lavfi: add earwax audio filter, ported from Sox
authorMina Nagy Zaki <mnzaki@gmail.com>
Sat, 26 Mar 2011 20:02:50 +0000 (21:02 +0100)
committerStefano Sabatini <stefasab@gmail.com>
Sat, 5 Nov 2011 01:16:45 +0000 (02:16 +0100)
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>
Changelog
doc/filters.texi
libavfilter/Makefile
libavfilter/af_earwax.c [new file with mode: 0644]
libavfilter/allfilters.c
libavfilter/avfilter.h

index f35477e..7bc4764 100644 (file)
--- a/Changelog
+++ b/Changelog
@@ -74,6 +74,7 @@ easier to use. The changes are:
 - replacement Indeo 3 decoder
 - new ffmpeg option: -map_channel
 - volume audio filter added
+- earwax audio filter added
 
 
 version 0.8:
index d21ddf1..92ef589 100644 (file)
@@ -224,6 +224,17 @@ expressed in the form "[@var{c0} @var{c1} @var{c2} @var{c3} @var{c4} @var{c5}
 @var{c6} @var{c7}]"
 @end table
 
+@section earwax
+
+Make audio easier to listen to on headphones.
+
+This filter adds `cues' to 44.1kHz stereo (i.e. audio CD format) audio
+so that when listened to on headphones the stereo image is moved from
+inside your head (standard for headphones) to outside and in front of
+the listener (standard for speakers).
+
+Ported from SoX.
+
 @section volume
 
 Adjust the input audio volume.
index edfb12f..bb30ccb 100644 (file)
@@ -28,6 +28,7 @@ OBJS-$(CONFIG_AFORMAT_FILTER)                += af_aformat.o
 OBJS-$(CONFIG_ANULL_FILTER)                  += af_anull.o
 OBJS-$(CONFIG_ARESAMPLE_FILTER)              += af_aresample.o
 OBJS-$(CONFIG_ASHOWINFO_FILTER)              += af_ashowinfo.o
+OBJS-$(CONFIG_EARWAX_FILTER)                 += af_earwax.o
 OBJS-$(CONFIG_VOLUME_FILTER)                 += af_volume.o
 
