avfilter: add audio compressor filter

Signed-off-by: Paul B Mahol <onemda@gmail.com>

Changelog

@@ -35,6 +35,7 @@ version <next>:
 - anoisesrc audio filter source
 - IVR demuxer
 - compensationdelay filter
+- acompressor filter
 
 
 version 2.8:

doc/filters.texi

@@ -318,6 +318,78 @@ build.
 
 Below is a description of the currently available audio filters.
 
+@section acompressor
+
+A compressor is mainly used to reduce the dynamic range of a signal.
+Especially modern music is mostly compressed at a high ratio to
+improve the overall loudness. It's done to get the highest attention
+of a listener, "fatten" the sound and bring more "power" to the track.
+If a signal is compressed too much it may sound dull or "dead"
+afterwards or it may start to "pump" (which could be a powerful effect
+but can also destroy a track completely).
+The right compression is the key to reach a professional sound and is
+the high art of mixing and mastering. Because of its complex settings
+it may take a long time to get the right feeling for this kind of effect.
+
+Compression is done by detecting the volume above a chosen level
+@code{threshold} and dividing it by the factor set with @code{ratio}.
+So if you set the threshold to -12dB and your signal reaches -6dB a ratio
+of 2:1 will result in a signal at -9dB. Because an exact manipulation of
+the signal would cause distortion of the waveform the reduction can be
+levelled over the time. This is done by setting "Attack" and "Release".
+@code{attack} determines how long the signal has to rise above the threshold
+before any reduction will occur and @code{release} sets the time the signal
+has to fall below the threshold to reduce the reduction again. Shorter signals
+than the chosen attack time will be left untouched.
+The overall reduction of the signal can be made up afterwards with the
+@code{makeup} setting. So compressing the peaks of a signal about 6dB and
+raising the makeup to this level results in a signal twice as loud than the
+source. To gain a softer entry in the compression the @code{knee} flattens the
+hard edge at the threshold in the range of the chosen decibels.
+
+The filter accepts the following options:
+
+@table @option
+@item threshold
+If a signal of second stream rises above this level it will affect the gain
+reduction of the first stream.
+By default it is 0.125. Range is between 0.00097563 and 1.
+
+@item ratio
+Set a ratio by which the signal is reduced. 1:2 means that if the level
+rose 4dB above the threshold, it will be only 2dB above after the reduction.
+Default is 2. Range is between 1 and 20.
+
+@item attack
+Amount of milliseconds the signal has to rise above the threshold before gain
+reduction starts. Default is 20. Range is between 0.01 and 2000.
+
+@item release
+Amount of milliseconds the signal has to fall below the threshold before
+reduction is decreased again. Default is 250. Range is between 0.01 and 9000.
+
+@item makeup
+Set the amount by how much signal will be amplified after processing.
+Default is 2. Range is from 1 and 64.
+
+@item knee
+Curve the sharp knee around the threshold to enter gain reduction more softly.
+Default is 2.82843. Range is between 1 and 8.
+
+@item link
+Choose if the @code{average} level between all channels of input stream
+or the louder(@code{maximum}) channel of input stream affects the
+reduction. Default is @code{average}.
+
+@item detection
+Should the exact signal be taken in case of @code{peak} or an RMS one in case
+of @code{rms}. Default is @code{rms} which is mostly smoother.
+
+@item mix
+How much to use compressed signal in output. Default is 1.
+Range is between 0 and 1.
+@end table
+
 @section acrossfade
 
 Apply cross fade from one input audio stream to another input audio stream.

libavfilter/Makefile

@@ -23,6 +23,7 @@ OBJS = allfilters.o                                                     \
        transform.o                                                      \
        video.o                                                          \
 
