdsp.put_pixels_tab[1][dxy](s->dest[0] + off, srcY, s->linesize, 8);
+         else
+             s->hdsp.put_no_rnd_pixels_tab[1][dxy](s->dest[0] + off, srcY, s->linesize, 8);
+     }
+ }
+ static av_always_inline int get_chroma_mv(int *mvx, int *mvy, int *a, int flag, int *tx, int *ty)
+ {
+     int idx, i;
+     static const int count[16] = { 0, 1, 1, 2, 1, 2, 2, 3, 1, 2, 2, 3, 2, 3, 3, 4};
+     idx =  ((a[3] != flag) << 3)
+          | ((a[2] != flag) << 2)
+          | ((a[1] != flag) << 1)
+          |  (a[0] != flag);
+     if (!idx) {
+         *tx = median4(mvx[0], mvx[1], mvx[2], mvx[3]);
+         *ty = median4(mvy[0], mvy[1], mvy[2], mvy[3]);
+         return 4;
+     } else if (count[idx] == 1) {
+         switch (idx) {
+         case 0x1:
+             *tx = mid_pred(mvx[1], mvx[2], mvx[3]);
+             *ty = mid_pred(mvy[1], mvy[2], mvy[3]);
+             return 3;
+         case 0x2:
+             *tx = mid_pred(mvx[0], mvx[2], mvx[3]);
+             *ty = mid_pred(mvy[0], mvy[2], mvy[3]);
+             return 3;
+         case 0x4:
+             *tx = mid_pred(mvx[0], mvx[1], mvx[3]);
+             *ty = mid_pred(mvy[0], mvy[1], mvy[3]);
+             return 3;
+         case 0x8:
+             *tx = mid_pred(mvx[0], mvx[1], mvx[2]);
+             *ty = mid_pred(mvy[0], mvy[1], mvy[2]);
+             return 3;
+         }
+     } else if (count[idx] == 2) {
+         int t1 = 0, t2 = 0;
+         for (i = 0; i < 3; i++)
+             if (!a[i]) {
+                 t1 = i;
+                 break;
+             }
+         for (i = t1 + 1; i < 4; i++)
+             if (!a[i]) {
+                 t2 = i;
+                 break;
+             }
+         *tx = (mvx[t1] + mvx[t2]) / 2;
+         *ty = (mvy[t1] + mvy[t2]) / 2;
+         return 2;
+     } else {
+         return 0;
+     }
+     return -1;
+ }
+ /** Do motion compensation for 4-MV macroblock - both chroma blocks
+  */
+ void ff_vc1_mc_4mv_chroma(VC1Context *v, int dir)
+ {
+     MpegEncContext *s = &v->s;
+     H264ChromaContext *h264chroma = &v->h264chroma;
+     uint8_t *srcU, *srcV;
+     int uvmx, uvmy, uvsrc_x, uvsrc_y;
+     int k, tx = 0, ty = 0;
+     int mvx[4], mvy[4], intra[4], mv_f[4];
+     int valid_count;
+     int chroma_ref_type = v->cur_field_type;
+     int v_edge_pos = s->v_edge_pos >> v->field_mode;
+     uint8_t (*lutuv)[256];
+     int use_ic;
+     if (!v->field_mode && !v->s.last_picture.f->data[0])
+         return;
+     if (s->flags & CODEC_FLAG_GRAY)
+         return;
+     for (k = 0; k < 4; k++) {
+         mvx[k] = s->mv[dir][k][0];
+         mvy[k] = s->mv[dir][k][1];
+         intra[k] = v->mb_type[0][s->block_index[k]];
+         if (v->field_mode)
+             mv_f[k] = v->mv_f[dir][s->block_index[k] + v->blocks_off];
+     }
+     /* calculate chroma MV vector from four luma MVs */
+     if (!v->field_mode || (v->field_mode && !v->numref)) {
+         valid_count = get_chroma_mv(mvx, mvy, intra, 0, &tx, &ty);
+         chroma_ref_type = v->reffield;
+         if (!valid_count) {
+             s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][0] = 0;
+             s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][1] = 0;
+             v->luma_mv[s->mb_x][0] = v->luma_mv[s->mb_x][1] = 0;
+             return; //no need to do MC for intra blocks
+         }
+     } else {
+         int dominant = 0;
+         if (mv_f[0] + mv_f[1] + mv_f[2] + mv_f[3] > 2)
+             dominant = 1;
+         valid_count = get_chroma_mv(mvx, mvy, mv_f, dominant, &tx, &ty);
+         if (dominant)
+             chroma_ref_type = !v->cur_field_type;
+     }
+     if (v->field_mode && chroma_ref_type == 1 && v->cur_field_type == 1 && !v->s.last_picture.f->data[0])
+         return;
+     s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][0] = tx;
+     s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][1] = ty;
+     uvmx = (tx + ((tx & 3) == 3)) >> 1;
+     uvmy = (ty + ((ty & 3) == 3)) >> 1;
+     v->luma_mv[s->mb_x][0] = uvmx;
+     v->luma_mv[s->mb_x][1] = uvmy;
+     if (v->fastuvmc) {
+         uvmx = uvmx + ((uvmx < 0) ? (uvmx & 1) : -(uvmx & 1));
+         uvmy = uvmy + ((uvmy < 0) ? (uvmy & 1) : -(uvmy & 1));
+     }
+     // Field conversion bias
+     if (v->cur_field_type != chroma_ref_type)
+         uvmy += 2 - 4 * chroma_ref_type;
+     uvsrc_x = s->mb_x * 8 + (uvmx >> 2);
+     uvsrc_y = s->mb_y * 8 + (uvmy >> 2);
+     if (v->profile != PROFILE_ADVANCED) {
+         uvsrc_x = av_clip(uvsrc_x, -8, s->mb_width  * 8);
+         uvsrc_y = av_clip(uvsrc_y, -8, s->mb_height * 8);
+     } else {
+         uvsrc_x = av_clip(uvsrc_x, -8, s->avctx->coded_width  >> 1);
+         uvsrc_y = av_clip(uvsrc_y, -8, s->avctx->coded_height >> 1);
+     }
+     if (!dir) {
+         if (v->field_mode && (v->cur_field_type != chroma_ref_type) && v->second_field) {
+             srcU = s->current_picture.f->data[1];
+             srcV = s->current_picture.f->data[2];
+             lutuv = v->curr_lutuv;
 -            srcU = s->next_picture.f->data[1] + uvsrc_y * s->uvlinesize + uvsrc_x;
 -            srcV = s->next_picture.f->data[2] + uvsrc_y * s->uvlinesize + uvsrc_x;
++            use_ic = *v->curr_use_ic;
+         } else {
+             srcU = s->last_picture.f->data[1];
+             srcV = s->last_picture.f->data[2];
+             lutuv = v->last_lutuv;
+             use_ic = v->last_use_ic;
+         }
+     } else {
+         srcU = s->next_picture.f->data[1];
+         srcV = s->next_picture.f->data[2];
+         lutuv = v->next_lutuv;
+         use_ic = v->next_use_ic;
+     }
+     if (!srcU) {
+         av_log(v->s.avctx, AV_LOG_ERROR, "Referenced frame missing.\n");
+         return;
+     }
+     srcU += uvsrc_y * s->uvlinesize + uvsrc_x;
+     srcV += uvsrc_y * s->uvlinesize + uvsrc_x;
+     if (v->field_mode) {
+         if (chroma_ref_type) {
+             srcU += s->current_picture_ptr->f->linesize[1];
+             srcV += s->current_picture_ptr->f->linesize[2];
+         }
+     }
+     if (v->rangeredfrm || use_ic
+         || s->h_edge_pos < 18 || v_edge_pos < 18
+         || (unsigned)uvsrc_x > (s->h_edge_pos >> 1) - 9
+         || (unsigned)uvsrc_y > (v_edge_pos    >> 1) - 9) {
+         s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcU,
+                                  s->uvlinesize, s->uvlinesize,
+                                  8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
+                                  s->h_edge_pos >> 1, v_edge_pos >> 1);
+         s->vdsp.emulated_edge_mc(s->edge_emu_buffer + 16, srcV,
+                                  s->uvlinesize, s->uvlinesize,
+                                  8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
+                                  s->h_edge_pos >> 1, v_edge_pos >> 1);
+         srcU = s->edge_emu_buffer;
+         srcV = s->edge_emu_buffer + 16;
+         /* if we deal with range reduction we need to scale source blocks */
+         if (v->rangeredfrm) {
+             int i, j;
+             uint8_t *src, *src2;
+             src  = srcU;
+             src2 = srcV;
+             for (j = 0; j < 9; j++) {
+                 for (i = 0; i < 9; i++) {
+                     src[i]  = ((src[i]  - 128) >> 1) + 128;
+                     src2[i] = ((src2[i] - 128) >> 1) + 128;
+                 }
+                 src  += s->uvlinesize;
+                 src2 += s->uvlinesize;
+             }
+         }
+         /* if we deal with intensity compensation we need to scale source blocks */
+         if (use_ic) {
+             int i, j;
+             uint8_t *src, *src2;
+             src  = srcU;
+             src2 = srcV;
+             for (j = 0; j < 9; j++) {
+                 int f = v->field_mode ? chroma_ref_type : ((j + uvsrc_y) & 1);
+                 for (i = 0; i < 9; i++) {
+                     src[i]  = lutuv[f][src[i]];
+                     src2[i] = lutuv[f][src2[i]];
+                 }
+                 src  += s->uvlinesize;
+                 src2 += s->uvlinesize;
+             }
+         }
+     }
+     /* Chroma MC always uses qpel bilinear */
+     uvmx = (uvmx & 3) << 1;
+     uvmy = (uvmy & 3) << 1;
+     if (!v->rnd) {
+         h264chroma->put_h264_chroma_pixels_tab[0](s->dest[1], srcU, s->uvlinesize, 8, uvmx, uvmy);
+         h264chroma->put_h264_chroma_pixels_tab[0](s->dest[2], srcV, s->uvlinesize, 8, uvmx, uvmy);
+     } else {
+         v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[0](s->dest[1], srcU, s->uvlinesize, 8, uvmx, uvmy);
+         v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[0](s->dest[2], srcV, s->uvlinesize, 8, uvmx, uvmy);
+     }
+ }
+ /** Do motion compensation for 4-MV interlaced frame chroma macroblock (both U and V)
+  */
+ void ff_vc1_mc_4mv_chroma4(VC1Context *v, int dir, int dir2, int avg)
+ {
+     MpegEncContext *s = &v->s;
+     H264ChromaContext *h264chroma = &v->h264chroma;
+     uint8_t *srcU, *srcV;
+     int uvsrc_x, uvsrc_y;
+     int uvmx_field[4], uvmy_field[4];
+     int i, off, tx, ty;
+     int fieldmv = v->blk_mv_type[s->block_index[0]];
+     static const int s_rndtblfield[16] = { 0, 0, 1, 2, 4, 4, 5, 6, 2, 2, 3, 8, 6, 6, 7, 12 };
+     int v_dist = fieldmv ? 1 : 4; // vertical offset for lower sub-blocks
+     int v_edge_pos = s->v_edge_pos >> 1;
+     int use_ic;
+     uint8_t (*lutuv)[256];
+     if (s->flags & CODEC_FLAG_GRAY)
+         return;
+     for (i = 0; i < 4; i++) {
+         int d = i < 2 ? dir: dir2;
+         tx = s->mv[d][i][0];
+         uvmx_field[i] = (tx + ((tx & 3) == 3)) >> 1;
+         ty = s->mv[d][i][1];
+         if (fieldmv)
+             uvmy_field[i] = (ty >> 4) * 8 + s_rndtblfield[ty & 0xF];
+         else
+             uvmy_field[i] = (ty + ((ty & 3) == 3)) >> 1;
+     }
+     for (i = 0; i < 4; i++) {
+         off = (i & 1) * 4 + ((i & 2) ? v_dist * s->uvlinesize : 0);
+         uvsrc_x = s->mb_x * 8 +  (i & 1) * 4           + (uvmx_field[i] >> 2);
+         uvsrc_y = s->mb_y * 8 + ((i & 2) ? v_dist : 0) + (uvmy_field[i] >> 2);
+         // FIXME: implement proper pull-back (see vc1cropmv.c, vc1CROPMV_ChromaPullBack())
+         uvsrc_x = av_clip(uvsrc_x, -8, s->avctx->coded_width  >> 1);
+         uvsrc_y = av_clip(uvsrc_y, -8, s->avctx->coded_height >> 1);
+         if (i < 2 ? dir : dir2) {
 -            srcU = s->last_picture.f->data[1] + uvsrc_y * s->uvlinesize + uvsrc_x;
 -            srcV = s->last_picture.f->data[2] + uvsrc_y * s->uvlinesize + uvsrc_x;
++            srcU = s->next_picture.f->data[1];
++            srcV = s->next_picture.f->data[2];
+             lutuv  = v->next_lutuv;
+             use_ic = v->next_use_ic;
+         } else {
 -            v_edge_pos--;
++            srcU = s->last_picture.f->data[1];
++            srcV = s->last_picture.f->data[2];
+             lutuv  = v->last_lutuv;
+             use_ic = v->last_use_ic;
+         }
++        if (!srcU)
++            return;
++        srcU += uvsrc_y * s->uvlinesize + uvsrc_x;
++        srcV += uvsrc_y * s->uvlinesize + uvsrc_x;
+         uvmx_field[i] = (uvmx_field[i] & 3) << 1;
+         uvmy_field[i] = (uvmy_field[i] & 3) << 1;
+         if (fieldmv && !(uvsrc_y & 1))
 -        if (v->cur_field_type != v->ref_field_type[1])
++            v_edge_pos = (s->v_edge_pos >> 1) - 1;
++
+         if (fieldmv && (uvsrc_y & 1) && uvsrc_y < 2)
+             uvsrc_y--;
+         if (use_ic
+             || s->h_edge_pos < 10 || v_edge_pos < (5 << fieldmv)
+             || (unsigned)uvsrc_x > (s->h_edge_pos >> 1) - 5
+             || (unsigned)uvsrc_y > v_edge_pos - (5 << fieldmv)) {
+             s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcU,
+                                      s->uvlinesize, s->uvlinesize,
+                                      5, (5 << fieldmv), uvsrc_x, uvsrc_y,
+                                      s->h_edge_pos >> 1, v_edge_pos);
+             s->vdsp.emulated_edge_mc(s->edge_emu_buffer + 16, srcV,
+                                      s->uvlinesize, s->uvlinesize,
+                                      5, (5 << fieldmv), uvsrc_x, uvsrc_y,
+                                      s->h_edge_pos >> 1, v_edge_pos);
+             srcU = s->edge_emu_buffer;
+             srcV = s->edge_emu_buffer + 16;
+             /* if we deal with intensity compensation we need to scale source blocks */
+             if (use_ic) {
+                 int i, j;
+                 uint8_t *src, *src2;
+                 src  = srcU;
+                 src2 = srcV;
+                 for (j = 0; j < 5; j++) {
+                     int f = (uvsrc_y + (j << fieldmv)) & 1;
+                     for (i = 0; i < 5; i++) {
+                         src[i]  = lutuv[f][src[i]];
+                         src2[i] = lutuv[f][src2[i]];
+                     }
+                     src  += s->uvlinesize << fieldmv;
+                     src2 += s->uvlinesize << fieldmv;
+                 }
+             }
+         }
+         if (avg) {
+             if (!v->rnd) {
+                 h264chroma->avg_h264_chroma_pixels_tab[1](s->dest[1] + off, srcU, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
+                 h264chroma->avg_h264_chroma_pixels_tab[1](s->dest[2] + off, srcV, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
+             } else {
+                 v->vc1dsp.avg_no_rnd_vc1_chroma_pixels_tab[1](s->dest[1] + off, srcU, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
+                 v->vc1dsp.avg_no_rnd_vc1_chroma_pixels_tab[1](s->dest[2] + off, srcV, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
+             }
+         } else {
+             if (!v->rnd) {
+                 h264chroma->put_h264_chroma_pixels_tab[1](s->dest[1] + off, srcU, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
+                 h264chroma->put_h264_chroma_pixels_tab[1](s->dest[2] + off, srcV, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
+             } else {
+                 v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[1](s->dest[1] + off, srcU, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
+                 v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[1](s->dest[2] + off, srcV, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
+             }
+         }
+     }
+ }
+ /** Motion compensation for direct or interpolated blocks in B-frames
+  */
+ void ff_vc1_interp_mc(VC1Context *v)
+ {
+     MpegEncContext *s = &v->s;
+     H264ChromaContext *h264chroma = &v->h264chroma;
+     uint8_t *srcY, *srcU, *srcV;
+     int dxy, mx, my, uvmx, uvmy, src_x, src_y, uvsrc_x, uvsrc_y;
+     int off, off_uv;
+     int v_edge_pos = s->v_edge_pos >> v->field_mode;
+     int use_ic = v->next_use_ic;
+     if (!v->field_mode && !v->s.next_picture.f->data[0])
+         return;
+     mx   = s->mv[1][0][0];
+     my   = s->mv[1][0][1];
+     uvmx = (mx + ((mx & 3) == 3)) >> 1;
+     uvmy = (my + ((my & 3) == 3)) >> 1;
+     if (v->field_mode) {
 -        uint8_t *uvbuf = s->edge_emu_buffer + 19 * s->linesize;
++        if (v->cur_field_type != v->ref_field_type[1]) {
+             my   = my   - 2 + 4 * v->cur_field_type;
+             uvmy = uvmy - 2 + 4 * v->cur_field_type;
++        }
+     }
+     if (v->fastuvmc) {
+         uvmx = uvmx + ((uvmx < 0) ? -(uvmx & 1) : (uvmx & 1));
+         uvmy = uvmy + ((uvmy < 0) ? -(uvmy & 1) : (uvmy & 1));
+     }
+     srcY = s->next_picture.f->data[0];
+     srcU = s->next_picture.f->data[1];