 OBJS-$(CONFIG_ABUFFER_FILTER)                += asrc_abuffer.o
diff --git a/libavfilter/af_earwax.c b/libavfilter/af_earwax.c
new file mode 100644 (file)
index 0000000..4e14c9f
--- /dev/null
@@ -0,0 +1,161 @@
+/*
+ * Copyright (c) 2011 Mina Nagy Zaki
+ * Copyright (c) 2000 Edward Beingessner And Sundry Contributors.
+ * This source code is freely redistributable and may be used for any purpose.
+ * This copyright notice must be maintained.  Edward Beingessner And Sundry
+ * Contributors are not responsible for the consequences of using this
+ * software.
+ *
+ * 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
+ */
+
+/**
+ * @file
+ * Stereo Widening Effect. Adds audio cues to move stereo image in
+ * front of the listener. Adapted from the libsox earwax effect.
+ */
+
+#include "libavutil/audioconvert.h"
+#include "avfilter.h"
+
+#define NUMTAPS 64
+
+static const int8_t filt[NUMTAPS] = {
+/* 30°  330° */
+    4,   -6,     /* 32 tap stereo FIR filter. */
+    4,  -11,     /* One side filters as if the */
+   -1,   -5,     /* signal was from 30 degrees */
+    3,    3,     /* from the ear, the other as */
+   -2,    5,     /* if 330 degrees. */
+   -5,    0,
+    9,    1,
+    6,    3,     /*                         Input                         */
+   -4,   -1,     /*                   Left         Right                  */
+   -5,   -3,     /*                __________   __________                */
+   -2,   -5,     /*               |          | |          |               */
+   -7,    1,     /*           .---|  Hh,0(f) | |  Hh,0(f) |---.           */
+    6,   -7,     /*          /    |__________| |__________|    \          */
+   30,  -29,     /*         /                \ /                \         */
+   12,   -3,     /*        /                  X                  \        */
+  -11,    4,     /*       /                  / \                  \       */
+   -3,    7,     /*  ____V_____   __________V   V__________   _____V____  */
+  -20,   23,     /* |          | |          |   |          | |          | */
+    2,    0,     /* | Hh,30(f) | | Hh,330(f)|   | Hh,330(f)| | Hh,30(f) | */
+    1,   -6,     /* |__________| |__________|   |__________| |__________| */
+  -14,   -5,     /*      \     ___      /           \      ___     /      */
+   15,  -18,     /*       \   /   \    /    _____    \    /   \   /       */
+    6,    7,     /*        `->| + |<--'    /     \    `-->| + |<-'        */
+   15,  -10,     /*           \___/      _/       \_      \___/           */
+  -14,   22,     /*               \     / \       / \     /               */
+   -7,   -2,     /*                `--->| |       | |<---'                */
+   -4,    9,     /*                     \_/       \_/                     */
+    6,  -12,     /*                                                       */
+    6,   -6,     /*                       Headphones                      */
+    0,  -11,
+    0,   -5,
+    4,    0};
+
+typedef struct {
+    int16_t taps[NUMTAPS * 2];
+} EarwaxContext;
+
+static int query_formats(AVFilterContext *ctx)
+{
+    AVFilterFormats *formats = NULL;
+    avfilter_add_format(&formats, AV_SAMPLE_FMT_S16);
+    avfilter_set_common_sample_formats(ctx, formats);
+    formats = NULL;
+    avfilter_add_format(&formats, AV_CH_LAYOUT_STEREO);
+    avfilter_set_common_channel_layouts(ctx, formats);
+    formats = NULL;
+    avfilter_add_format(&formats, AVFILTER_PACKED);
+    avfilter_set_common_packing_formats(ctx, formats);
+
+    return 0;
+}
+
+static int config_input(AVFilterLink *inlink)
+{
+    if (inlink->sample_rate != 44100) {
+        av_log(inlink->dst, AV_LOG_ERROR,
+               "The earwax filter only works for 44.1kHz audio. Insert "
+               "a resample filter before this\n");
+        return AVERROR(EINVAL);
+    }
+    return 0;
+}
+
+//FIXME: replace with DSPContext.scalarproduct_int16
+static inline int16_t *scalarproduct(const int16_t *in, const int16_t *endin, int16_t *out)
+{
+    int32_t sample;
+    int16_t j;
+
+    while (in < endin) {
+        sample = 32;
+        for (j = 0; j < NUMTAPS; j++)
+            sample += in[j] * filt[j];
+        *out = sample >> 6;
+        out++;
+        in++;
+    }
+
+    return out;
+}
+
+static void filter_samples(AVFilterLink *inlink, AVFilterBufferRef *insamples)
+{
+    AVFilterLink *outlink = inlink->dst->outputs[0];
+    int16_t *taps, *endin, *in, *out;
+    AVFilterBufferRef *outsamples =
+        avfilter_get_audio_buffer(inlink, AV_PERM_WRITE,
+                                  insamples->audio->nb_samples);
+    taps  = ((EarwaxContext *)inlink->dst->priv)->taps;
+    out   = (int16_t *)outsamples->data[0];
+    in    = (int16_t *)insamples ->data[0];
+
+    // copy part of new input and process with saved input
+    memcpy(taps+NUMTAPS, in, NUMTAPS * sizeof(*taps));
+    out   = scalarproduct(taps, taps + NUMTAPS, out);
+
+    // process current input
+    endin = in + insamples->audio->nb_samples * 2 - NUMTAPS;
+    out   = scalarproduct(in, endin, out);
+
+    // save part of input for next round
+    memcpy(taps, endin, NUMTAPS * sizeof(*taps));
+
+    avfilter_filter_samples(outlink, outsamples);
+    avfilter_unref_buffer(insamples);
+}
+
+AVFilter avfilter_af_earwax = {
+    .name           = "earwax",
+    .description    = NULL_IF_CONFIG_SMALL("Widen the stereo image."),
+    .query_formats  = query_formats,
+    .priv_size      = sizeof(EarwaxContext),
+    .inputs  = (AVFilterPad[])  {{  .name           = "default",
+                                    .type           = AVMEDIA_TYPE_AUDIO,
+                                    .filter_samples = filter_samples,
+                                    .config_props   = config_input,
+                                    .min_perms      = AV_PERM_READ, },
+                                 {  .name = NULL}},
+
+    .outputs = (AVFilterPad[])  {{  .name           = "default",
+                                    .type           = AVMEDIA_TYPE_AUDIO, },
+                                 {  .name = NULL}},
+};
index e80fc17..de54bbb 100644 (file)
@@ -39,6 +39,7 @@ void avfilter_register_all(void)
     REGISTER_FILTER (ANULL,       anull,       af);
     REGISTER_FILTER (ARESAMPLE,   aresample,   af);
     REGISTER_FILTER (ASHOWINFO,   ashowinfo,   af);
+    REGISTER_FILTER (EARWAX,      earwax,      af);
     REGISTER_FILTER (VOLUME,      volume,      af);
 
     REGISTER_FILTER (ABUFFER,     abuffer,     asrc);
index 4021522..b8205fd 100644 (file)
@@ -29,7 +29,7 @@
 #include "libavutil/rational.h"
 
 #define LIBAVFILTER_VERSION_MAJOR  2
-#define LIBAVFILTER_VERSION_MINOR 46
+#define LIBAVFILTER_VERSION_MINOR 47
 #define LIBAVFILTER_VERSION_MICRO  0
 
 #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \