Zoom & Pan filter
authorPaul B Mahol <onemda@gmail.com>
Mon, 16 Sep 2013 15:31:15 +0000 (15:31 +0000)
committerPaul B Mahol <onemda@gmail.com>
Mon, 9 Jun 2014 12:41:33 +0000 (12:41 +0000)
Signed-off-by: Paul B Mahol <onemda@gmail.com>
Changelog
configure
doc/filters.texi
libavfilter/Makefile
libavfilter/allfilters.c
libavfilter/version.h
libavfilter/vf_zoompan.c [new file with mode: 0644]

index 7f3704e..56a5ddf 100644 (file)
--- a/Changelog
+++ b/Changelog
@@ -27,6 +27,7 @@ version <next>:
 - display matrix export and rotation api
 - WebVTT encoder
 - showcqt multimedia filter
+- zoompan filter
 
 
 version 2.2:
index 76edb6b..5f53af1 100755 (executable)
--- a/configure
+++ b/configure
@@ -2540,6 +2540,7 @@ pixfmts_super2xsai_test_deps="super2xsai_filter"
 tinterlace_merge_test_deps="tinterlace_filter"
 tinterlace_pad_test_deps="tinterlace_filter"
 zmq_filter_deps="libzmq"
+zoompan_filter_deps="swscale"
 
 # examples
 avio_reading="avformat avcodec avutil"
index 9dc97d7..d8ed7f7 100644 (file)
@@ -8680,6 +8680,94 @@ Only deinterlace frames marked as interlaced.
 The default value is @code{all}.
 @end table
 
+@section zoompan
+
+Apply Zoom & Pan effect.
+
+This filter accepts the following options:
+
+@table @option
+@item zoom, z
+Set the zoom expression. Default is 1.
+
+@item x
+@item y
+Set the x and y expression. Default is 0.
+
+@item d
+Set the duration expression in number of frames.
+This sets for how many number of frames effect will last for
+single input image.
+
+@item s
+Set the output image size, default is 'hd720'.
+@end table
+
+Each expression can contain the following constants:
+
+@table @option
+@item in_w, iw
+Input width.
+
+@item in_h, ih
+Input height.
+
+@item out_w, ow
+Output width.
+
+@item out_h, oh
+Output height.
+
+@item in
+Input frame count.
+
+@item on
+Output frame count.
+
+@item x
+@item y
+Last calculated 'x' and 'y' position from 'x' and 'y' expression
+for current input frame.
+
+@item px
+@item py
+'x' and 'y' of last output frame of previous input frame or 0 when there was
+not yet such frame (first input frame).
+
+@item zoom
+Last calculated zoom from 'z' expression for current input frame.
+
+@item pzoom
+Last calculated zoom of last output frame of previous input frame.
+
+@item duration
+Number of output frames for current input frame. Calculated from 'd' expression
+for each input frame.
+
+@item pduration
+number of output frames created for previous input frame
+
+@item a
+Rational number: input width / input height
+
+@item sar
+sample aspect ratio
+
+@item dar
+display aspect ratio
+
+@end table
+
+@subsection Examples
+
+@itemize
+@item
+Zoom-in up to 1.5 and pan at same time to some spot near center of picture:
+@example
+zoompan=z=min(zoom+0.0015\,1.5):d=700:x=if(gte(zoom\,1.5)\,x\,x+1/a):y=if(gte(zoom\,1.5)\,y\,y+1):s=640x360"
+@end example
+@end itemize
+
 @c man end VIDEO FILTERS
 
 @chapter Video Sources
index 8ba0312..9552b9f 100644 (file)
@@ -195,6 +195,7 @@ OBJS-$(CONFIG_VIGNETTE_FILTER)               += vf_vignette.o
 OBJS-$(CONFIG_W3FDIF_FILTER)                 += vf_w3fdif.o
 OBJS-$(CONFIG_YADIF_FILTER)                  += vf_yadif.o
 OBJS-$(CONFIG_ZMQ_FILTER)                    += f_zmq.o
+OBJS-$(CONFIG_ZOOMPAN_FILTER)                += vf_zoompan.o
 
 OBJS-$(CONFIG_CELLAUTO_FILTER)               += vsrc_cellauto.o
 OBJS-$(CONFIG_COLOR_FILTER)                  += vsrc_testsrc.o
index 55d505a..9ef4f67 100644 (file)
@@ -212,6 +212,7 @@ void avfilter_register_all(void)
     REGISTER_FILTER(W3FDIF,         w3fdif,         vf);
     REGISTER_FILTER(YADIF,          yadif,          vf);
     REGISTER_FILTER(ZMQ,            zmq,            vf);
+    REGISTER_FILTER(ZOOMPAN,        zoompan,        vf);
 
     REGISTER_FILTER(CELLAUTO,       cellauto,       vsrc);
     REGISTER_FILTER(COLOR,          color,          vsrc);
index d7acf72..80a7b66 100644 (file)
@@ -30,7 +30,7 @@
 #include "libavutil/version.h"
 
 #define LIBAVFILTER_VERSION_MAJOR   4
-#define LIBAVFILTER_VERSION_MINOR   6
+#define LIBAVFILTER_VERSION_MINOR   7
 #define LIBAVFILTER_VERSION_MICRO 100
 
 #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \
diff --git a/libavfilter/vf_zoompan.c b/libavfilter/vf_zoompan.c
new file mode 100644 (file)
index 0000000..9ac925c
--- /dev/null
@@ -0,0 +1,310 @@
+/*
+ * Copyright (c) 2013 Paul B Mahol
+ *
+ * This file is part of FFmpeg.
+ *
+ * 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.
+ *
+ * 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
+ */
+
+#include "libavutil/eval.h"
+#include "libavutil/opt.h"
+#include "libavutil/pixdesc.h"
+#include "avfilter.h"
+#include "formats.h"
+#include "internal.h"
+#include "video.h"
+#include "libswscale/swscale.h"
+
+static const char *const var_names[] = {
+    "in_w",   "iw",
+    "in_h",   "ih",
+    "out_w",  "ow",
+    "out_h",  "oh",
+    "in",
+    "on",
+    "duration",
+    "pduration",
+    "time",
+    "frame",
+    "zoom",
+    "pzoom",
+    "x", "px",
+    "y", "py",
+    "a",
+    "sar",
+    "dar",
+    "hsub",
+    "vsub",
+    NULL
+};
+
+enum var_name {
+    VAR_IN_W,   VAR_IW,
+    VAR_IN_H,   VAR_IH,
+    VAR_OUT_W,  VAR_OW,
+    VAR_OUT_H,  VAR_OH,
+    VAR_IN,
+    VAR_ON,
+    VAR_DURATION,
+    VAR_PDURATION,
+    VAR_TIME,
+    VAR_FRAME,
+    VAR_ZOOM,
+    VAR_PZOOM,
+    VAR_X, VAR_PX,
+    VAR_Y, VAR_PY,
+    VAR_A,
+    VAR_SAR,
+    VAR_DAR,
+    VAR_HSUB,
+    VAR_VSUB,
+    VARS_NB
+};
+
+typedef struct ZPcontext {
+    const AVClass *class;
+    char *zoom_expr_str;
+    char *x_expr_str;
+    char *y_expr_str;
+    char *duration_expr_str;
+    int w, h;
+    double x, y;
+    double prev_zoom;
+    int prev_nb_frames;
+    struct SwsContext *sws;
+    int64_t frame_count;
+} ZPContext;
+
+#define OFFSET(x) offsetof(ZPContext, x)
+#define FLAGS AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_FILTERING_PARAM
+static const AVOption zoompan_options[] = {
+    { "zoom", "set the zoom expression", OFFSET(zoom_expr_str), AV_OPT_TYPE_STRING, {.str = "1" }, .flags = FLAGS },
+    { "z", "set the zoom expression", OFFSET(zoom_expr_str), AV_OPT_TYPE_STRING, {.str = "1" }, .flags = FLAGS },
+    { "x", "set the x expression", OFFSET(x_expr_str), AV_OPT_TYPE_STRING, {.str="0"}, .flags = FLAGS },
+    { "y", "set the y expression", OFFSET(y_expr_str), AV_OPT_TYPE_STRING, {.str="0"}, .flags = FLAGS },
+    { "d", "set the duration expression", OFFSET(duration_expr_str), AV_OPT_TYPE_STRING, {.str="90"}, .flags = FLAGS },
+    { "s", "set the output image size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="hd720"}, .flags = FLAGS },
+    { NULL }
+};
+
+AVFILTER_DEFINE_CLASS(zoompan);
+
+static av_cold int init(AVFilterContext *ctx)
+{
+    ZPContext *s = ctx->priv;
+
+    s->prev_zoom = 1;
+    return 0;
+}
+
+static int config_output(AVFilterLink *outlink)
+{
+    AVFilterContext *ctx = outlink->src;
+    ZPContext *s = ctx->priv;
+
+    outlink->w = s->w;
+    outlink->h = s->h;
+
+    return 0;
+}
+
+static int filter_frame(AVFilterLink *inlink, AVFrame *in)
+{
+    AVFilterContext *ctx = inlink->dst;
+    AVFilterLink *outlink = ctx->outputs[0];
+    ZPContext *s = ctx->priv;
+    double var_values[VARS_NB], nb_frames, zoom, dx, dy;
+    const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(in->format);
+    AVFrame *out;
+    int i, k, x, y, w, h, ret = 0;
+
+    var_values[VAR_IN_W]  = var_values[VAR_IW] = in->width;
+    var_values[VAR_IN_H]  = var_values[VAR_IH] = in->height;
+    var_values[VAR_OUT_W] = var_values[VAR_OW] = s->w;
+    var_values[VAR_OUT_H] = var_values[VAR_OH] = s->h;
+    var_values[VAR_IN]    = inlink->frame_count + 1;
+    var_values[VAR_ON]    = outlink->frame_count + 1;
+    var_values[VAR_PX]    = s->x;
+    var_values[VAR_PY]    = s->y;
+    var_values[VAR_X]     = 0;
+    var_values[VAR_Y]     = 0;
+    var_values[VAR_PZOOM] = s->prev_zoom;
+    var_values[VAR_ZOOM]  = 1;
+    var_values[VAR_PDURATION] = s->prev_nb_frames;
+    var_values[VAR_A]     = (double) in->width / in->height;
+    var_values[VAR_SAR]   = inlink->sample_aspect_ratio.num ?
+        (double) inlink->sample_aspect_ratio.num / inlink->sample_aspect_ratio.den : 1;
+    var_values[VAR_DAR]   = var_values[VAR_A] * var_values[VAR_SAR];
+    var_values[VAR_HSUB]  = 1 << desc->log2_chroma_w;
+    var_values[VAR_VSUB]  = 1 << desc->log2_chroma_h;
+
+    if ((ret = av_expr_parse_and_eval(&nb_frames, s->duration_expr_str,
+                                      var_names, var_values,
+                                      NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0)
+        goto fail;
+
+    var_values[VAR_DURATION] = nb_frames;
+    for (i = 0; i < nb_frames; i++) {
+        int px[4];
+        int py[4];
+        int64_t pts = av_rescale_q(in->pts, inlink->time_base,
+                                   outlink->time_base) + s->frame_count;
+
+        var_values[VAR_TIME] = pts * av_q2d(outlink->time_base);
+        var_values[VAR_FRAME] = i;
+        var_values[VAR_ON] = outlink->frame_count + 1;
+        if ((ret = av_expr_parse_and_eval(&zoom, s->zoom_expr_str,
+                                          var_names, var_values,
+                                          NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0)
+            goto fail;
+
+        zoom = av_clipd(zoom, 1, 10);
+        var_values[VAR_ZOOM] = zoom;
+        w = in->width * (1.0 / zoom);
+        h = in->height * (1.0 / zoom);
+
+        if ((ret = av_expr_parse_and_eval(&dx, s->x_expr_str,
+                                          var_names, var_values,
+                                          NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0)
+            goto fail;
+        x = dx = av_clipd(dx, 0, FFMAX(in->width - w, 0));
+        var_values[VAR_X] = dx;
+        x &= ~((1 << desc->log2_chroma_w) - 1);
+
+        if ((ret = av_expr_parse_and_eval(&dy, s->y_expr_str,
+                                          var_names, var_values,
+                                          NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0)
+            goto fail;
+        y = dy = av_clipd(dy, 0, FFMAX(in->height - h, 0));
+        var_values[VAR_Y] = dy;
+        y &= ~((1 << desc->log2_chroma_h) - 1);
+
+        out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
+        if (!out) {
+            ret = AVERROR(ENOMEM);
+            goto fail;
+        }
+
+        px[1] = px[2] = FF_CEIL_RSHIFT(x, desc->log2_chroma_w);
+        px[0] = px[3] = x;
+
+        py[1] = py[2] = FF_CEIL_RSHIFT(y, desc->log2_chroma_h);
+        py[0] = py[3] = y;
+
+        s->sws = sws_alloc_context();
+        if (!s->sws) {
+            ret = AVERROR(ENOMEM);
+            goto fail;
+        }
+
+        uint8_t *input[4];
+
+        for (k = 0; in->data[k]; k++)
+            input[k] = in->data[k] + py[k] * in->linesize[k] + px[k];
+
+        av_opt_set_int(s->sws, "srcw", w, 0);
+        av_opt_set_int(s->sws, "srch", h, 0);
+        av_opt_set_int(s->sws, "src_format", in->format, 0);
+        av_opt_set_int(s->sws, "dstw", outlink->w, 0);
+        av_opt_set_int(s->sws, "dsth", outlink->h, 0);
+        av_opt_set_int(s->sws, "dst_format", outlink->format, 0);
+        av_opt_set_int(s->sws, "sws_flags", SWS_BICUBIC, 0);
+
+        if ((ret = sws_init_context(s->sws, NULL, NULL)) < 0)
+            goto fail;
+
+        sws_scale(s->sws, (const uint8_t *const *)&input, in->linesize, 0, h, out->data, out->linesize);
+
+        out->pts = pts;
+        s->frame_count++;
+
+        ret = ff_filter_frame(outlink, out);
+        if (ret < 0)
+            break;
+
+        sws_freeContext(s->sws);
+        s->sws = NULL;
+    }
+
+    s->x = dx;
+    s->y = dy;
+    s->prev_zoom = zoom;
+    s->prev_nb_frames = nb_frames;
+
+fail:
+    sws_freeContext(s->sws);
+    s->sws = NULL;
+    av_frame_free(&in);
+    return ret;
+}
+
+static int query_formats(AVFilterContext *ctx)
+{
+    static const enum AVPixelFormat pix_fmts[] = {
+        AV_PIX_FMT_YUV444P,  AV_PIX_FMT_YUV422P,
+        AV_PIX_FMT_YUV420P,  AV_PIX_FMT_YUV411P,
+        AV_PIX_FMT_YUV410P,  AV_PIX_FMT_YUV440P,
+        AV_PIX_FMT_YUVA444P, AV_PIX_FMT_YUVA422P,
+        AV_PIX_FMT_YUVA420P,
+        AV_PIX_FMT_YUVJ444P, AV_PIX_FMT_YUVJ440P,
+        AV_PIX_FMT_YUVJ422P, AV_PIX_FMT_YUVJ420P,
+        AV_PIX_FMT_YUVJ411P,
+        AV_PIX_FMT_GRAY8,
+        AV_PIX_FMT_NONE
+    };
+
+
+    ff_set_common_formats(ctx, ff_make_format_list(pix_fmts));
+    return 0;
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+    ZPContext *s = ctx->priv;
+
+    sws_freeContext(s->sws);
+    s->sws = NULL;
+}
+
+static const AVFilterPad inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_VIDEO,
+        .filter_frame = filter_frame,
+    },
+    { NULL }
+};
+
+static const AVFilterPad outputs[] = {
+    {
+        .name          = "default",
+        .type          = AVMEDIA_TYPE_VIDEO,
+        .config_props  = config_output,
+    },
+    { NULL }
+};
+
+AVFilter ff_vf_zoompan = {
+    .name          = "zoompan",
+    .description   = NULL_IF_CONFIG_SMALL("Apply Zoom & Pan effect."),
+    .priv_size     = sizeof(ZPContext),
+    .priv_class    = &zoompan_class,
+    .init          = init,
+    .uninit        = uninit,
+    .query_formats = query_formats,
+    .inputs        = inputs,
+    .outputs       = outputs,
+    .flags         = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC,
+};