+     srcV = s->next_picture.f->data[2];
+     src_x   = s->mb_x * 16 + (mx   >> 2);
+     src_y   = s->mb_y * 16 + (my   >> 2);
+     uvsrc_x = s->mb_x *  8 + (uvmx >> 2);
+     uvsrc_y = s->mb_y *  8 + (uvmy >> 2);
+     if (v->profile != PROFILE_ADVANCED) {
+         src_x   = av_clip(  src_x, -16, s->mb_width  * 16);
+         src_y   = av_clip(  src_y, -16, s->mb_height * 16);
+         uvsrc_x = av_clip(uvsrc_x,  -8, s->mb_width  *  8);
+         uvsrc_y = av_clip(uvsrc_y,  -8, s->mb_height *  8);
+     } else {
+         src_x   = av_clip(  src_x, -17, s->avctx->coded_width);
+         src_y   = av_clip(  src_y, -18, s->avctx->coded_height + 1);
+         uvsrc_x = av_clip(uvsrc_x,  -8, s->avctx->coded_width  >> 1);
+         uvsrc_y = av_clip(uvsrc_y,  -8, s->avctx->coded_height >> 1);
+     }
+     srcY += src_y   * s->linesize   + src_x;
+     srcU += uvsrc_y * s->uvlinesize + uvsrc_x;
+     srcV += uvsrc_y * s->uvlinesize + uvsrc_x;
+     if (v->field_mode && v->ref_field_type[1]) {
+         srcY += s->current_picture_ptr->f->linesize[0];
+         srcU += s->current_picture_ptr->f->linesize[1];
+         srcV += s->current_picture_ptr->f->linesize[2];
+     }
+     /* for grayscale we should not try to read from unknown area */
+     if (s->flags & CODEC_FLAG_GRAY) {
+         srcU = s->edge_emu_buffer + 18 * s->linesize;
+         srcV = s->edge_emu_buffer + 18 * s->linesize;
+     }
+     if (v->rangeredfrm || s->h_edge_pos < 22 || v_edge_pos < 22 || use_ic
+         || (unsigned)(src_x - 1) > s->h_edge_pos - (mx & 3) - 16 - 3
+         || (unsigned)(src_y - 1) > v_edge_pos    - (my & 3) - 16 - 3) {
 -        s->vdsp.emulated_edge_mc(uvbuf, srcU,
++        uint8_t *ubuf = s->edge_emu_buffer + 19 * s->linesize;
++        uint8_t *vbuf = ubuf + 9 * s->uvlinesize;
+         srcY -= s->mspel * (1 + s->linesize);
+         s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
+                                  s->linesize, s->linesize,
+                                  17 + s->mspel * 2, 17 + s->mspel * 2,
+                                  src_x - s->mspel, src_y - s->mspel,
+                                  s->h_edge_pos, v_edge_pos);
+         srcY = s->edge_emu_buffer;
 -                                 uvsrc_x, uvsrc_y, s->h_edge_pos >> 1, v_edge_pos >> 1);
 -        s->vdsp.emulated_edge_mc(uvbuf + 16, srcV,
++        s->vdsp.emulated_edge_mc(ubuf, srcU,
+                                  s->uvlinesize, s->uvlinesize,
+                                  8 + 1, 8 + 1,
 -                                 uvsrc_x, uvsrc_y, s->h_edge_pos >> 1, v_edge_pos >> 1);
 -        srcU = uvbuf;
 -        srcV = uvbuf + 16;
++                                 uvsrc_x, uvsrc_y,
++                                 s->h_edge_pos >> 1, v_edge_pos >> 1);
++        s->vdsp.emulated_edge_mc(vbuf, srcV,
+                                  s->uvlinesize, s->uvlinesize,
+                                  8 + 1, 8 + 1,
 -        v->vc1dsp.avg_vc1_mspel_pixels_tab[dxy](s->dest[0] + off    , srcY    , s->linesize, v->rnd);
 -        v->vc1dsp.avg_vc1_mspel_pixels_tab[dxy](s->dest[0] + off + 8, srcY + 8, s->linesize, v->rnd);
 -        srcY += s->linesize * 8;
 -        v->vc1dsp.avg_vc1_mspel_pixels_tab[dxy](s->dest[0] + off + 8 * s->linesize    , srcY    , s->linesize, v->rnd);
 -        v->vc1dsp.avg_vc1_mspel_pixels_tab[dxy](s->dest[0] + off + 8 * s->linesize + 8, srcY + 8, s->linesize, v->rnd);
++                                 uvsrc_x, uvsrc_y,
++                                 s->h_edge_pos >> 1, v_edge_pos >> 1);
++        srcU = ubuf;
++        srcV = vbuf;
+         /* if we deal with range reduction we need to scale source blocks */
+         if (v->rangeredfrm) {
+             int i, j;
+             uint8_t *src, *src2;
+             src = srcY;
+             for (j = 0; j < 17 + s->mspel * 2; j++) {
+                 for (i = 0; i < 17 + s->mspel * 2; i++)
+                     src[i] = ((src[i] - 128) >> 1) + 128;
+                 src += s->linesize;
+             }
+             src = srcU;
+             src2 = srcV;
+             for (j = 0; j < 9; j++) {
+                 for (i = 0; i < 9; i++) {
+                     src[i]  = ((src[i]  - 128) >> 1) + 128;
+                     src2[i] = ((src2[i] - 128) >> 1) + 128;
+                 }
+                 src  += s->uvlinesize;
+                 src2 += s->uvlinesize;
+             }
+         }
+         if (use_ic) {
+             uint8_t (*luty )[256] = v->next_luty;
+             uint8_t (*lutuv)[256] = v->next_lutuv;
+             int i, j;
+             uint8_t *src, *src2;
+             src = srcY;
+             for (j = 0; j < 17 + s->mspel * 2; j++) {
+                 int f = v->field_mode ? v->ref_field_type[1] : ((j+src_y - s->mspel) & 1);
+                 for (i = 0; i < 17 + s->mspel * 2; i++)
+                     src[i] = luty[f][src[i]];
+                 src += s->linesize;
+             }
+             src  = srcU;
+             src2 = srcV;
+             for (j = 0; j < 9; j++) {
+                 int f = v->field_mode ? v->ref_field_type[1] : ((j+uvsrc_y) & 1);
+                 for (i = 0; i < 9; i++) {
+                     src[i]  = lutuv[f][src[i]];
+                     src2[i] = lutuv[f][src2[i]];
+                 }
+                 src  += s->uvlinesize;
+                 src2 += s->uvlinesize;
+             }
+         }
+         srcY += s->mspel * (1 + s->linesize);
+     }
+     off    = 0;
+     off_uv = 0;
+     if (s->mspel) {
+         dxy = ((my & 3) << 2) | (mx & 3);
++        v->vc1dsp.avg_vc1_mspel_pixels_tab[0][dxy](s->dest[0] + off    , srcY    , s->linesize, v->rnd);
+     } else { // hpel mc
+         dxy = (my & 2) | ((mx & 2) >> 1);
+         if (!v->rnd)
+             s->hdsp.avg_pixels_tab[0][dxy](s->dest[0] + off, srcY, s->linesize, 16);
+         else
+             s->hdsp.avg_no_rnd_pixels_tab[dxy](s->dest[0] + off, srcY, s->linesize, 16);
+     }
+     if (s->flags & CODEC_FLAG_GRAY) return;
+     /* Chroma MC always uses qpel blilinear */
+     uvmx = (uvmx & 3) << 1;
+     uvmy = (uvmy & 3) << 1;
+     if (!v->rnd) {
+         h264chroma->avg_h264_chroma_pixels_tab[0](s->dest[1] + off_uv, srcU, s->uvlinesize, 8, uvmx, uvmy);
+         h264chroma->avg_h264_chroma_pixels_tab[0](s->dest[2] + off_uv, srcV, s->uvlinesize, 8, uvmx, uvmy);
+     } else {
+         v->vc1dsp.avg_no_rnd_vc1_chroma_pixels_tab[0](s->dest[1] + off_uv, srcU, s->uvlinesize, 8, uvmx, uvmy);
+         v->vc1dsp.avg_no_rnd_vc1_chroma_pixels_tab[0](s->dest[2] + off_uv, srcV, s->uvlinesize, 8, uvmx, uvmy);
+     }
+ }
index 0000000,6a54fe4..cf1dce6
mode 000000,100644..100644
--- /dev/null
@@@ -1,0 -1,959 +1,961 @@@
 - * This file is part of Libav.
+ /*
+  * VC-1 and WMV3 decoder
+  * Copyright (c) 2011 Mashiat Sarker Shakkhar
+  * Copyright (c) 2006-2007 Konstantin Shishkov
+  * Partly based on vc9.c (c) 2005 Anonymous, Alex Beregszaszi, Michael Niedermayer
+  *
 - * Libav is free software; you can redistribute it and/or
++ * This file is part of FFmpeg.
+  *
 - * Libav is distributed in the hope that it will be useful,
++ * FFmpeg is free software; you can redistribute it and/or
+  * modify it under the terms of the GNU Lesser General Public
+  * License as published by the Free Software Foundation; either
+  * version 2.1 of the License, or (at your option) any later version.
+  *
 - * License along with Libav; if not, write to the Free Software
++ * FFmpeg is distributed in the hope that it will be useful,
+  * but WITHOUT ANY WARRANTY; without even the implied warranty of
+  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+  * Lesser General Public License for more details.
+  *
+  * You should have received a copy of the GNU Lesser General Public
 -    int px, py;
++ * License along with FFmpeg; if not, write to the Free Software
+  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+  */
+ /**
+  * @file
+  * VC-1 and WMV3 block decoding routines
+  */
+ #include "mathops.h"
+ #include "mpegutils.h"
+ #include "mpegvideo.h"
+ #include "vc1.h"
+ #include "vc1_pred.h"
+ #include "vc1data.h"
+ static av_always_inline int scaleforsame_x(VC1Context *v, int n /* MV */, int dir)
+ {
+     int scaledvalue, refdist;
+     int scalesame1, scalesame2;
+     int scalezone1_x, zone1offset_x;
+     int table_index = dir ^ v->second_field;
+     if (v->s.pict_type != AV_PICTURE_TYPE_B)
+         refdist = v->refdist;
+     else
+         refdist = dir ? v->brfd : v->frfd;
+     if (refdist > 3)
+         refdist = 3;
+     scalesame1    = ff_vc1_field_mvpred_scales[table_index][1][refdist];
+     scalesame2    = ff_vc1_field_mvpred_scales[table_index][2][refdist];
+     scalezone1_x  = ff_vc1_field_mvpred_scales[table_index][3][refdist];
+     zone1offset_x = ff_vc1_field_mvpred_scales[table_index][5][refdist];
+     if (FFABS(n) > 255)
+         scaledvalue = n;
+     else {
+         if (FFABS(n) < scalezone1_x)
+             scaledvalue = (n * scalesame1) >> 8;
+         else {
+             if (n < 0)
+                 scaledvalue = ((n * scalesame2) >> 8) - zone1offset_x;
+             else
+                 scaledvalue = ((n * scalesame2) >> 8) + zone1offset_x;
+         }
+     }
+     return av_clip(scaledvalue, -v->range_x, v->range_x - 1);
+ }
+ static av_always_inline int scaleforsame_y(VC1Context *v, int i, int n /* MV */, int dir)
+ {
+     int scaledvalue, refdist;
+     int scalesame1, scalesame2;
+     int scalezone1_y, zone1offset_y;
+     int table_index = dir ^ v->second_field;
+     if (v->s.pict_type != AV_PICTURE_TYPE_B)
+         refdist = v->refdist;
+     else
+         refdist = dir ? v->brfd : v->frfd;
+     if (refdist > 3)
+         refdist = 3;
+     scalesame1    = ff_vc1_field_mvpred_scales[table_index][1][refdist];
+     scalesame2    = ff_vc1_field_mvpred_scales[table_index][2][refdist];
+     scalezone1_y  = ff_vc1_field_mvpred_scales[table_index][4][refdist];
+     zone1offset_y = ff_vc1_field_mvpred_scales[table_index][6][refdist];
+     if (FFABS(n) > 63)
+         scaledvalue = n;
+     else {
+         if (FFABS(n) < scalezone1_y)
+             scaledvalue = (n * scalesame1) >> 8;
+         else {
+             if (n < 0)
+                 scaledvalue = ((n * scalesame2) >> 8) - zone1offset_y;
+             else
+                 scaledvalue = ((n * scalesame2) >> 8) + zone1offset_y;
+         }
+     }
+     if (v->cur_field_type && !v->ref_field_type[dir])
+         return av_clip(scaledvalue, -v->range_y / 2 + 1, v->range_y / 2);
+     else
+         return av_clip(scaledvalue, -v->range_y / 2, v->range_y / 2 - 1);
+ }
+ static av_always_inline int scaleforopp_x(VC1Context *v, int n /* MV */)
+ {
+     int scalezone1_x, zone1offset_x;
+     int scaleopp1, scaleopp2, brfd;
+     int scaledvalue;
+     brfd = FFMIN(v->brfd, 3);
+     scalezone1_x  = ff_vc1_b_field_mvpred_scales[3][brfd];
+     zone1offset_x = ff_vc1_b_field_mvpred_scales[5][brfd];
+     scaleopp1     = ff_vc1_b_field_mvpred_scales[1][brfd];
+     scaleopp2     = ff_vc1_b_field_mvpred_scales[2][brfd];
+     if (FFABS(n) > 255)
+         scaledvalue = n;
+     else {
+         if (FFABS(n) < scalezone1_x)
+             scaledvalue = (n * scaleopp1) >> 8;
+         else {
+             if (n < 0)
+                 scaledvalue = ((n * scaleopp2) >> 8) - zone1offset_x;
+             else
+                 scaledvalue = ((n * scaleopp2) >> 8) + zone1offset_x;
+         }
+     }
+     return av_clip(scaledvalue, -v->range_x, v->range_x - 1);
+ }
+ static av_always_inline int scaleforopp_y(VC1Context *v, int n /* MV */, int dir)
+ {
+     int scalezone1_y, zone1offset_y;
+     int scaleopp1, scaleopp2, brfd;
+     int scaledvalue;
+     brfd = FFMIN(v->brfd, 3);
+     scalezone1_y  = ff_vc1_b_field_mvpred_scales[4][brfd];
+     zone1offset_y = ff_vc1_b_field_mvpred_scales[6][brfd];
+     scaleopp1     = ff_vc1_b_field_mvpred_scales[1][brfd];
+     scaleopp2     = ff_vc1_b_field_mvpred_scales[2][brfd];
+     if (FFABS(n) > 63)
+         scaledvalue = n;
+     else {
+         if (FFABS(n) < scalezone1_y)
+             scaledvalue = (n * scaleopp1) >> 8;
+         else {
+             if (n < 0)
+                 scaledvalue = ((n * scaleopp2) >> 8) - zone1offset_y;
+             else
+                 scaledvalue = ((n * scaleopp2) >> 8) + zone1offset_y;
+         }
+     }
+     if (v->cur_field_type && !v->ref_field_type[dir]) {
+         return av_clip(scaledvalue, -v->range_y / 2 + 1, v->range_y / 2);
+     } else {
+         return av_clip(scaledvalue, -v->range_y / 2, v->range_y / 2 - 1);
+     }
+ }
+ static av_always_inline int scaleforsame(VC1Context *v, int i, int n /* MV */,
+                                          int dim, int dir)
+ {
+     int brfd, scalesame;
+     int hpel = 1 - v->s.quarter_sample;
+     n >>= hpel;
+     if (v->s.pict_type != AV_PICTURE_TYPE_B || v->second_field || !dir) {
+         if (dim)
+             n = scaleforsame_y(v, i, n, dir) << hpel;
+         else
+             n = scaleforsame_x(v, n, dir) << hpel;
+         return n;
+     }
+     brfd      = FFMIN(v->brfd, 3);
+     scalesame = ff_vc1_b_field_mvpred_scales[0][brfd];
+     n = (n * scalesame >> 8) << hpel;
+     return n;
+ }
+ static av_always_inline int scaleforopp(VC1Context *v, int n /* MV */,
+                                         int dim, int dir)
+ {
+     int refdist, scaleopp;
+     int hpel = 1 - v->s.quarter_sample;
+     n >>= hpel;
+     if (v->s.pict_type == AV_PICTURE_TYPE_B && !v->second_field && dir == 1) {
+         if (dim)
+             n = scaleforopp_y(v, n, dir) << hpel;
+         else
+             n = scaleforopp_x(v, n) << hpel;
+         return n;
+     }
+     if (v->s.pict_type != AV_PICTURE_TYPE_B)
+         refdist = FFMIN(v->refdist, 3);
+     else
+         refdist = dir ? v->brfd : v->frfd;
+     scaleopp = ff_vc1_field_mvpred_scales[dir ^ v->second_field][0][refdist];
+     n = (n * scaleopp >> 8) << hpel;
+     return n;
+ }
+ /** Predict and set motion vector
+  */
+ void ff_vc1_pred_mv(VC1Context *v, int n, int dmv_x, int dmv_y,
+                     int mv1, int r_x, int r_y, uint8_t* is_intra,
+                     int pred_flag, int dir)
+ {
+     MpegEncContext *s = &v->s;
+     int xy, wrap, off = 0;
+     int16_t *A, *B, *C;
+     int px, py;
+     int sum;
+     int mixedmv_pic, num_samefield = 0, num_oppfield = 0;
+     int opposite, a_f, b_f, c_f;
+     int16_t field_predA[2];
+     int16_t field_predB[2];
+     int16_t field_predC[2];
+     int a_valid, b_valid, c_valid;
+     int hybridmv_thresh, y_bias = 0;
+     if (v->mv_mode == MV_PMODE_MIXED_MV ||
+         ((v->mv_mode == MV_PMODE_INTENSITY_COMP) && (v->mv_mode2 == MV_PMODE_MIXED_MV)))
+         mixedmv_pic = 1;
+     else
+         mixedmv_pic = 0;
+     /* scale MV difference to be quad-pel */
+     dmv_x <<= 1 - s->quarter_sample;
+     dmv_y <<= 1 - s->quarter_sample;
+     wrap = s->b8_stride;
+     xy   = s->block_index[n];
+     if (s->mb_intra) {
+         s->mv[0][n][0] = s->current_picture.motion_val[0][xy + v->blocks_off][0] = 0;
+         s->mv[0][n][1] = s->current_picture.motion_val[0][xy + v->blocks_off][1] = 0;
+         s->current_picture.motion_val[1][xy + v->blocks_off][0] = 0;
+         s->current_picture.motion_val[1][xy + v->blocks_off][1] = 0;
+         if (mv1) { /* duplicate motion data for 1-MV block */
+             s->current_picture.motion_val[0][xy + 1 + v->blocks_off][0]        = 0;
+             s->current_picture.motion_val[0][xy + 1 + v->blocks_off][1]        = 0;
+             s->current_picture.motion_val[0][xy + wrap + v->blocks_off][0]     = 0;
+             s->current_picture.motion_val[0][xy + wrap + v->blocks_off][1]     = 0;
+             s->current_picture.motion_val[0][xy + wrap + 1 + v->blocks_off][0] = 0;
+             s->current_picture.motion_val[0][xy + wrap + 1 + v->blocks_off][1] = 0;
+             v->luma_mv[s->mb_x][0] = v->luma_mv[s->mb_x][1] = 0;
+             s->current_picture.motion_val[1][xy + 1 + v->blocks_off][0]        = 0;
+             s->current_picture.motion_val[1][xy + 1 + v->blocks_off][1]        = 0;
+             s->current_picture.motion_val[1][xy + wrap][0]                     = 0;
+             s->current_picture.motion_val[1][xy + wrap + v->blocks_off][1]     = 0;
+             s->current_picture.motion_val[1][xy + wrap + 1 + v->blocks_off][0] = 0;
+             s->current_picture.motion_val[1][xy + wrap + 1 + v->blocks_off][1] = 0;
+         }
+         return;
+     }
+     C = s->current_picture.motion_val[dir][xy -    1 + v->blocks_off];
+     A = s->current_picture.motion_val[dir][xy - wrap + v->blocks_off];
+     if (mv1) {
+         if (v->field_mode && mixedmv_pic)
+             off = (s->mb_x == (s->mb_width - 1)) ? -2 : 2;
+         else
+             off = (s->mb_x == (s->mb_width - 1)) ? -1 : 2;
+     } else {
+         //in 4-MV mode different blocks have different B predictor position
+         switch (n) {
+         case 0:
+             off = (s->mb_x > 0) ? -1 : 1;
+             break;
+         case 1:
+             off = (s->mb_x == (s->mb_width - 1)) ? -1 : 1;
+             break;
+         case 2:
+             off = 1;
+             break;
+         case 3:
+             off = -1;
+         }
+     }
+     B = s->current_picture.motion_val[dir][xy - wrap + off + v->blocks_off];
+     a_valid = !s->first_slice_line || (n == 2 || n == 3);
+     b_valid = a_valid && (s->mb_width > 1);
+     c_valid = s->mb_x || (n == 1 || n == 3);
+     if (v->field_mode) {
+         a_valid = a_valid && !is_intra[xy - wrap];
+         b_valid = b_valid && !is_intra[xy - wrap + off];
+         c_valid = c_valid && !is_intra[xy - 1];
+     }
+     if (a_valid) {
+         a_f = v->mv_f[dir][xy - wrap + v->blocks_off];
+         num_oppfield  += a_f;
+         num_samefield += 1 - a_f;
+         field_predA[0] = A[0];
+         field_predA[1] = A[1];
+     } else {
+         field_predA[0] = field_predA[1] = 0;
+         a_f = 0;
+     }
+     if (b_valid) {
+         b_f = v->mv_f[dir][xy - wrap + off + v->blocks_off];
+         num_oppfield  += b_f;
+         num_samefield += 1 - b_f;
+         field_predB[0] = B[0];
+         field_predB[1] = B[1];
+     } else {
+         field_predB[0] = field_predB[1] = 0;
+         b_f = 0;
+     }
+     if (c_valid) {
+         c_f = v->mv_f[dir][xy - 1 + v->blocks_off];
+         num_oppfield  += c_f;
+         num_samefield += 1 - c_f;
+         field_predC[0] = C[0];
+         field_predC[1] = C[1];
+     } else {
+         field_predC[0] = field_predC[1] = 0;
+         c_f = 0;
+     }
+     if (v->field_mode) {
+         if (!v->numref)
+             // REFFIELD determines if the last field or the second-last field is
+             // to be used as reference
+             opposite = 1 - v->reffield;
+         else {
+             if (num_samefield <= num_oppfield)
+                 opposite = 1 - pred_flag;
+             else
+                 opposite = pred_flag;
+         }
+     } else
+         opposite = 0;
+     if (opposite) {
+         if (a_valid && !a_f) {
+             field_predA[0] = scaleforopp(v, field_predA[0], 0, dir);
+             field_predA[1] = scaleforopp(v, field_predA[1], 1, dir);
+         }
+         if (b_valid && !b_f) {
+             field_predB[0] = scaleforopp(v, field_predB[0], 0, dir);
+             field_predB[1] = scaleforopp(v, field_predB[1], 1, dir);
+         }
+         if (c_valid && !c_f) {
+             field_predC[0] = scaleforopp(v, field_predC[0], 0, dir);
+             field_predC[1] = scaleforopp(v, field_predC[1], 1, dir);
+         }
+         v->mv_f[dir][xy + v->blocks_off] = 1;
+         v->ref_field_type[dir] = !v->cur_field_type;
+     } else {
+         if (a_valid && a_f) {
+             field_predA[0] = scaleforsame(v, n, field_predA[0], 0, dir);
+             field_predA[1] = scaleforsame(v, n, field_predA[1], 1, dir);
+         }
+         if (b_valid && b_f) {
+             field_predB[0] = scaleforsame(v, n, field_predB[0], 0, dir);
+             field_predB[1] = scaleforsame(v, n, field_predB[1], 1, dir);
+         }
+         if (c_valid && c_f) {
+             field_predC[0] = scaleforsame(v, n, field_predC[0], 0, dir);
+             field_predC[1] = scaleforsame(v, n, field_predC[1], 1, dir);
+         }
+         v->mv_f[dir][xy + v->blocks_off] = 0;
+         v->ref_field_type[dir] = v->cur_field_type;
+     }
+     if (a_valid) {
+         px = field_predA[0];
+         py = field_predA[1];
+     } else if (c_valid) {
+         px = field_predC[0];
+         py = field_predC[1];
+     } else if (b_valid) {
+         px = field_predB[0];
+         py = field_predB[1];
+     } else {
+         px = 0;
+         py = 0;
+     }
+     if (num_samefield + num_oppfield > 1) {
+         px = mid_pred(field_predA[0], field_predB[0], field_predC[0]);
+         py = mid_pred(field_predA[1], field_predB[1], field_predC[1]);
+     }
+     /* Pullback MV as specified in 8.3.5.3.4 */
+     if (!v->field_mode) {
+         int qx, qy, X, Y;
+         qx = (s->mb_x << 6) + ((n == 1 || n == 3) ? 32 : 0);
+         qy = (s->mb_y << 6) + ((n == 2 || n == 3) ? 32 : 0);
+         X  = (s->mb_width  << 6) - 4;
+         Y  = (s->mb_height << 6) - 4;
+         if (mv1) {
+             if (qx + px < -60) px = -60 - qx;
+             if (qy + py < -60) py = -60 - qy;
+         } else {
+             if (qx + px < -28) px = -28 - qx;
+             if (qy + py < -28) py = -28 - qy;
+         }
+         if (qx + px > X) px = X - qx;
+         if (qy + py > Y) py = Y - qy;
+     }
+     if (!v->field_mode || s->pict_type != AV_PICTURE_TYPE_B) {
+         /* Calculate hybrid prediction as specified in 8.3.5.3.5 (also 10.3.5.4.3.5) */
+         hybridmv_thresh = 32;
+         if (a_valid && c_valid) {
+             if (is_intra[xy - wrap])
+                 sum = FFABS(px) + FFABS(py);
+             else
+                 sum = FFABS(px - field_predA[0]) + FFABS(py - field_predA[1]);
+             if (sum > hybridmv_thresh) {
+                 if (get_bits1(&s->gb)) {     // read HYBRIDPRED bit
+                     px = field_predA[0];
+                     py = field_predA[1];
+                 } else {
+                     px = field_predC[0];
+                     py = field_predC[1];
+                 }
+             } else {
+                 if (is_intra[xy - 1])
+                     sum = FFABS(px) + FFABS(py);
+                 else
+                     sum = FFABS(px - field_predC[0]) + FFABS(py - field_predC[1]);
+                 if (sum > hybridmv_thresh) {
+                     if (get_bits1(&s->gb)) {
+                         px = field_predA[0];
+                         py = field_predA[1];
+                     } else {
+                         px = field_predC[0];
+                         py = field_predC[1];
+                     }
+                 }
+             }
+         }
+     }
+     if (v->field_mode && v->numref)
+         r_y >>= 1;
+     if (v->field_mode && v->cur_field_type && v->ref_field_type[dir] == 0)
+         y_bias = 1;
+     /* store MV using signed modulus of MV range defined in 4.11 */
+     s->mv[dir][n][0] = s->current_picture.motion_val[dir][xy + v->blocks_off][0] = ((px + dmv_x + r_x) & ((r_x << 1) - 1)) - r_x;
+     s->mv[dir][n][1] = s->current_picture.motion_val[dir][xy + v->blocks_off][1] = ((py + dmv_y + r_y - y_bias) & ((r_y << 1) - 1)) - r_y + y_bias;
+     if (mv1) { /* duplicate motion data for 1-MV block */
+         s->current_picture.motion_val[dir][xy +    1 +     v->blocks_off][0] = s->current_picture.motion_val[dir][xy + v->blocks_off][0];
+         s->current_picture.motion_val[dir][xy +    1 +     v->blocks_off][1] = s->current_picture.motion_val[dir][xy + v->blocks_off][1];
+         s->current_picture.motion_val[dir][xy + wrap +     v->blocks_off][0] = s->current_picture.motion_val[dir][xy + v->blocks_off][0];
+         s->current_picture.motion_val[dir][xy + wrap +     v->blocks_off][1] = s->current_picture.motion_val[dir][xy + v->blocks_off][1];
+         s->current_picture.motion_val[dir][xy + wrap + 1 + v->blocks_off][0] = s->current_picture.motion_val[dir][xy + v->blocks_off][0];
+         s->current_picture.motion_val[dir][xy + wrap + 1 + v->blocks_off][1] = s->current_picture.motion_val[dir][xy + v->blocks_off][1];
+         v->mv_f[dir][xy +    1 + v->blocks_off] = v->mv_f[dir][xy +            v->blocks_off];
+         v->mv_f[dir][xy + wrap + v->blocks_off] = v->mv_f[dir][xy + wrap + 1 + v->blocks_off] = v->mv_f[dir][xy + v->blocks_off];
+     }
+ }
+ /** Predict and set motion vector for interlaced frame picture MBs
+  */
+ void ff_vc1_pred_mv_intfr(VC1Context *v, int n, int dmv_x, int dmv_y,
+                           int mvn, int r_x, int r_y, uint8_t* is_intra, int dir)
+ {
+     MpegEncContext *s = &v->s;
+     int xy, wrap, off = 0;
+     int A[2], B[2], C[2];
 -                if (a_valid) { px = A[0]; py = A[1]; }
 -                if (b_valid) { px = B[0]; py = B[1]; }
 -                if (c_valid) { px = C[0]; py = C[1]; }
 -            } else
 -                px = py = 0;
++    int px = 0, py = 0;
+     int a_valid = 0, b_valid = 0, c_valid = 0;
+     int field_a, field_b, field_c; // 0: same, 1: opposit
+     int total_valid, num_samefield, num_oppfield;
+     int pos_c, pos_b, n_adj;
+     wrap = s->b8_stride;
+     xy = s->block_index[n];
+     if (s->mb_intra) {
+         s->mv[0][n][0] = s->current_picture.motion_val[0][xy][0] = 0;
+         s->mv[0][n][1] = s->current_picture.motion_val[0][xy][1] = 0;
+         s->current_picture.motion_val[1][xy][0] = 0;
+         s->current_picture.motion_val[1][xy][1] = 0;
+         if (mvn == 1) { /* duplicate motion data for 1-MV block */
+             s->current_picture.motion_val[0][xy + 1][0]        = 0;
+             s->current_picture.motion_val[0][xy + 1][1]        = 0;
+             s->current_picture.motion_val[0][xy + wrap][0]     = 0;
+             s->current_picture.motion_val[0][xy + wrap][1]     = 0;
+             s->current_picture.motion_val[0][xy + wrap + 1][0] = 0;
+             s->current_picture.motion_val[0][xy + wrap + 1][1] = 0;
+             v->luma_mv[s->mb_x][0] = v->luma_mv[s->mb_x][1] = 0;
+             s->current_picture.motion_val[1][xy + 1][0]        = 0;
+             s->current_picture.motion_val[1][xy + 1][1]        = 0;
+             s->current_picture.motion_val[1][xy + wrap][0]     = 0;
+             s->current_picture.motion_val[1][xy + wrap][1]     = 0;
+             s->current_picture.motion_val[1][xy + wrap + 1][0] = 0;
+             s->current_picture.motion_val[1][xy + wrap + 1][1] = 0;
+         }
+         return;
+     }
+     off = ((n == 0) || (n == 1)) ? 1 : -1;
+     /* predict A */
+     if (s->mb_x || (n == 1) || (n == 3)) {
+         if ((v->blk_mv_type[xy]) // current block (MB) has a field MV
+             || (!v->blk_mv_type[xy] && !v->blk_mv_type[xy - 1])) { // or both have frame MV
+             A[0] = s->current_picture.motion_val[dir][xy - 1][0];
+             A[1] = s->current_picture.motion_val[dir][xy - 1][1];
+             a_valid = 1;
+         } else { // current block has frame mv and cand. has field MV (so average)
+             A[0] = (s->current_picture.motion_val[dir][xy - 1][0]
+                     + s->current_picture.motion_val[dir][xy - 1 + off * wrap][0] + 1) >> 1;
+             A[1] = (s->current_picture.motion_val[dir][xy - 1][1]
+                     + s->current_picture.motion_val[dir][xy - 1 + off * wrap][1] + 1) >> 1;
+             a_valid = 1;
+         }
+         if (!(n & 1) && v->is_intra[s->mb_x - 1]) {
+             a_valid = 0;
+             A[0] = A[1] = 0;
+         }
+     } else
+         A[0] = A[1] = 0;
+     /* Predict B and C */
+     B[0] = B[1] = C[0] = C[1] = 0;
+     if (n == 0 || n == 1 || v->blk_mv_type[xy]) {
+         if (!s->first_slice_line) {
+             if (!v->is_intra[s->mb_x - s->mb_stride]) {
+                 b_valid = 1;
+                 n_adj   = n | 2;
+                 pos_b   = s->block_index[n_adj] - 2 * wrap;
+                 if (v->blk_mv_type[pos_b] && v->blk_mv_type[xy]) {
+                     n_adj = (n & 2) | (n & 1);
+                 }
+                 B[0] = s->current_picture.motion_val[dir][s->block_index[n_adj] - 2 * wrap][0];
+                 B[1] = s->current_picture.motion_val[dir][s->block_index[n_adj] - 2 * wrap][1];
+                 if (v->blk_mv_type[pos_b] && !v->blk_mv_type[xy]) {
+                     B[0] = (B[0] + s->current_picture.motion_val[dir][s->block_index[n_adj ^ 2] - 2 * wrap][0] + 1) >> 1;
+                     B[1] = (B[1] + s->current_picture.motion_val[dir][s->block_index[n_adj ^ 2] - 2 * wrap][1] + 1) >> 1;
+                 }
+             }
+             if (s->mb_width > 1) {
+                 if (!v->is_intra[s->mb_x - s->mb_stride + 1]) {
+                     c_valid = 1;
+                     n_adj   = 2;
+                     pos_c   = s->block_index[2] - 2 * wrap + 2;
+                     if (v->blk_mv_type[pos_c] && v->blk_mv_type[xy]) {
+                         n_adj = n & 2;
+                     }
+                     C[0] = s->current_picture.motion_val[dir][s->block_index[n_adj] - 2 * wrap + 2][0];
+                     C[1] = s->current_picture.motion_val[dir][s->block_index[n_adj] - 2 * wrap + 2][1];
+                     if (v->blk_mv_type[pos_c] && !v->blk_mv_type[xy]) {
+                         C[0] = (1 + C[0] + (s->current_picture.motion_val[dir][s->block_index[n_adj ^ 2] - 2 * wrap + 2][0])) >> 1;
+                         C[1] = (1 + C[1] + (s->current_picture.motion_val[dir][s->block_index[n_adj ^ 2] - 2 * wrap + 2][1])) >> 1;
+                     }
+                     if (s->mb_x == s->mb_width - 1) {
+                         if (!v->is_intra[s->mb_x - s->mb_stride - 1]) {
+                             c_valid = 1;
+                             n_adj   = 3;
+                             pos_c   = s->block_index[3] - 2 * wrap - 2;
+                             if (v->blk_mv_type[pos_c] && v->blk_mv_type[xy]) {
+                                 n_adj = n | 1;
+                             }
+                             C[0] = s->current_picture.motion_val[dir][s->block_index[n_adj] - 2 * wrap - 2][0];
+                             C[1] = s->current_picture.motion_val[dir][s->block_index[n_adj] - 2 * wrap - 2][1];
+                             if (v->blk_mv_type[pos_c] && !v->blk_mv_type[xy]) {
+                                 C[0] = (1 + C[0] + s->current_picture.motion_val[dir][s->block_index[1] - 2 * wrap - 2][0]) >> 1;
+                                 C[1] = (1 + C[1] + s->current_picture.motion_val[dir][s->block_index[1] - 2 * wrap - 2][1]) >> 1;
+                             }
+                         } else
+                             c_valid = 0;
+                     }
+                 }
+             }
+         }
+     } else {
+         pos_b   = s->block_index[1];
+         b_valid = 1;
+         B[0]    = s->current_picture.motion_val[dir][pos_b][0];
+         B[1]    = s->current_picture.motion_val[dir][pos_b][1];
+         pos_c   = s->block_index[0];
+         c_valid = 1;
+         C[0]    = s->current_picture.motion_val[dir][pos_c][0];
+         C[1]    = s->current_picture.motion_val[dir][pos_c][1];
+     }
+     total_valid = a_valid + b_valid + c_valid;
+     // check if predictor A is out of bounds
+     if (!s->mb_x && !(n == 1 || n == 3)) {
+         A[0] = A[1] = 0;
+     }
+     // check if predictor B is out of bounds
+     if ((s->first_slice_line && v->blk_mv_type[xy]) || (s->first_slice_line && !(n & 2))) {
+         B[0] = B[1] = C[0] = C[1] = 0;
+     }
+     if (!v->blk_mv_type[xy]) {
+         if (s->mb_width == 1) {
+             px = B[0];
+             py = B[1];
+         } else {
+             if (total_valid >= 2) {
+                 px = mid_pred(A[0], B[0], C[0]);
+                 py = mid_pred(A[1], B[1], C[1]);
+             } else if (total_valid) {
 -                } else if (c_valid) {
++                if      (a_valid) { px = A[0]; py = A[1]; }
++                else if (b_valid) { px = B[0]; py = B[1]; }
++                else              { px = C[0]; py = C[1]; }
++            }
+         }
+     } else {
+         if (a_valid)
+             field_a = (A[1] & 4) ? 1 : 0;
+         else
+             field_a = 0;
+         if (b_valid)
+             field_b = (B[1] & 4) ? 1 : 0;
+         else
+             field_b = 0;
+         if (c_valid)
+             field_c = (C[1] & 4) ? 1 : 0;
+         else
+             field_c = 0;
+         num_oppfield  = field_a + field_b + field_c;
+         num_samefield = total_valid - num_oppfield;
+         if (total_valid == 3) {
+             if ((num_samefield == 3) || (num_oppfield == 3)) {
+                 px = mid_pred(A[0], B[0], C[0]);
+                 py = mid_pred(A[1], B[1], C[1]);
+             } else if (num_samefield >= num_oppfield) {
+                 /* take one MV from same field set depending on priority
+                 the check for B may not be necessary */
+                 px = !field_a ? A[0] : B[0];
+                 py = !field_a ? A[1] : B[1];
+             } else {
+                 px =  field_a ? A[0] : B[0];
+                 py =  field_a ? A[1] : B[1];
+             }
+         } else if (total_valid == 2) {
+             if (num_samefield >= num_oppfield) {
+                 if (!field_a && a_valid) {
+                     px = A[0];
+                     py = A[1];
+                 } else if (!field_b && b_valid) {
+                     px = B[0];
+                     py = B[1];
 -                } else px = py = 0;
++                } else /*if (c_valid)*/ {
++                    av_assert1(c_valid);
+                     px = C[0];
+                     py = C[1];
 -                } else if (field_b && b_valid) {
++                } /*else px = py = 0;*/
+             } else {
+                 if (field_a && a_valid) {
+                     px = A[0];
+                     py = A[1];
 -                } else if (c_valid) {
++                } else /*if (field_b && b_valid)*/ {
++                    av_assert1(field_b && b_valid);
+                     px = B[0];
+                     py = B[1];
 -                } else
 -                    px = py = 0;
++                } /*else if (c_valid) {
+                     px = C[0];
+                     py = C[1];
 -        } else
 -            px = py = 0;
++                }*/
+             }
+         } else if (total_valid == 1) {
+             px = (a_valid) ? A[0] : ((b_valid) ? B[0] : C[0]);
+             py = (a_valid) ? A[1] : ((b_valid) ? B[1] : C[1]);
 -        s->current_picture.motion_val[0][xy + v->blocks_off][0] =
 -        s->current_picture.motion_val[0][xy + v->blocks_off][1] =
 -        s->current_picture.motion_val[1][xy + v->blocks_off][0] =
 -        s->current_picture.motion_val[1][xy + v->blocks_off][1] = 0;
++        }
+     }
+     /* store MV using signed modulus of MV range defined in 4.11 */
+     s->mv[dir][n][0] = s->current_picture.motion_val[dir][xy][0] = ((px + dmv_x + r_x) & ((r_x << 1) - 1)) - r_x;
+     s->mv[dir][n][1] = s->current_picture.motion_val[dir][xy][1] = ((py + dmv_y + r_y) & ((r_y << 1) - 1)) - r_y;
+     if (mvn == 1) { /* duplicate motion data for 1-MV block */
+         s->current_picture.motion_val[dir][xy +    1    ][0] = s->current_picture.motion_val[dir][xy][0];
+         s->current_picture.motion_val[dir][xy +    1    ][1] = s->current_picture.motion_val[dir][xy][1];
+         s->current_picture.motion_val[dir][xy + wrap    ][0] = s->current_picture.motion_val[dir][xy][0];
+         s->current_picture.motion_val[dir][xy + wrap    ][1] = s->current_picture.motion_val[dir][xy][1];
+         s->current_picture.motion_val[dir][xy + wrap + 1][0] = s->current_picture.motion_val[dir][xy][0];
+         s->current_picture.motion_val[dir][xy + wrap + 1][1] = s->current_picture.motion_val[dir][xy][1];
+     } else if (mvn == 2) { /* duplicate motion data for 2-Field MV block */
+         s->current_picture.motion_val[dir][xy + 1][0] = s->current_picture.motion_val[dir][xy][0];
+         s->current_picture.motion_val[dir][xy + 1][1] = s->current_picture.motion_val[dir][xy][1];
+         s->mv[dir][n + 1][0] = s->mv[dir][n][0];
+         s->mv[dir][n + 1][1] = s->mv[dir][n][1];
+     }
+ }
+ void ff_vc1_pred_b_mv(VC1Context *v, int dmv_x[2], int dmv_y[2],
+                       int direct, int mvtype)
+ {
+     MpegEncContext *s = &v->s;
+     int xy, wrap, off = 0;
+     int16_t *A, *B, *C;
+     int px, py;
+     int sum;
+     int r_x, r_y;
+     const uint8_t *is_intra = v->mb_type[0];
++    av_assert0(!v->field_mode);
++
+     r_x = v->range_x;
+     r_y = v->range_y;
+     /* scale MV difference to be quad-pel */
+     dmv_x[0] <<= 1 - s->quarter_sample;
+     dmv_y[0] <<= 1 - s->quarter_sample;
+     dmv_x[1] <<= 1 - s->quarter_sample;
+     dmv_y[1] <<= 1 - s->quarter_sample;
+     wrap = s->b8_stride;
+     xy = s->block_index[0];
+     if (s->mb_intra) {
 -    if (!v->field_mode) {
++        s->current_picture.motion_val[0][xy][0] =
++        s->current_picture.motion_val[0][xy][1] =
++        s->current_picture.motion_val[1][xy][0] =
++        s->current_picture.motion_val[1][xy][1] = 0;
+         return;
+     }
 -    }
++        if (direct && s->next_picture_ptr->field_picture)
++            av_log(s->avctx, AV_LOG_WARNING, "Mixed frame/field direct mode not supported\n");
++
+         s->mv[0][0][0] = scale_mv(s->next_picture.motion_val[1][xy][0], v->bfraction, 0, s->quarter_sample);
+         s->mv[0][0][1] = scale_mv(s->next_picture.motion_val[1][xy][1], v->bfraction, 0, s->quarter_sample);
+         s->mv[1][0][0] = scale_mv(s->next_picture.motion_val[1][xy][0], v->bfraction, 1, s->quarter_sample);
+         s->mv[1][0][1] = scale_mv(s->next_picture.motion_val[1][xy][1], v->bfraction, 1, s->quarter_sample);
+         /* Pullback predicted motion vectors as specified in 8.4.5.4 */
+         s->mv[0][0][0] = av_clip(s->mv[0][0][0], -60 - (s->mb_x << 6), (s->mb_width  << 6) - 4 - (s->mb_x << 6));
+         s->mv[0][0][1] = av_clip(s->mv[0][0][1], -60 - (s->mb_y << 6), (s->mb_height << 6) - 4 - (s->mb_y << 6));
+         s->mv[1][0][0] = av_clip(s->mv[1][0][0], -60 - (s->mb_x << 6), (s->mb_width  << 6) - 4 - (s->mb_x << 6));
+         s->mv[1][0][1] = av_clip(s->mv[1][0][1], -60 - (s->mb_y << 6), (s->mb_height << 6) - 4 - (s->mb_y << 6));
 -        s->current_picture.motion_val[0][xy + v->blocks_off][0] = s->mv[0][0][0];
 -        s->current_picture.motion_val[0][xy + v->blocks_off][1] = s->mv[0][0][1];
 -        s->current_picture.motion_val[1][xy + v->blocks_off][0] = s->mv[1][0][0];
 -        s->current_picture.motion_val[1][xy + v->blocks_off][1] = s->mv[1][0][1];
+     if (direct) {
++        s->current_picture.motion_val[0][xy][0] = s->mv[0][0][0];
++        s->current_picture.motion_val[0][xy][1] = s->mv[0][0][1];
++        s->current_picture.motion_val[1][xy][0] = s->mv[1][0][0];
++        s->current_picture.motion_val[1][xy][1] = s->mv[1][0][1];
+         return;
+     }
+     if ((mvtype == BMV_TYPE_FORWARD) || (mvtype == BMV_TYPE_INTERPOLATED)) {
+         C   = s->current_picture.motion_val[0][xy - 2];
+         A   = s->current_picture.motion_val[0][xy - wrap * 2];
+         off = (s->mb_x == (s->mb_width - 1)) ? -2 : 2;
+         B   = s->current_picture.motion_val[0][xy - wrap * 2 + off];
+         if (!s->mb_x) C[0] = C[1] = 0;
+         if (!s->first_slice_line) { // predictor A is not out of bounds
+             if (s->mb_width == 1) {
+                 px = A[0];
+                 py = A[1];
+             } else {
+                 px = mid_pred(A[0], B[0], C[0]);
+                 py = mid_pred(A[1], B[1], C[1]);
+             }
+         } else if (s->mb_x) { // predictor C is not out of bounds
+             px = C[0];
+             py = C[1];
+         } else {
+             px = py = 0;
+         }
+         /* Pullback MV as specified in 8.3.5.3.4 */
+         {
+             int qx, qy, X, Y;
+             if (v->profile < PROFILE_ADVANCED) {
+                 qx = (s->mb_x << 5);
+                 qy = (s->mb_y << 5);
+                 X  = (s->mb_width  << 5) - 4;
+                 Y  = (s->mb_height << 5) - 4;
+                 if (qx + px < -28) px = -28 - qx;
+                 if (qy + py < -28) py = -28 - qy;
+                 if (qx + px > X) px = X - qx;
+                 if (qy + py > Y) py = Y - qy;
+             } else {
+                 qx = (s->mb_x << 6);
+                 qy = (s->mb_y << 6);
+                 X  = (s->mb_width  << 6) - 4;
+                 Y  = (s->mb_height << 6) - 4;
+                 if (qx + px < -60) px = -60 - qx;
+                 if (qy + py < -60) py = -60 - qy;
+                 if (qx + px > X) px = X - qx;
+                 if (qy + py > Y) py = Y - qy;
+             }
+         }
+         /* Calculate hybrid prediction as specified in 8.3.5.3.5 */
+         if (0 && !s->first_slice_line && s->mb_x) {
+             if (is_intra[xy - wrap])
+                 sum = FFABS(px) + FFABS(py);
+             else
+                 sum = FFABS(px - A[0]) + FFABS(py - A[1]);
+             if (sum > 32) {
+                 if (get_bits1(&s->gb)) {
+                     px = A[0];
+                     py = A[1];
+                 } else {
+                     px = C[0];
+                     py = C[1];
+                 }
+             } else {
+                 if (is_intra[xy - 2])
+                     sum = FFABS(px) + FFABS(py);
+                 else
+                     sum = FFABS(px - C[0]) + FFABS(py - C[1]);
+                 if (sum > 32) {
+                     if (get_bits1(&s->gb)) {
+                         px = A[0];
+                         py = A[1];
+                     } else {
+                         px = C[0];
+                         py = C[1];
+                     }
+                 }
+             }
+         }
+         /* store MV using signed modulus of MV range defined in 4.11 */
+         s->mv[0][0][0] = ((px + dmv_x[0] + r_x) & ((r_x << 1) - 1)) - r_x;
+         s->mv[0][0][1] = ((py + dmv_y[0] + r_y) & ((r_y << 1) - 1)) - r_y;
+     }
+     if ((mvtype == BMV_TYPE_BACKWARD) || (mvtype == BMV_TYPE_INTERPOLATED)) {
+         C   = s->current_picture.motion_val[1][xy - 2];
+         A   = s->current_picture.motion_val[1][xy - wrap * 2];
+         off = (s->mb_x == (s->mb_width - 1)) ? -2 : 2;
+         B   = s->current_picture.motion_val[1][xy - wrap * 2 + off];
+         if (!s->mb_x)
+             C[0] = C[1] = 0;
+         if (!s->first_slice_line) { // predictor A is not out of bounds
+             if (s->mb_width == 1) {
+                 px = A[0];
+                 py = A[1];
+             } else {
+                 px = mid_pred(A[0], B[0], C[0]);
+                 py = mid_pred(A[1], B[1], C[1]);
+             }
+         } else if (s->mb_x) { // predictor C is not out of bounds
+             px = C[0];
+             py = C[1];
+         } else {
+             px = py = 0;
+         }
+         /* Pullback MV as specified in 8.3.5.3.4 */
+         {
+             int qx, qy, X, Y;
+             if (v->profile < PROFILE_ADVANCED) {
+                 qx = (s->mb_x << 5);
+                 qy = (s->mb_y << 5);
+                 X  = (s->mb_width  << 5) - 4;
+                 Y  = (s->mb_height << 5) - 4;
+                 if (qx + px < -28) px = -28 - qx;
+                 if (qy + py < -28) py = -28 - qy;
+                 if (qx + px > X) px = X - qx;
+                 if (qy + py > Y) py = Y - qy;
+             } else {
+                 qx = (s->mb_x << 6);
+                 qy = (s->mb_y << 6);
+                 X  = (s->mb_width  << 6) - 4;
+                 Y  = (s->mb_height << 6) - 4;
+                 if (qx + px < -60) px = -60 - qx;
+                 if (qy + py < -60) py = -60 - qy;
+                 if (qx + px > X) px = X - qx;
+                 if (qy + py > Y) py = Y - qy;
+             }
+         }
+         /* Calculate hybrid prediction as specified in 8.3.5.3.5 */
+         if (0 && !s->first_slice_line && s->mb_x) {
+             if (is_intra[xy - wrap])
+                 sum = FFABS(px) + FFABS(py);
+             else
+                 sum = FFABS(px - A[0]) + FFABS(py - A[1]);
+             if (sum > 32) {
+                 if (get_bits1(&s->gb)) {
+                     px = A[0];
+                     py = A[1];
+                 } else {
+                     px = C[0];
+                     py = C[1];
+                 }
+             } else {
+                 if (is_intra[xy - 2])
+                     sum = FFABS(px) + FFABS(py);
+                 else
+                     sum = FFABS(px - C[0]) + FFABS(py - C[1]);
+                 if (sum > 32) {
+                     if (get_bits1(&s->gb)) {
+                         px = A[0];
+                         py = A[1];
+                     } else {
+                         px = C[0];
+                         py = C[1];
+                     }
+                 }
+             }
+         }
+         /* store MV using signed modulus of MV range defined in 4.11 */
+         s->mv[1][0][0] = ((px + dmv_x[1] + r_x) & ((r_x << 1) - 1)) - r_x;
+         s->mv[1][0][1] = ((py + dmv_y[1] + r_y) & ((r_y << 1) - 1)) - r_y;
+     }
+     s->current_picture.motion_val[0][xy][0] = s->mv[0][0][0];
+     s->current_picture.motion_val[0][xy][1] = s->mv[0][0][1];
+     s->current_picture.motion_val[1][xy][0] = s->mv[1][0][0];
+     s->current_picture.motion_val[1][xy][1] = s->mv[1][0][1];
+ }
+ void ff_vc1_pred_b_mv_intfi(VC1Context *v, int n, int *dmv_x, int *dmv_y,
+                             int mv1, int *pred_flag)
+ {
+     int dir = (v->bmvtype == BMV_TYPE_BACKWARD) ? 1 : 0;
+     MpegEncContext *s = &v->s;
+     int mb_pos = s->mb_x + s->mb_y * s->mb_stride;
+     if (v->bmvtype == BMV_TYPE_DIRECT) {
+         int total_opp, k, f;
+         if (s->next_picture.mb_type[mb_pos + v->mb_off] != MB_TYPE_INTRA) {
+             s->mv[0][0][0] = scale_mv(s->next_picture.motion_val[1][s->block_index[0] + v->blocks_off][0],
+                                       v->bfraction, 0, s->quarter_sample);