+OBJS-$(CONFIG_ACOMPRESSOR_FILTER)            += af_sidechaincompress.o
 OBJS-$(CONFIG_ACROSSFADE_FILTER)             += af_afade.o
 OBJS-$(CONFIG_ADELAY_FILTER)                 += af_adelay.o
 OBJS-$(CONFIG_AECHO_FILTER)                  += af_aecho.o

libavfilter/af_sidechaincompress.c

@@ -21,7 +21,7 @@
 
 /**
  * @file
- * Sidechain compressor filter
+ * Audio (Sidechain) Compressor filter
  */
 
 #include "libavutil/avassert.h"
@@ -61,7 +61,7 @@ typedef struct SidechainCompressContext {
 #define A AV_OPT_FLAG_AUDIO_PARAM
 #define F AV_OPT_FLAG_FILTERING_PARAM
 
-static const AVOption sidechaincompress_options[] = {
+static const AVOption options[] = {
     { "threshold", "set threshold",    OFFSET(threshold), AV_OPT_TYPE_DOUBLE, {.dbl=0.125}, 0.000976563,    1, A|F },
     { "ratio",     "set ratio",        OFFSET(ratio),     AV_OPT_TYPE_DOUBLE, {.dbl=2},               1,   20, A|F },
     { "attack",    "set attack",       OFFSET(attack),    AV_OPT_TYPE_DOUBLE, {.dbl=20},           0.01, 2000, A|F },
@@ -78,6 +78,7 @@ static const AVOption sidechaincompress_options[] = {
     { NULL }
 };
 
+#define sidechaincompress_options options
 AVFILTER_DEFINE_CLASS(sidechaincompress);
 
 static av_cold int init(AVFilterContext *ctx)
@@ -126,33 +127,24 @@ static double output_gain(double lin_slope, double ratio, double thres,
     return exp(gain - slope);
 }
 
-static int filter_frame(AVFilterLink *link, AVFrame *frame)
+static int compressor_config_output(AVFilterLink *outlink)
 {
-    AVFilterContext *ctx = link->dst;
+    AVFilterContext *ctx = outlink->src;
     SidechainCompressContext *s = ctx->priv;
-    AVFilterLink *sclink = ctx->inputs[1];
-    AVFilterLink *outlink = ctx->outputs[0];
-    const double makeup = s->makeup;
-    const double mix = s->mix;
-    const double *scsrc;
-    double *sample;
-    int nb_samples;
-    int ret, i, c;
 
-    for (i = 0; i < 2; i++)
-        if (link == ctx->inputs[i])
-            break;
-    av_assert0(i < 2 && !s->input_frame[i]);
-    s->input_frame[i] = frame;
+    s->attack_coeff = FFMIN(1., 1. / (s->attack * outlink->sample_rate / 4000.));
+    s->release_coeff = FFMIN(1., 1. / (s->release * outlink->sample_rate / 4000.));
 
-    if (!s->input_frame[0] || !s->input_frame[1])
     return 0;
+}
 
-    nb_samples = FFMIN(s->input_frame[0]->nb_samples,
-                       s->input_frame[1]->nb_samples);
-
-    sample = (double *)s->input_frame[0]->data[0];
-    scsrc = (const double *)s->input_frame[1]->data[0];
+static void compressor(SidechainCompressContext *s,
+                       double *sample, const double *scsrc, int nb_samples,
+                       AVFilterLink *inlink, AVFilterLink *sclink)
+{
+    const double makeup = s->makeup;
+    const double mix = s->mix;
+    int i, c;
 
     for (i = 0; i < nb_samples; i++) {
         double abs_sample, gain = 1.0;
@@ -179,13 +171,42 @@ static int filter_frame(AVFilterLink *link, AVFrame *frame)
                                s->knee_start, s->knee_stop,
                                s->compressed_knee_stop, s->detection);
 
-        for (c = 0; c < outlink->channels; c++)
+        for (c = 0; c < inlink->channels; c++)
             sample[c] *= (gain * makeup * mix + (1. - mix));
 