+             s->mv[0][0][1] = scale_mv(s->next_picture.motion_val[1][s->block_index[0] + v->blocks_off][1],
+                                       v->bfraction, 0, s->quarter_sample);
+             s->mv[1][0][0] = scale_mv(s->next_picture.motion_val[1][s->block_index[0] + v->blocks_off][0],
+                                       v->bfraction, 1, s->quarter_sample);
+             s->mv[1][0][1] = scale_mv(s->next_picture.motion_val[1][s->block_index[0] + v->blocks_off][1],
+                                       v->bfraction, 1, s->quarter_sample);
+             total_opp = v->mv_f_next[0][s->block_index[0] + v->blocks_off]
+                       + v->mv_f_next[0][s->block_index[1] + v->blocks_off]
+                       + v->mv_f_next[0][s->block_index[2] + v->blocks_off]
+                       + v->mv_f_next[0][s->block_index[3] + v->blocks_off];
+             f = (total_opp > 2) ? 1 : 0;
+         } else {
+             s->mv[0][0][0] = s->mv[0][0][1] = 0;
+             s->mv[1][0][0] = s->mv[1][0][1] = 0;
+             f = 0;
+         }
+         v->ref_field_type[0] = v->ref_field_type[1] = v->cur_field_type ^ f;
+         for (k = 0; k < 4; k++) {
+             s->current_picture.motion_val[0][s->block_index[k] + v->blocks_off][0] = s->mv[0][0][0];
+             s->current_picture.motion_val[0][s->block_index[k] + v->blocks_off][1] = s->mv[0][0][1];
+             s->current_picture.motion_val[1][s->block_index[k] + v->blocks_off][0] = s->mv[1][0][0];
+             s->current_picture.motion_val[1][s->block_index[k] + v->blocks_off][1] = s->mv[1][0][1];
+             v->mv_f[0][s->block_index[k] + v->blocks_off] = f;
+             v->mv_f[1][s->block_index[k] + v->blocks_off] = f;
+         }
+         return;
+     }
+     if (v->bmvtype == BMV_TYPE_INTERPOLATED) {
+         ff_vc1_pred_mv(v, 0, dmv_x[0], dmv_y[0],   1, v->range_x, v->range_y, v->mb_type[0], pred_flag[0], 0);
+         ff_vc1_pred_mv(v, 0, dmv_x[1], dmv_y[1],   1, v->range_x, v->range_y, v->mb_type[0], pred_flag[1], 1);
+         return;
+     }
+     if (dir) { // backward
+         ff_vc1_pred_mv(v, n, dmv_x[1], dmv_y[1], mv1, v->range_x, v->range_y, v->mb_type[0], pred_flag[1], 1);
+         if (n == 3 || mv1) {
+             ff_vc1_pred_mv(v, 0, dmv_x[0], dmv_y[0],   1, v->range_x, v->range_y, v->mb_type[0], 0, 0);
+         }
+     } else { // forward
+         ff_vc1_pred_mv(v, n, dmv_x[0], dmv_y[0], mv1, v->range_x, v->range_y, v->mb_type[0], pred_flag[0], 0);
+         if (n == 3 || mv1) {
+             ff_vc1_pred_mv(v, 0, dmv_x[1], dmv_y[1],   1, v->range_x, v->range_y, v->mb_type[0], 0, 1);
+         }
+     }
+ }
index 0000000,34c9c1a..4d47f86
mode 000000,100644..100644
--- /dev/null
@@@ -1,0 -1,59 +1,59 @@@
 - * This file is part of Libav.
+ /*
+  * VC-1 and WMV3 decoder
+  * Copyright (c) 2006-2007 Konstantin Shishkov
+  * Partly based on vc9.c (c) 2005 Anonymous, Alex Beregszaszi, Michael Niedermayer
+  *
 - * Libav is free software; you can redistribute it and/or
++ * This file is part of FFmpeg.
+  *
 - * Libav is distributed in the hope that it will be useful,
++ * FFmpeg is free software; you can redistribute it and/or
+  * modify it under the terms of the GNU Lesser General Public
+  * License as published by the Free Software Foundation; either
+  * version 2.1 of the License, or (at your option) any later version.
+  *
 - * License along with Libav; if not, write to the Free Software
++ * FFmpeg is distributed in the hope that it will be useful,
+  * but WITHOUT ANY WARRANTY; without even the implied warranty of
+  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+  * Lesser General Public License for more details.
+  *
+  * You should have received a copy of the GNU Lesser General Public
++ * License along with FFmpeg; if not, write to the Free Software
+  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+  */
+ #ifndef AVCODEC_VC1_PRED_H
+ #define AVCODEC_VC1_PRED_H
+ #include "vc1.h"
+ #include "vc1data.h"
+ void ff_vc1_pred_mv(VC1Context *v, int n, int dmv_x, int dmv_y,
+                     int mv1, int r_x, int r_y, uint8_t* is_intra,
+                     int pred_flag, int dir);
+ void ff_vc1_pred_mv_intfr(VC1Context *v, int n, int dmv_x, int dmv_y,
+                           int mvn, int r_x, int r_y, uint8_t* is_intra,
+                           int dir);
+ void ff_vc1_pred_b_mv(VC1Context *v, int dmv_x[2], int dmv_y[2],
+                       int direct, int mvtype);
+ void ff_vc1_pred_b_mv_intfi(VC1Context *v, int n, int *dmv_x, int *dmv_y,
+                             int mv1, int *pred_flag);
+ static av_always_inline int scale_mv(int value, int bfrac, int inv, int qs)
+ {
+     int n = bfrac;
+ #if B_FRACTION_DEN==256
+     if (inv)
+         n -= 256;
+     if (!qs)
+         return 2 * ((value * n + 255) >> 9);
+     return (value * n + 128) >> 8;
+ #else
+     if (inv)
+         n -= B_FRACTION_DEN;
+     if (!qs)
+         return 2 * ((value * n + B_FRACTION_DEN - 1) / (2 * B_FRACTION_DEN));
+     return (value * n + B_FRACTION_DEN/2) / B_FRACTION_DEN;
+ #endif
+ }
+ #endif /* AVCODEC_VC1_PRED_H */
Simple merge
   * VC-1 and WMV3 decoder
   */
  
- #include "internal.h"
  #include "avcodec.h"
- #include "error_resilience.h"
+ #include "blockdsp.h"
+ #include "get_bits.h"
+ #include "internal.h"
  #include "mpeg_er.h"
- #include "mpegutils.h"
  #include "mpegvideo.h"
- #include "h263.h"
- #include "h264chroma.h"
- #include "qpeldsp.h"
+ #include "msmpeg4data.h"
  #include "vc1.h"
  #include "vc1data.h"
- #include "vc1acdata.h"
- #include "msmpeg4data.h"
- #include "unary.h"
- #include "mathops.h"
 +#include "vdpau_internal.h"
 +#include "libavutil/avassert.h"
 +
- #undef NDEBUG
- #include <assert.h>
- #define MB_INTRA_VLC_BITS 9
- #define DC_VLC_BITS 9
- // offset tables for interlaced picture MVDATA decoding
- static const int offset_table1[9] = {  0,  1,  2,  4,  8, 16, 32,  64, 128 };
- static const int offset_table2[9] = {  0,  1,  3,  7, 15, 31, 63, 127, 255 };
- /***********************************************************************/
- /**
-  * @name VC-1 Bitplane decoding
-  * @see 8.7, p56
-  * @{
-  */
- static void init_block_index(VC1Context *v)
- {
-     MpegEncContext *s = &v->s;
-     ff_init_block_index(s);
-     if (v->field_mode && !(v->second_field ^ v->tff)) {
-         s->dest[0] += s->current_picture_ptr->f->linesize[0];
-         s->dest[1] += s->current_picture_ptr->f->linesize[1];
-         s->dest[2] += s->current_picture_ptr->f->linesize[2];
-     }
- }
- /** @} */ //Bitplane group
- static void vc1_put_signed_blocks_clamped(VC1Context *v)
- {
-     MpegEncContext *s = &v->s;
-     int topleft_mb_pos, top_mb_pos;
-     int stride_y, fieldtx = 0;
-     int v_dist;
-     /* The put pixels loop is always one MB row behind the decoding loop,
-      * because we can only put pixels when overlap filtering is done, and
-      * for filtering of the bottom edge of a MB, we need the next MB row
-      * present as well.
-      * Within the row, the put pixels loop is also one MB col behind the
-      * decoding loop. The reason for this is again, because for filtering
-      * of the right MB edge, we need the next MB present. */
-     if (!s->first_slice_line) {
-         if (s->mb_x) {
-             topleft_mb_pos = (s->mb_y - 1) * s->mb_stride + s->mb_x - 1;
-             if (v->fcm == ILACE_FRAME)
-                 fieldtx = v->fieldtx_plane[topleft_mb_pos];
-             stride_y       = s->linesize << fieldtx;
-             v_dist         = (16 - fieldtx) >> (fieldtx == 0);
-             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][0],
-                                               s->dest[0] - 16 * s->linesize - 16,
-                                               stride_y);
-             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][1],
-                                               s->dest[0] - 16 * s->linesize - 8,
-                                               stride_y);
-             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][2],
-                                               s->dest[0] - v_dist * s->linesize - 16,
-                                               stride_y);
-             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][3],
-                                               s->dest[0] - v_dist * s->linesize - 8,
-                                               stride_y);
-             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][4],
-                                               s->dest[1] - 8 * s->uvlinesize - 8,
-                                               s->uvlinesize);
-             s->idsp.put_signed_pixels_clamped(v->block[v->topleft_blk_idx][5],
-                                               s->dest[2] - 8 * s->uvlinesize - 8,
-                                               s->uvlinesize);
-         }
-         if (s->mb_x == s->mb_width - 1) {
-             top_mb_pos = (s->mb_y - 1) * s->mb_stride + s->mb_x;
-             if (v->fcm == ILACE_FRAME)
-                 fieldtx = v->fieldtx_plane[top_mb_pos];
-             stride_y   = s->linesize << fieldtx;
-             v_dist     = fieldtx ? 15 : 8;
-             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][0],
-                                               s->dest[0] - 16 * s->linesize,
-                                               stride_y);
-             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][1],
-                                               s->dest[0] - 16 * s->linesize + 8,
-                                               stride_y);
-             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][2],
-                                               s->dest[0] - v_dist * s->linesize,
-                                               stride_y);
-             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][3],
-                                               s->dest[0] - v_dist * s->linesize + 8,
-                                               stride_y);
-             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][4],
-                                               s->dest[1] - 8 * s->uvlinesize,
-                                               s->uvlinesize);
-             s->idsp.put_signed_pixels_clamped(v->block[v->top_blk_idx][5],
-                                               s->dest[2] - 8 * s->uvlinesize,
-                                               s->uvlinesize);
-         }
-     }
- #define inc_blk_idx(idx) do { \
-         idx++; \
-         if (idx >= v->n_allocated_blks) \
-             idx = 0; \
-     } while (0)
-     inc_blk_idx(v->topleft_blk_idx);
-     inc_blk_idx(v->top_blk_idx);
-     inc_blk_idx(v->left_blk_idx);
-     inc_blk_idx(v->cur_blk_idx);
- }
- static void vc1_loop_filter_iblk(VC1Context *v, int pq)
- {
-     MpegEncContext *s = &v->s;
-     int j;
-     if (!s->first_slice_line) {
-         v->vc1dsp.vc1_v_loop_filter16(s->dest[0], s->linesize, pq);
-         if (s->mb_x)
-             v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize, s->linesize, pq);
-         v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize + 8, s->linesize, pq);
-         for (j = 0; j < 2; j++) {
-             v->vc1dsp.vc1_v_loop_filter8(s->dest[j + 1], s->uvlinesize, pq);
-             if (s->mb_x)
-                 v->vc1dsp.vc1_h_loop_filter8(s->dest[j + 1] - 8 * s->uvlinesize, s->uvlinesize, pq);
-         }
-     }
-     v->vc1dsp.vc1_v_loop_filter16(s->dest[0] + 8 * s->linesize, s->linesize, pq);
-     if (s->mb_y == s->end_mb_y - 1) {
-         if (s->mb_x) {
-             v->vc1dsp.vc1_h_loop_filter16(s->dest[0], s->linesize, pq);
-             v->vc1dsp.vc1_h_loop_filter8(s->dest[1], s->uvlinesize, pq);
-             v->vc1dsp.vc1_h_loop_filter8(s->dest[2], s->uvlinesize, pq);
-         }
-         v->vc1dsp.vc1_h_loop_filter16(s->dest[0] + 8, s->linesize, pq);
-     }
- }
- static void vc1_loop_filter_iblk_delayed(VC1Context *v, int pq)
- {
-     MpegEncContext *s = &v->s;
-     int j;
-     /* The loopfilter runs 1 row and 1 column behind the overlap filter, which
-      * means it runs two rows/cols behind the decoding loop. */
-     if (!s->first_slice_line) {
-         if (s->mb_x) {
-             if (s->mb_y >= s->start_mb_y + 2) {
-                 v->vc1dsp.vc1_v_loop_filter16(s->dest[0] - 16 * s->linesize - 16, s->linesize, pq);
-                 if (s->mb_x >= 2)
-                     v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 32 * s->linesize - 16, s->linesize, pq);
-                 v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 32 * s->linesize - 8, s->linesize, pq);
-                 for (j = 0; j < 2; j++) {
-                     v->vc1dsp.vc1_v_loop_filter8(s->dest[j + 1] - 8 * s->uvlinesize - 8, s->uvlinesize, pq);
-                     if (s->mb_x >= 2) {
-                         v->vc1dsp.vc1_h_loop_filter8(s->dest[j + 1] - 16 * s->uvlinesize - 8, s->uvlinesize, pq);
-                     }
-                 }
-             }
-             v->vc1dsp.vc1_v_loop_filter16(s->dest[0] - 8 * s->linesize - 16, s->linesize, pq);
-         }
-         if (s->mb_x == s->mb_width - 1) {
-             if (s->mb_y >= s->start_mb_y + 2) {
-                 v->vc1dsp.vc1_v_loop_filter16(s->dest[0] - 16 * s->linesize, s->linesize, pq);
-                 if (s->mb_x)
-                     v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 32 * s->linesize, s->linesize, pq);
-                 v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 32 * s->linesize + 8, s->linesize, pq);
-                 for (j = 0; j < 2; j++) {
-                     v->vc1dsp.vc1_v_loop_filter8(s->dest[j + 1] - 8 * s->uvlinesize, s->uvlinesize, pq);
-                     if (s->mb_x >= 2) {
-                         v->vc1dsp.vc1_h_loop_filter8(s->dest[j + 1] - 16 * s->uvlinesize, s->uvlinesize, pq);
-                     }
-                 }
-             }
-             v->vc1dsp.vc1_v_loop_filter16(s->dest[0] - 8 * s->linesize, s->linesize, pq);
-         }
-         if (s->mb_y == s->end_mb_y) {
-             if (s->mb_x) {
-                 if (s->mb_x >= 2)
-                     v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize - 16, s->linesize, pq);
-                 v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize - 8, s->linesize, pq);
-                 if (s->mb_x >= 2) {
-                     for (j = 0; j < 2; j++) {
-                         v->vc1dsp.vc1_h_loop_filter8(s->dest[j + 1] - 8 * s->uvlinesize - 8, s->uvlinesize, pq);
-                     }
-                 }
-             }
-             if (s->mb_x == s->mb_width - 1) {
-                 if (s->mb_x)
-                     v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize, s->linesize, pq);
-                 v->vc1dsp.vc1_h_loop_filter16(s->dest[0] - 16 * s->linesize + 8, s->linesize, pq);
-                 if (s->mb_x) {
-                     for (j = 0; j < 2; j++) {
-                         v->vc1dsp.vc1_h_loop_filter8(s->dest[j + 1] - 8 * s->uvlinesize, s->uvlinesize, pq);
-                     }
-                 }
-             }
-         }
-     }
- }
- static void vc1_smooth_overlap_filter_iblk(VC1Context *v)
- {
-     MpegEncContext *s = &v->s;
-     int mb_pos;
-     if (v->condover == CONDOVER_NONE)
-         return;
-     mb_pos = s->mb_x + s->mb_y * s->mb_stride;
-     /* Within a MB, the horizontal overlap always runs before the vertical.
-      * To accomplish that, we run the H on left and internal borders of the
-      * currently decoded MB. Then, we wait for the next overlap iteration
-      * to do H overlap on the right edge of this MB, before moving over and
-      * running the V overlap. Therefore, the V overlap makes us trail by one
-      * MB col and the H overlap filter makes us trail by one MB row. This
-      * is reflected in the time at which we run the put_pixels loop. */
-     if (v->condover == CONDOVER_ALL || v->pq >= 9 || v->over_flags_plane[mb_pos]) {
-         if (s->mb_x && (v->condover == CONDOVER_ALL || v->pq >= 9 ||
-                         v->over_flags_plane[mb_pos - 1])) {
-             v->vc1dsp.vc1_h_s_overlap(v->block[v->left_blk_idx][1],
-                                       v->block[v->cur_blk_idx][0]);
-             v->vc1dsp.vc1_h_s_overlap(v->block[v->left_blk_idx][3],
-                                       v->block[v->cur_blk_idx][2]);
-             if (!(s->flags & CODEC_FLAG_GRAY)) {
-                 v->vc1dsp.vc1_h_s_overlap(v->block[v->left_blk_idx][4],
-                                           v->block[v->cur_blk_idx][4]);
-                 v->vc1dsp.vc1_h_s_overlap(v->block[v->left_blk_idx][5],
-                                           v->block[v->cur_blk_idx][5]);
-             }
-         }
-         v->vc1dsp.vc1_h_s_overlap(v->block[v->cur_blk_idx][0],
-                                   v->block[v->cur_blk_idx][1]);
-         v->vc1dsp.vc1_h_s_overlap(v->block[v->cur_blk_idx][2],
-                                   v->block[v->cur_blk_idx][3]);
-         if (s->mb_x == s->mb_width - 1) {
-             if (!s->first_slice_line && (v->condover == CONDOVER_ALL || v->pq >= 9 ||
-                                          v->over_flags_plane[mb_pos - s->mb_stride])) {
-                 v->vc1dsp.vc1_v_s_overlap(v->block[v->top_blk_idx][2],
-                                           v->block[v->cur_blk_idx][0]);
-                 v->vc1dsp.vc1_v_s_overlap(v->block[v->top_blk_idx][3],
-                                           v->block[v->cur_blk_idx][1]);
-                 if (!(s->flags & CODEC_FLAG_GRAY)) {
-                     v->vc1dsp.vc1_v_s_overlap(v->block[v->top_blk_idx][4],
-                                               v->block[v->cur_blk_idx][4]);
-                     v->vc1dsp.vc1_v_s_overlap(v->block[v->top_blk_idx][5],
-                                               v->block[v->cur_blk_idx][5]);
-                 }
-             }
-             v->vc1dsp.vc1_v_s_overlap(v->block[v->cur_blk_idx][0],
-                                       v->block[v->cur_blk_idx][2]);
-             v->vc1dsp.vc1_v_s_overlap(v->block[v->cur_blk_idx][1],
-                                       v->block[v->cur_blk_idx][3]);
-         }
-     }
-     if (s->mb_x && (v->condover == CONDOVER_ALL || v->over_flags_plane[mb_pos - 1])) {
-         if (!s->first_slice_line && (v->condover == CONDOVER_ALL || v->pq >= 9 ||
-                                      v->over_flags_plane[mb_pos - s->mb_stride - 1])) {
-             v->vc1dsp.vc1_v_s_overlap(v->block[v->topleft_blk_idx][2],
-                                       v->block[v->left_blk_idx][0]);
-             v->vc1dsp.vc1_v_s_overlap(v->block[v->topleft_blk_idx][3],
-                                       v->block[v->left_blk_idx][1]);
-             if (!(s->flags & CODEC_FLAG_GRAY)) {
-                 v->vc1dsp.vc1_v_s_overlap(v->block[v->topleft_blk_idx][4],
-                                           v->block[v->left_blk_idx][4]);
-                 v->vc1dsp.vc1_v_s_overlap(v->block[v->topleft_blk_idx][5],
-                                           v->block[v->left_blk_idx][5]);
-             }
-         }
-         v->vc1dsp.vc1_v_s_overlap(v->block[v->left_blk_idx][0],
-                                   v->block[v->left_blk_idx][2]);
-         v->vc1dsp.vc1_v_s_overlap(v->block[v->left_blk_idx][1],
-                                   v->block[v->left_blk_idx][3]);
-     }
- }
- /** Do motion compensation over 1 macroblock
-  * Mostly adapted hpel_motion and qpel_motion from mpegvideo.c
-  */
- static void vc1_mc_1mv(VC1Context *v, int dir)
- {
-     MpegEncContext *s = &v->s;
-     H264ChromaContext *h264chroma = &v->h264chroma;
-     uint8_t *srcY, *srcU, *srcV;
-     int dxy, mx, my, uvmx, uvmy, src_x, src_y, uvsrc_x, uvsrc_y;
-     int v_edge_pos = s->v_edge_pos >> v->field_mode;
-     int i;
-     uint8_t (*luty)[256], (*lutuv)[256];
-     int use_ic;
-     if ((!v->field_mode ||
-          (v->ref_field_type[dir] == 1 && v->cur_field_type == 1)) &&
-         !v->s.last_picture.f->data[0])
-         return;
-     mx = s->mv[dir][0][0];
-     my = s->mv[dir][0][1];
-     // store motion vectors for further use in B frames
-     if (s->pict_type == AV_PICTURE_TYPE_P) {
-         for (i = 0; i < 4; i++) {
-             s->current_picture.motion_val[1][s->block_index[i] + v->blocks_off][0] = mx;
-             s->current_picture.motion_val[1][s->block_index[i] + v->blocks_off][1] = my;
-         }
-     }
-     uvmx = (mx + ((mx & 3) == 3)) >> 1;
-     uvmy = (my + ((my & 3) == 3)) >> 1;
-     v->luma_mv[s->mb_x][0] = uvmx;
-     v->luma_mv[s->mb_x][1] = uvmy;
-     if (v->field_mode &&
-         v->cur_field_type != v->ref_field_type[dir]) {
-         my   = my   - 2 + 4 * v->cur_field_type;
-         uvmy = uvmy - 2 + 4 * v->cur_field_type;
-     }
-     // fastuvmc shall be ignored for interlaced frame picture
-     if (v->fastuvmc && (v->fcm != ILACE_FRAME)) {
-         uvmx = uvmx + ((uvmx < 0) ? (uvmx & 1) : -(uvmx & 1));
-         uvmy = uvmy + ((uvmy < 0) ? (uvmy & 1) : -(uvmy & 1));
-     }
-     if (!dir) {
-         if (v->field_mode && (v->cur_field_type != v->ref_field_type[dir]) && v->second_field) {
-             srcY = s->current_picture.f->data[0];
-             srcU = s->current_picture.f->data[1];
-             srcV = s->current_picture.f->data[2];
-             luty  = v->curr_luty;
-             lutuv = v->curr_lutuv;
-             use_ic = *v->curr_use_ic;
-         } else {
-             srcY = s->last_picture.f->data[0];
-             srcU = s->last_picture.f->data[1];
-             srcV = s->last_picture.f->data[2];
-             luty  = v->last_luty;
-             lutuv = v->last_lutuv;
-             use_ic = v->last_use_ic;
-         }
-     } else {
-         srcY = s->next_picture.f->data[0];
-         srcU = s->next_picture.f->data[1];
-         srcV = s->next_picture.f->data[2];
-         luty  = v->next_luty;
-         lutuv = v->next_lutuv;
-         use_ic = v->next_use_ic;
-     }
-     if (!srcY || !srcU) {
-         av_log(v->s.avctx, AV_LOG_ERROR, "Referenced frame missing.\n");
-         return;
-     }
-     src_x   = s->mb_x * 16 + (mx   >> 2);
-     src_y   = s->mb_y * 16 + (my   >> 2);
-     uvsrc_x = s->mb_x *  8 + (uvmx >> 2);
-     uvsrc_y = s->mb_y *  8 + (uvmy >> 2);
-     if (v->profile != PROFILE_ADVANCED) {
-         src_x   = av_clip(  src_x, -16, s->mb_width  * 16);
-         src_y   = av_clip(  src_y, -16, s->mb_height * 16);
-         uvsrc_x = av_clip(uvsrc_x,  -8, s->mb_width  *  8);
-         uvsrc_y = av_clip(uvsrc_y,  -8, s->mb_height *  8);
-     } else {
-         src_x   = av_clip(  src_x, -17, s->avctx->coded_width);
-         src_y   = av_clip(  src_y, -18, s->avctx->coded_height + 1);
-         uvsrc_x = av_clip(uvsrc_x,  -8, s->avctx->coded_width  >> 1);
-         uvsrc_y = av_clip(uvsrc_y,  -8, s->avctx->coded_height >> 1);
-     }
-     srcY += src_y   * s->linesize   + src_x;
-     srcU += uvsrc_y * s->uvlinesize + uvsrc_x;
-     srcV += uvsrc_y * s->uvlinesize + uvsrc_x;
-     if (v->field_mode && v->ref_field_type[dir]) {
-         srcY += s->current_picture_ptr->f->linesize[0];
-         srcU += s->current_picture_ptr->f->linesize[1];
-         srcV += s->current_picture_ptr->f->linesize[2];
-     }
-     /* for grayscale we should not try to read from unknown area */
-     if (s->flags & CODEC_FLAG_GRAY) {
-         srcU = s->edge_emu_buffer + 18 * s->linesize;
-         srcV = s->edge_emu_buffer + 18 * s->linesize;
-     }
-     if (v->rangeredfrm || use_ic
-         || s->h_edge_pos < 22 || v_edge_pos < 22
-         || (unsigned)(src_x - s->mspel) > s->h_edge_pos - (mx&3) - 16 - s->mspel * 3
-         || (unsigned)(src_y - 1)        > v_edge_pos    - (my&3) - 16 - 3) {
-         uint8_t *ubuf = s->edge_emu_buffer + 19 * s->linesize;
-         uint8_t *vbuf = ubuf + 9 * s->uvlinesize;
-         srcY -= s->mspel * (1 + s->linesize);
-         s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
-                                  s->linesize, s->linesize,
-                                  17 + s->mspel * 2, 17 + s->mspel * 2,
-                                  src_x - s->mspel, src_y - s->mspel,
-                                  s->h_edge_pos, v_edge_pos);
-         srcY = s->edge_emu_buffer;
-         s->vdsp.emulated_edge_mc(ubuf, srcU,
-                                  s->uvlinesize, s->uvlinesize,
-                                  8 + 1, 8 + 1,
-                                  uvsrc_x, uvsrc_y,
-                                  s->h_edge_pos >> 1, v_edge_pos >> 1);
-         s->vdsp.emulated_edge_mc(vbuf, srcV,
-                                  s->uvlinesize, s->uvlinesize,
-                                  8 + 1, 8 + 1,
-                                  uvsrc_x, uvsrc_y,
-                                  s->h_edge_pos >> 1, v_edge_pos >> 1);
-         srcU = ubuf;
-         srcV = vbuf;
-         /* if we deal with range reduction we need to scale source blocks */
-         if (v->rangeredfrm) {
-             int i, j;
-             uint8_t *src, *src2;
-             src = srcY;
-             for (j = 0; j < 17 + s->mspel * 2; j++) {
-                 for (i = 0; i < 17 + s->mspel * 2; i++)
-                     src[i] = ((src[i] - 128) >> 1) + 128;
-                 src += s->linesize;
-             }
-             src  = srcU;
-             src2 = srcV;
-             for (j = 0; j < 9; j++) {
-                 for (i = 0; i < 9; i++) {
-                     src[i]  = ((src[i]  - 128) >> 1) + 128;
-                     src2[i] = ((src2[i] - 128) >> 1) + 128;
-                 }
-                 src  += s->uvlinesize;
-                 src2 += s->uvlinesize;
-             }
-         }
-         /* if we deal with intensity compensation we need to scale source blocks */
-         if (use_ic) {
-             int i, j;
-             uint8_t *src, *src2;
-             src = srcY;
-             for (j = 0; j < 17 + s->mspel * 2; j++) {
-                 int f = v->field_mode ? v->ref_field_type[dir] : ((j + src_y - s->mspel) & 1) ;
-                 for (i = 0; i < 17 + s->mspel * 2; i++)
-                     src[i] = luty[f][src[i]];
-                 src += s->linesize;
-             }
-             src  = srcU;
-             src2 = srcV;
-             for (j = 0; j < 9; j++) {
-                 int f = v->field_mode ? v->ref_field_type[dir] : ((j + uvsrc_y) & 1);
-                 for (i = 0; i < 9; i++) {
-                     src[i]  = lutuv[f][src[i]];
-                     src2[i] = lutuv[f][src2[i]];
-                 }
-                 src  += s->uvlinesize;
-                 src2 += s->uvlinesize;
-             }
-         }
-         srcY += s->mspel * (1 + s->linesize);
-     }
-     if (s->mspel) {
-         dxy = ((my & 3) << 2) | (mx & 3);
-         v->vc1dsp.put_vc1_mspel_pixels_tab[0][dxy](s->dest[0]    , srcY    , s->linesize, v->rnd);
-     } else { // hpel mc - always used for luma
-         dxy = (my & 2) | ((mx & 2) >> 1);
-         if (!v->rnd)
-             s->hdsp.put_pixels_tab[0][dxy](s->dest[0], srcY, s->linesize, 16);
-         else
-             s->hdsp.put_no_rnd_pixels_tab[0][dxy](s->dest[0], srcY, s->linesize, 16);
-     }
-     if (s->flags & CODEC_FLAG_GRAY) return;
-     /* Chroma MC always uses qpel bilinear */
-     uvmx = (uvmx & 3) << 1;
-     uvmy = (uvmy & 3) << 1;
-     if (!v->rnd) {
-         h264chroma->put_h264_chroma_pixels_tab[0](s->dest[1], srcU, s->uvlinesize, 8, uvmx, uvmy);
-         h264chroma->put_h264_chroma_pixels_tab[0](s->dest[2], srcV, s->uvlinesize, 8, uvmx, uvmy);
-     } else {
-         v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[0](s->dest[1], srcU, s->uvlinesize, 8, uvmx, uvmy);
-         v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[0](s->dest[2], srcV, s->uvlinesize, 8, uvmx, uvmy);
-     }
- }
- static inline int median4(int a, int b, int c, int d)
- {
-     if (a < b) {
-         if (c < d) return (FFMIN(b, d) + FFMAX(a, c)) / 2;
-         else       return (FFMIN(b, c) + FFMAX(a, d)) / 2;
-     } else {
-         if (c < d) return (FFMIN(a, d) + FFMAX(b, c)) / 2;
-         else       return (FFMIN(a, c) + FFMAX(b, d)) / 2;
-     }
- }
- /** Do motion compensation for 4-MV macroblock - luminance block
-  */
- static void vc1_mc_4mv_luma(VC1Context *v, int n, int dir, int avg)
- {
-     MpegEncContext *s = &v->s;
-     uint8_t *srcY;
-     int dxy, mx, my, src_x, src_y;
-     int off;
-     int fieldmv = (v->fcm == ILACE_FRAME) ? v->blk_mv_type[s->block_index[n]] : 0;
-     int v_edge_pos = s->v_edge_pos >> v->field_mode;
-     uint8_t (*luty)[256];
-     int use_ic;
-     if ((!v->field_mode ||
-          (v->ref_field_type[dir] == 1 && v->cur_field_type == 1)) &&
-         !v->s.last_picture.f->data[0])
-         return;
-     mx = s->mv[dir][n][0];
-     my = s->mv[dir][n][1];
-     if (!dir) {
-         if (v->field_mode && (v->cur_field_type != v->ref_field_type[dir]) && v->second_field) {
-             srcY = s->current_picture.f->data[0];
-             luty = v->curr_luty;
-             use_ic = *v->curr_use_ic;
-         } else {
-             srcY = s->last_picture.f->data[0];
-             luty = v->last_luty;
-             use_ic = v->last_use_ic;
-         }
-     } else {
-         srcY = s->next_picture.f->data[0];
-         luty = v->next_luty;
-         use_ic = v->next_use_ic;
-     }
-     if (!srcY) {
-         av_log(v->s.avctx, AV_LOG_ERROR, "Referenced frame missing.\n");
-         return;
-     }
-     if (v->field_mode) {
-         if (v->cur_field_type != v->ref_field_type[dir])
-             my = my - 2 + 4 * v->cur_field_type;
-     }
-     if (s->pict_type == AV_PICTURE_TYPE_P && n == 3 && v->field_mode) {
-         int same_count = 0, opp_count = 0, k;
-         int chosen_mv[2][4][2], f;
-         int tx, ty;
-         for (k = 0; k < 4; k++) {
-             f = v->mv_f[0][s->block_index[k] + v->blocks_off];
-             chosen_mv[f][f ? opp_count : same_count][0] = s->mv[0][k][0];
-             chosen_mv[f][f ? opp_count : same_count][1] = s->mv[0][k][1];
-             opp_count  += f;
-             same_count += 1 - f;
-         }
-         f = opp_count > same_count;
-         switch (f ? opp_count : same_count) {
-         case 4:
-             tx = median4(chosen_mv[f][0][0], chosen_mv[f][1][0],
-                          chosen_mv[f][2][0], chosen_mv[f][3][0]);
-             ty = median4(chosen_mv[f][0][1], chosen_mv[f][1][1],
-                          chosen_mv[f][2][1], chosen_mv[f][3][1]);
-             break;
-         case 3:
-             tx = mid_pred(chosen_mv[f][0][0], chosen_mv[f][1][0], chosen_mv[f][2][0]);
-             ty = mid_pred(chosen_mv[f][0][1], chosen_mv[f][1][1], chosen_mv[f][2][1]);
-             break;
-         case 2:
-             tx = (chosen_mv[f][0][0] + chosen_mv[f][1][0]) / 2;
-             ty = (chosen_mv[f][0][1] + chosen_mv[f][1][1]) / 2;
-             break;
-         default:
-             av_assert0(0);
-         }
-         s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][0] = tx;
-         s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][1] = ty;
-         for (k = 0; k < 4; k++)
-             v->mv_f[1][s->block_index[k] + v->blocks_off] = f;
-     }
-     if (v->fcm == ILACE_FRAME) {  // not sure if needed for other types of picture
-         int qx, qy;
-         int width  = s->avctx->coded_width;
-         int height = s->avctx->coded_height >> 1;
-         if (s->pict_type == AV_PICTURE_TYPE_P) {
-             s->current_picture.motion_val[1][s->block_index[n] + v->blocks_off][0] = mx;
-             s->current_picture.motion_val[1][s->block_index[n] + v->blocks_off][1] = my;
-         }
-         qx = (s->mb_x * 16) + (mx >> 2);
-         qy = (s->mb_y *  8) + (my >> 3);
-         if (qx < -17)
-             mx -= 4 * (qx + 17);
-         else if (qx > width)
-             mx -= 4 * (qx - width);
-         if (qy < -18)
-             my -= 8 * (qy + 18);
-         else if (qy > height + 1)
-             my -= 8 * (qy - height - 1);
-     }
-     if ((v->fcm == ILACE_FRAME) && fieldmv)
-         off = ((n > 1) ? s->linesize : 0) + (n & 1) * 8;
-     else
-         off = s->linesize * 4 * (n & 2) + (n & 1) * 8;
-     src_x = s->mb_x * 16 + (n & 1) * 8 + (mx >> 2);
-     if (!fieldmv)
-         src_y = s->mb_y * 16 + (n & 2) * 4 + (my >> 2);
-     else
-         src_y = s->mb_y * 16 + ((n > 1) ? 1 : 0) + (my >> 2);
-     if (v->profile != PROFILE_ADVANCED) {
-         src_x = av_clip(src_x, -16, s->mb_width  * 16);
-         src_y = av_clip(src_y, -16, s->mb_height * 16);
-     } else {
-         src_x = av_clip(src_x, -17, s->avctx->coded_width);
-         if (v->fcm == ILACE_FRAME) {
-             if (src_y & 1)
-                 src_y = av_clip(src_y, -17, s->avctx->coded_height + 1);
-             else
-                 src_y = av_clip(src_y, -18, s->avctx->coded_height);
-         } else {
-             src_y = av_clip(src_y, -18, s->avctx->coded_height + 1);
-         }
-     }
-     srcY += src_y * s->linesize + src_x;
-     if (v->field_mode && v->ref_field_type[dir])
-         srcY += s->current_picture_ptr->f->linesize[0];
-     if (fieldmv && !(src_y & 1))
-         v_edge_pos--;
-     if (fieldmv && (src_y & 1) && src_y < 4)
-         src_y--;
-     if (v->rangeredfrm || use_ic
-         || s->h_edge_pos < 13 || v_edge_pos < 23
-         || (unsigned)(src_x - s->mspel) > s->h_edge_pos - (mx & 3) - 8 - s->mspel * 2
-         || (unsigned)(src_y - (s->mspel << fieldmv)) > v_edge_pos - (my & 3) - ((8 + s->mspel * 2) << fieldmv)) {
-         srcY -= s->mspel * (1 + (s->linesize << fieldmv));
-         /* check emulate edge stride and offset */
-         s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcY,
-                                  s->linesize, s->linesize,
-                                  9 + s->mspel * 2, (9 + s->mspel * 2) << fieldmv,
-                                  src_x - s->mspel, src_y - (s->mspel << fieldmv),
-                                  s->h_edge_pos, v_edge_pos);
-         srcY = s->edge_emu_buffer;
-         /* if we deal with range reduction we need to scale source blocks */
-         if (v->rangeredfrm) {
-             int i, j;
-             uint8_t *src;
-             src = srcY;
-             for (j = 0; j < 9 + s->mspel * 2; j++) {
-                 for (i = 0; i < 9 + s->mspel * 2; i++)
-                     src[i] = ((src[i] - 128) >> 1) + 128;
-                 src += s->linesize << fieldmv;
-             }
-         }
-         /* if we deal with intensity compensation we need to scale source blocks */
-         if (use_ic) {
-             int i, j;
-             uint8_t *src;
-             src = srcY;
-             for (j = 0; j < 9 + s->mspel * 2; j++) {
-                 int f = v->field_mode ? v->ref_field_type[dir] : (((j<<fieldmv)+src_y - (s->mspel << fieldmv)) & 1);
-                 for (i = 0; i < 9 + s->mspel * 2; i++)
-                     src[i] = luty[f][src[i]];
-                 src += s->linesize << fieldmv;
-             }
-         }
-         srcY += s->mspel * (1 + (s->linesize << fieldmv));
-     }
-     if (s->mspel) {
-         dxy = ((my & 3) << 2) | (mx & 3);
-         if (avg)
-             v->vc1dsp.avg_vc1_mspel_pixels_tab[1][dxy](s->dest[0] + off, srcY, s->linesize << fieldmv, v->rnd);
-         else
-             v->vc1dsp.put_vc1_mspel_pixels_tab[1][dxy](s->dest[0] + off, srcY, s->linesize << fieldmv, v->rnd);
-     } else { // hpel mc - always used for luma
-         dxy = (my & 2) | ((mx & 2) >> 1);
-         if (!v->rnd)
-             s->hdsp.put_pixels_tab[1][dxy](s->dest[0] + off, srcY, s->linesize, 8);
-         else
-             s->hdsp.put_no_rnd_pixels_tab[1][dxy](s->dest[0] + off, srcY, s->linesize, 8);
-     }
- }
- static av_always_inline int get_chroma_mv(int *mvx, int *mvy, int *a, int flag, int *tx, int *ty)
- {
-     int idx, i;
-     static const int count[16] = { 0, 1, 1, 2, 1, 2, 2, 3, 1, 2, 2, 3, 2, 3, 3, 4};
-     idx =  ((a[3] != flag) << 3)
-          | ((a[2] != flag) << 2)
-          | ((a[1] != flag) << 1)
-          |  (a[0] != flag);
-     if (!idx) {
-         *tx = median4(mvx[0], mvx[1], mvx[2], mvx[3]);
-         *ty = median4(mvy[0], mvy[1], mvy[2], mvy[3]);
-         return 4;
-     } else if (count[idx] == 1) {
-         switch (idx) {
-         case 0x1:
-             *tx = mid_pred(mvx[1], mvx[2], mvx[3]);
-             *ty = mid_pred(mvy[1], mvy[2], mvy[3]);
-             return 3;
-         case 0x2:
-             *tx = mid_pred(mvx[0], mvx[2], mvx[3]);
-             *ty = mid_pred(mvy[0], mvy[2], mvy[3]);
-             return 3;
-         case 0x4:
-             *tx = mid_pred(mvx[0], mvx[1], mvx[3]);
-             *ty = mid_pred(mvy[0], mvy[1], mvy[3]);
-             return 3;
-         case 0x8:
-             *tx = mid_pred(mvx[0], mvx[1], mvx[2]);
-             *ty = mid_pred(mvy[0], mvy[1], mvy[2]);
-             return 3;
-         }
-     } else if (count[idx] == 2) {
-         int t1 = 0, t2 = 0;
-         for (i = 0; i < 3; i++)
-             if (!a[i]) {
-                 t1 = i;
-                 break;
-             }
-         for (i = t1 + 1; i < 4; i++)
-             if (!a[i]) {
-                 t2 = i;
-                 break;
-             }
-         *tx = (mvx[t1] + mvx[t2]) / 2;
-         *ty = (mvy[t1] + mvy[t2]) / 2;
-         return 2;
-     } else {
-         return 0;
-     }
-     return -1;
- }
- /** Do motion compensation for 4-MV macroblock - both chroma blocks
-  */
- static void vc1_mc_4mv_chroma(VC1Context *v, int dir)
- {
-     MpegEncContext *s = &v->s;
-     H264ChromaContext *h264chroma = &v->h264chroma;
-     uint8_t *srcU, *srcV;
-     int uvmx, uvmy, uvsrc_x, uvsrc_y;
-     int k, tx = 0, ty = 0;
-     int mvx[4], mvy[4], intra[4], mv_f[4];
-     int valid_count;
-     int chroma_ref_type = v->cur_field_type;
-     int v_edge_pos = s->v_edge_pos >> v->field_mode;
-     uint8_t (*lutuv)[256];
-     int use_ic;
-     if (!v->field_mode && !v->s.last_picture.f->data[0])
-         return;
-     if (s->flags & CODEC_FLAG_GRAY)
-         return;
-     for (k = 0; k < 4; k++) {
-         mvx[k] = s->mv[dir][k][0];
-         mvy[k] = s->mv[dir][k][1];
-         intra[k] = v->mb_type[0][s->block_index[k]];
-         if (v->field_mode)
-             mv_f[k] = v->mv_f[dir][s->block_index[k] + v->blocks_off];
-     }
-     /* calculate chroma MV vector from four luma MVs */
-     if (!v->field_mode || (v->field_mode && !v->numref)) {
-         valid_count = get_chroma_mv(mvx, mvy, intra, 0, &tx, &ty);
-         chroma_ref_type = v->reffield;
-         if (!valid_count) {
-             s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][0] = 0;
-             s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][1] = 0;
-             v->luma_mv[s->mb_x][0] = v->luma_mv[s->mb_x][1] = 0;
-             return; //no need to do MC for intra blocks
-         }
-     } else {
-         int dominant = 0;
-         if (mv_f[0] + mv_f[1] + mv_f[2] + mv_f[3] > 2)
-             dominant = 1;
-         valid_count = get_chroma_mv(mvx, mvy, mv_f, dominant, &tx, &ty);
-         if (dominant)
-             chroma_ref_type = !v->cur_field_type;
-     }
-     if (v->field_mode && chroma_ref_type == 1 && v->cur_field_type == 1 && !v->s.last_picture.f->data[0])
-         return;
-     s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][0] = tx;
-     s->current_picture.motion_val[1][s->block_index[0] + v->blocks_off][1] = ty;
-     uvmx = (tx + ((tx & 3) == 3)) >> 1;
-     uvmy = (ty + ((ty & 3) == 3)) >> 1;
-     v->luma_mv[s->mb_x][0] = uvmx;
-     v->luma_mv[s->mb_x][1] = uvmy;
-     if (v->fastuvmc) {
-         uvmx = uvmx + ((uvmx < 0) ? (uvmx & 1) : -(uvmx & 1));
-         uvmy = uvmy + ((uvmy < 0) ? (uvmy & 1) : -(uvmy & 1));
-     }
-     // Field conversion bias
-     if (v->cur_field_type != chroma_ref_type)
-         uvmy += 2 - 4 * chroma_ref_type;
-     uvsrc_x = s->mb_x * 8 + (uvmx >> 2);
-     uvsrc_y = s->mb_y * 8 + (uvmy >> 2);
-     if (v->profile != PROFILE_ADVANCED) {
-         uvsrc_x = av_clip(uvsrc_x, -8, s->mb_width  * 8);
-         uvsrc_y = av_clip(uvsrc_y, -8, s->mb_height * 8);
-     } else {
-         uvsrc_x = av_clip(uvsrc_x, -8, s->avctx->coded_width  >> 1);
-         uvsrc_y = av_clip(uvsrc_y, -8, s->avctx->coded_height >> 1);
-     }
-     if (!dir) {
-         if (v->field_mode && (v->cur_field_type != chroma_ref_type) && v->second_field) {
-             srcU = s->current_picture.f->data[1];
-             srcV = s->current_picture.f->data[2];
-             lutuv = v->curr_lutuv;
-             use_ic = *v->curr_use_ic;
-         } else {
-             srcU = s->last_picture.f->data[1];
-             srcV = s->last_picture.f->data[2];
-             lutuv = v->last_lutuv;
-             use_ic = v->last_use_ic;
-         }
-     } else {
-         srcU = s->next_picture.f->data[1];
-         srcV = s->next_picture.f->data[2];
-         lutuv = v->next_lutuv;
-         use_ic = v->next_use_ic;
-     }
-     if (!srcU) {
-         av_log(v->s.avctx, AV_LOG_ERROR, "Referenced frame missing.\n");
-         return;
-     }
-     srcU += uvsrc_y * s->uvlinesize + uvsrc_x;
-     srcV += uvsrc_y * s->uvlinesize + uvsrc_x;
-     if (v->field_mode) {
-         if (chroma_ref_type) {
-             srcU += s->current_picture_ptr->f->linesize[1];
-             srcV += s->current_picture_ptr->f->linesize[2];
-         }
-     }
-     if (v->rangeredfrm || use_ic
-         || s->h_edge_pos < 18 || v_edge_pos < 18
-         || (unsigned)uvsrc_x > (s->h_edge_pos >> 1) - 9
-         || (unsigned)uvsrc_y > (v_edge_pos    >> 1) - 9) {
-         s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcU,
-                                  s->uvlinesize, s->uvlinesize,
-                                  8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
-                                  s->h_edge_pos >> 1, v_edge_pos >> 1);
-         s->vdsp.emulated_edge_mc(s->edge_emu_buffer + 16, srcV,
-                                  s->uvlinesize, s->uvlinesize,
-                                  8 + 1, 8 + 1, uvsrc_x, uvsrc_y,
-                                  s->h_edge_pos >> 1, v_edge_pos >> 1);
-         srcU = s->edge_emu_buffer;
-         srcV = s->edge_emu_buffer + 16;
-         /* if we deal with range reduction we need to scale source blocks */
-         if (v->rangeredfrm) {
-             int i, j;
-             uint8_t *src, *src2;
-             src  = srcU;
-             src2 = srcV;
-             for (j = 0; j < 9; j++) {
-                 for (i = 0; i < 9; i++) {
-                     src[i]  = ((src[i]  - 128) >> 1) + 128;
-                     src2[i] = ((src2[i] - 128) >> 1) + 128;
-                 }
-                 src  += s->uvlinesize;
-                 src2 += s->uvlinesize;
-             }
-         }
-         /* if we deal with intensity compensation we need to scale source blocks */
-         if (use_ic) {
-             int i, j;
-             uint8_t *src, *src2;
-             src  = srcU;
-             src2 = srcV;
-             for (j = 0; j < 9; j++) {
-                 int f = v->field_mode ? chroma_ref_type : ((j + uvsrc_y) & 1);
-                 for (i = 0; i < 9; i++) {
-                     src[i]  = lutuv[f][src[i]];
-                     src2[i] = lutuv[f][src2[i]];
-                 }
-                 src  += s->uvlinesize;
-                 src2 += s->uvlinesize;
-             }
-         }
-     }
-     /* Chroma MC always uses qpel bilinear */
-     uvmx = (uvmx & 3) << 1;
-     uvmy = (uvmy & 3) << 1;
-     if (!v->rnd) {
-         h264chroma->put_h264_chroma_pixels_tab[0](s->dest[1], srcU, s->uvlinesize, 8, uvmx, uvmy);
-         h264chroma->put_h264_chroma_pixels_tab[0](s->dest[2], srcV, s->uvlinesize, 8, uvmx, uvmy);
-     } else {
-         v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[0](s->dest[1], srcU, s->uvlinesize, 8, uvmx, uvmy);
-         v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[0](s->dest[2], srcV, s->uvlinesize, 8, uvmx, uvmy);
-     }
- }
- /** Do motion compensation for 4-MV interlaced frame chroma macroblock (both U and V)
-  */
- static void vc1_mc_4mv_chroma4(VC1Context *v, int dir, int dir2, int avg)
- {
-     MpegEncContext *s = &v->s;
-     H264ChromaContext *h264chroma = &v->h264chroma;
-     uint8_t *srcU, *srcV;
-     int uvsrc_x, uvsrc_y;
-     int uvmx_field[4], uvmy_field[4];
-     int i, off, tx, ty;
-     int fieldmv = v->blk_mv_type[s->block_index[0]];
-     static const int s_rndtblfield[16] = { 0, 0, 1, 2, 4, 4, 5, 6, 2, 2, 3, 8, 6, 6, 7, 12 };
-     int v_dist = fieldmv ? 1 : 4; // vertical offset for lower sub-blocks
-     int v_edge_pos = s->v_edge_pos >> 1;
-     int use_ic;
-     uint8_t (*lutuv)[256];
-     if (s->flags & CODEC_FLAG_GRAY)
-         return;
-     for (i = 0; i < 4; i++) {
-         int d = i < 2 ? dir: dir2;
-         tx = s->mv[d][i][0];
-         uvmx_field[i] = (tx + ((tx & 3) == 3)) >> 1;
-         ty = s->mv[d][i][1];
-         if (fieldmv)
-             uvmy_field[i] = (ty >> 4) * 8 + s_rndtblfield[ty & 0xF];
-         else
-             uvmy_field[i] = (ty + ((ty & 3) == 3)) >> 1;
-     }
-     for (i = 0; i < 4; i++) {
-         off = (i & 1) * 4 + ((i & 2) ? v_dist * s->uvlinesize : 0);
-         uvsrc_x = s->mb_x * 8 +  (i & 1) * 4           + (uvmx_field[i] >> 2);
-         uvsrc_y = s->mb_y * 8 + ((i & 2) ? v_dist : 0) + (uvmy_field[i] >> 2);
-         // FIXME: implement proper pull-back (see vc1cropmv.c, vc1CROPMV_ChromaPullBack())
-         uvsrc_x = av_clip(uvsrc_x, -8, s->avctx->coded_width  >> 1);
-         uvsrc_y = av_clip(uvsrc_y, -8, s->avctx->coded_height >> 1);
-         if (i < 2 ? dir : dir2) {
-             srcU = s->next_picture.f->data[1];
-             srcV = s->next_picture.f->data[2];
-             lutuv  = v->next_lutuv;
-             use_ic = v->next_use_ic;
-         } else {
-             srcU = s->last_picture.f->data[1];
-             srcV = s->last_picture.f->data[2];
-             lutuv  = v->last_lutuv;
-             use_ic = v->last_use_ic;
-         }
-         if (!srcU)
-             return;
-         srcU += uvsrc_y * s->uvlinesize + uvsrc_x;
-         srcV += uvsrc_y * s->uvlinesize + uvsrc_x;
-         uvmx_field[i] = (uvmx_field[i] & 3) << 1;
-         uvmy_field[i] = (uvmy_field[i] & 3) << 1;
-         if (fieldmv && !(uvsrc_y & 1))
-             v_edge_pos = (s->v_edge_pos >> 1) - 1;
-         if (fieldmv && (uvsrc_y & 1) && uvsrc_y < 2)
-             uvsrc_y--;
-         if (use_ic
-             || s->h_edge_pos < 10 || v_edge_pos < (5 << fieldmv)
-             || (unsigned)uvsrc_x > (s->h_edge_pos >> 1) - 5
-             || (unsigned)uvsrc_y > v_edge_pos - (5 << fieldmv)) {
-             s->vdsp.emulated_edge_mc(s->edge_emu_buffer, srcU,
-                                      s->uvlinesize, s->uvlinesize,
-                                      5, (5 << fieldmv), uvsrc_x, uvsrc_y,
-                                      s->h_edge_pos >> 1, v_edge_pos);
-             s->vdsp.emulated_edge_mc(s->edge_emu_buffer + 16, srcV,
-                                      s->uvlinesize, s->uvlinesize,
-                                      5, (5 << fieldmv), uvsrc_x, uvsrc_y,
-                                      s->h_edge_pos >> 1, v_edge_pos);
-             srcU = s->edge_emu_buffer;
-             srcV = s->edge_emu_buffer + 16;
-             /* if we deal with intensity compensation we need to scale source blocks */
-             if (use_ic) {
-                 int i, j;
-                 uint8_t *src, *src2;
-                 src  = srcU;
-                 src2 = srcV;
-                 for (j = 0; j < 5; j++) {
-                     int f = (uvsrc_y + (j << fieldmv)) & 1;
-                     for (i = 0; i < 5; i++) {
-                         src[i]  = lutuv[f][src[i]];
-                         src2[i] = lutuv[f][src2[i]];
-                     }
-                     src  += s->uvlinesize << fieldmv;
-                     src2 += s->uvlinesize << fieldmv;
-                 }
-             }
-         }
-         if (avg) {
-             if (!v->rnd) {
-                 h264chroma->avg_h264_chroma_pixels_tab[1](s->dest[1] + off, srcU, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
-                 h264chroma->avg_h264_chroma_pixels_tab[1](s->dest[2] + off, srcV, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
-             } else {
-                 v->vc1dsp.avg_no_rnd_vc1_chroma_pixels_tab[1](s->dest[1] + off, srcU, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
-                 v->vc1dsp.avg_no_rnd_vc1_chroma_pixels_tab[1](s->dest[2] + off, srcV, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
-             }
-         } else {
-             if (!v->rnd) {
-                 h264chroma->put_h264_chroma_pixels_tab[1](s->dest[1] + off, srcU, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
-                 h264chroma->put_h264_chroma_pixels_tab[1](s->dest[2] + off, srcV, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
-             } else {
-                 v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[1](s->dest[1] + off, srcU, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
-                 v->vc1dsp.put_no_rnd_vc1_chroma_pixels_tab[1](s->dest[2] + off, srcV, s->uvlinesize << fieldmv, 4, uvmx_field[i], uvmy_field[i]);
-             }
-         }
-     }
- }
- /***********************************************************************/
- /**
-  * @name VC-1 Block-level functions
-  * @see 7.1.4, p91 and 8.1.1.7, p(1)04
-  * @{
-  */
- /**
-  * @def GET_MQUANT
-  * @brief Get macroblock-level quantizer scale
-  */
- #define GET_MQUANT()                                           \
-     if (v->dquantfrm) {                                        \
-         int edges = 0;                                         \
-         if (v->dqprofile == DQPROFILE_ALL_MBS) {               \
-             if (v->dqbilevel) {                                \
-                 mquant = (get_bits1(gb)) ? v->altpq : v->pq;   \
-             } else {                                           \
-                 mqdiff = get_bits(gb, 3);                      \
-                 if (mqdiff != 7)                               \
-                     mquant = v->pq + mqdiff;                   \
-                 else                                           \
-                     mquant = get_bits(gb, 5);                  \
-             }                                                  \
-         }                                                      \
-         if (v->dqprofile == DQPROFILE_SINGLE_EDGE)             \
-             edges = 1 << v->dqsbedge;                          \
-         else if (v->dqprofile == DQPROFILE_DOUBLE_EDGES)       \
-             edges = (3 << v->dqsbedge) % 15;                   \
-         else if (v->dqprofile == DQPROFILE_FOUR_EDGES)         \
-             edges = 15;                                        \
-         if ((edges&1) && !s->mb_x)                             \
-             mquant = v->altpq;                                 \
-         if ((edges&2) && s->first_slice_line)                  \
-             mquant = v->altpq;                                 \
-         if ((edges&4) && s->mb_x == (s->mb_width - 1))         \
-             mquant = v->altpq;                                 \
-         if ((edges&8) && s->mb_y == (s->mb_height - 1))        \
-             mquant = v->altpq;                                 \
-         if (!mquant || mquant > 31) {                          \
-             av_log(v->s.avctx, AV_LOG_ERROR,                   \
-                    "Overriding invalid mquant %d\n", mquant);  \
-             mquant = 1;                                        \
-         }                                                      \
-     }
- /**
-  * @def GET_MVDATA(_dmv_x, _dmv_y)
-  * @brief Get MV differentials
-  * @see MVDATA decoding from 8.3.5.2, p(1)20
-  * @param _dmv_x Horizontal differential for decoded MV
-  * @param _dmv_y Vertical differential for decoded MV
-  */
- #define GET_MVDATA(_dmv_x, _dmv_y)                                      \
-     index = 1 + get_vlc2(gb, ff_vc1_mv_diff_vlc[s->mv_table_index].table, \
-                          VC1_MV_DIFF_VLC_BITS, 2);                      \
-     if (index > 36) {                                                   \
-         mb_has_coeffs = 1;                                              \
-         index -= 37;                                                    \
-     } else                                                              \
-         mb_has_coeffs = 0;                                              \
-     s->mb_intra = 0;                                                    \
-     if (!index) {                                                       \
-         _dmv_x = _dmv_y = 0;                                            \
-     } else if (index == 35) {                                           \
-         _dmv_x = get_bits(gb, v->k_x - 1 + s->quarter_sample);          \
-         _dmv_y = get_bits(gb, v->k_y - 1 + s->quarter_sample);          \
-     } else if (index == 36) {                                           \
-         _dmv_x = 0;                                                     \
-         _dmv_y = 0;                                                     \
-         s->mb_intra = 1;                                                \
-     } else {                                                            \
-         index1 = index % 6;                                             \
-         if (!s->quarter_sample && index1 == 5) val = 1;                 \
-         else                                   val = 0;                 \
-         if (size_table[index1] - val > 0)                               \
-             val = get_bits(gb, size_table[index1] - val);               \
-         else                                   val = 0;                 \
-         sign = 0 - (val&1);                                             \
-         _dmv_x = (sign ^ ((val>>1) + offset_table[index1])) - sign;     \
-                                                                         \
-         index1 = index / 6;                                             \
-         if (!s->quarter_sample && index1 == 5) val = 1;                 \
-         else                                   val = 0;                 \
-         if (size_table[index1] - val > 0)                               \
-             val = get_bits(gb, size_table[index1] - val);               \
-         else                                   val = 0;                 \
-         sign = 0 - (val & 1);                                           \
-         _dmv_y = (sign ^ ((val >> 1) + offset_table[index1])) - sign;   \
-     }
- static av_always_inline void get_mvdata_interlaced(VC1Context *v, int *dmv_x,
-                                                    int *dmv_y, int *pred_flag)
- {
-     int index, index1;
-     int extend_x = 0, extend_y = 0;
-     GetBitContext *gb = &v->s.gb;
-     int bits, esc;
-     int val, sign;
-     const int* offs_tab;
-     if (v->numref) {
-         bits = VC1_2REF_MVDATA_VLC_BITS;
-         esc  = 125;
-     } else {
-         bits = VC1_1REF_MVDATA_VLC_BITS;
-         esc  = 71;
-     }
-     switch (v->dmvrange) {
-     case 1:
-         extend_x = 1;
-         break;
-     case 2:
-         extend_y = 1;
-         break;
-     case 3:
-         extend_x = extend_y = 1;
-         break;
-     }
-     index = get_vlc2(gb, v->imv_vlc->table, bits, 3);
-     if (index == esc) {
-         *dmv_x = get_bits(gb, v->k_x);
-         *dmv_y = get_bits(gb, v->k_y);
-         if (v->numref) {
-             if (pred_flag) {
-                 *pred_flag = *dmv_y & 1;
-                 *dmv_y     = (*dmv_y + *pred_flag) >> 1;
-             } else {
-                 *dmv_y     = (*dmv_y + (*dmv_y & 1)) >> 1;
-             }
-         }
-     }
-     else {
-         av_assert0(index < esc);
-         if (extend_x)
-             offs_tab = offset_table2;
-         else
-             offs_tab = offset_table1;
-         index1 = (index + 1) % 9;
-         if (index1 != 0) {
-             val    = get_bits(gb, index1 + extend_x);
-             sign   = 0 -(val & 1);
-             *dmv_x = (sign ^ ((val >> 1) + offs_tab[index1])) - sign;
-         } else
-             *dmv_x = 0;
-         if (extend_y)
-             offs_tab = offset_table2;
-         else
-             offs_tab = offset_table1;
-         index1 = (index + 1) / 9;
-         if (index1 > v->numref) {
-             val    = get_bits(gb, (index1 + (extend_y << v->numref)) >> v->numref);
-             sign   = 0 - (val & 1);
-             *dmv_y = (sign ^ ((val >> 1) + offs_tab[index1 >> v->numref])) - sign;
-         } else
-             *dmv_y = 0;
-         if (v->numref && pred_flag)
-             *pred_flag = index1 & 1;
-     }
- }
- static av_always_inline int scaleforsame_x(VC1Context *v, int n /* MV */, int dir)
- {
-     int scaledvalue, refdist;
-     int scalesame1, scalesame2;
-     int scalezone1_x, zone1offset_x;
-     int table_index = dir ^ v->second_field;
-     if (v->s.pict_type != AV_PICTURE_TYPE_B)
-         refdist = v->refdist;
-     else
-         refdist = dir ? v->brfd : v->frfd;
-     if (refdist > 3)
-         refdist = 3;
-     scalesame1    = ff_vc1_field_mvpred_scales[table_index][1][refdist];
-     scalesame2    = ff_vc1_field_mvpred_scales[table_index][2][refdist];
-     scalezone1_x  = ff_vc1_field_mvpred_scales[table_index][3][refdist];
-     zone1offset_x = ff_vc1_field_mvpred_scales[table_index][5][refdist];
-     if (FFABS(n) > 255)
-         scaledvalue = n;
-     else {
-         if (FFABS(n) < scalezone1_x)
-             scaledvalue = (n * scalesame1) >> 8;
-         else {
-             if (n < 0)
-                 scaledvalue = ((n * scalesame2) >> 8) - zone1offset_x;
-             else
-                 scaledvalue = ((n * scalesame2) >> 8) + zone1offset_x;
-         }
-     }
-     return av_clip(scaledvalue, -v->range_x, v->range_x - 1);
- }
- static av_always_inline int scaleforsame_y(VC1Context *v, int i, int n /* MV */, int dir)
- {
-     int scaledvalue, refdist;
-     int scalesame1, scalesame2;
-     int scalezone1_y, zone1offset_y;
-     int table_index = dir ^ v->second_field;
-     if (v->s.pict_type != AV_PICTURE_TYPE_B)
-         refdist = v->refdist;
-     else
-         refdist = dir ? v->brfd : v->frfd;
-     if (refdist > 3)
-         refdist = 3;
-     scalesame1    = ff_vc1_field_mvpred_scales[table_index][1][refdist];
-     scalesame2    = ff_vc1_field_mvpred_scales[table_index][2][refdist];
-     scalezone1_y  = ff_vc1_field_mvpred_scales[table_index][4][refdist];
-     zone1offset_y = ff_vc1_field_mvpred_scales[table_index][6][refdist];
-     if (FFABS(n) > 63)
-         scaledvalue = n;
-     else {
-         if (FFABS(n) < scalezone1_y)
-             scaledvalue = (n * scalesame1) >> 8;
-         else {
-             if (n < 0)
-                 scaledvalue = ((n * scalesame2) >> 8) - zone1offset_y;
-             else
-                 scaledvalue = ((n * scalesame2) >> 8) + zone1offset_y;
-         }
-     }
-     if (v->cur_field_type && !v->ref_field_type[dir])
-         return av_clip(scaledvalue, -v->range_y / 2 + 1, v->range_y / 2);
-     else
-         return av_clip(scaledvalue, -v->range_y / 2, v->range_y / 2 - 1);
- }
- static av_always_inline int scaleforopp_x(VC1Context *v, int n /* MV */)
- {
-     int scalezone1_x, zone1offset_x;