-        sample += outlink->channels;
+        sample += inlink->channels;
         scsrc += sclink->channels;
     }
+}
+
+#if CONFIG_SIDECHAINCOMPRESS_FILTER
+static int filter_frame(AVFilterLink *link, AVFrame *frame)
+{
+    AVFilterContext *ctx = link->dst;
+    SidechainCompressContext *s = ctx->priv;
+    AVFilterLink *outlink = ctx->outputs[0];
+    const double *scsrc;
+    double *sample;
+    int nb_samples;
+    int ret, i;
+
+    for (i = 0; i < 2; i++)
+        if (link == ctx->inputs[i])
+            break;
+    av_assert0(i < 2 && !s->input_frame[i]);
+    s->input_frame[i] = frame;
+
+    if (!s->input_frame[0] || !s->input_frame[1])
+        return 0;
+
+    nb_samples = FFMIN(s->input_frame[0]->nb_samples,
+                       s->input_frame[1]->nb_samples);
+
+    sample = (double *)s->input_frame[0]->data[0];
+    scsrc = (const double *)s->input_frame[1]->data[0];
 
+    compressor(s, sample, scsrc, nb_samples,
+               ctx->inputs[0], ctx->inputs[1]);
     ret = ff_filter_frame(outlink, s->input_frame[0]);
 
     s->input_frame[0] = NULL;
@@ -253,7 +274,6 @@ static int query_formats(AVFilterContext *ctx)
 static int config_output(AVFilterLink *outlink)
 {
     AVFilterContext *ctx = outlink->src;
-    SidechainCompressContext *s = ctx->priv;
 
     if (ctx->inputs[0]->sample_rate != ctx->inputs[1]->sample_rate) {
         av_log(ctx, AV_LOG_ERROR,
@@ -268,8 +288,7 @@ static int config_output(AVFilterLink *outlink)
     outlink->channel_layout = ctx->inputs[0]->channel_layout;
     outlink->channels = ctx->inputs[0]->channels;
 
-    s->attack_coeff = FFMIN(1., 1. / (s->attack * outlink->sample_rate / 4000.));
-    s->release_coeff = FFMIN(1., 1. / (s->release * outlink->sample_rate / 4000.));
+    compressor_config_output(outlink);
 
     return 0;
 }
@@ -310,3 +329,83 @@ AVFilter ff_af_sidechaincompress = {
     .inputs         = sidechaincompress_inputs,
     .outputs        = sidechaincompress_outputs,
 };
+#endif  /* CONFIG_SIDECHAINCOMPRESS_FILTER */
+
+#if CONFIG_ACOMPRESSOR_FILTER
+static int acompressor_filter_frame(AVFilterLink *inlink, AVFrame *frame)
+{
+    AVFilterContext *ctx = inlink->dst;
+    SidechainCompressContext *s = ctx->priv;
+    AVFilterLink *outlink = ctx->outputs[0];
+    double *sample;
+
+    sample = (double *)frame->data[0];
+    compressor(s, sample, sample, frame->nb_samples,
+               inlink, inlink);
+
+    return ff_filter_frame(outlink, frame);
+}
+
+static int acompressor_query_formats(AVFilterContext *ctx)
+{
+    AVFilterFormats *formats;
+    AVFilterChannelLayouts *layouts;
+    static const enum AVSampleFormat sample_fmts[] = {
+        AV_SAMPLE_FMT_DBL,
+        AV_SAMPLE_FMT_NONE
+    };
+    int ret;
+
+    layouts = ff_all_channel_counts();
+    if (!layouts)
+        return AVERROR(ENOMEM);
+    ret = ff_set_common_channel_layouts(ctx, layouts);
+    if (ret < 0)
+        return ret;
+
+    formats = ff_make_format_list(sample_fmts);
+    if (!formats)
+        return AVERROR(ENOMEM);
+    ret = ff_set_common_formats(ctx, formats);
+    if (ret < 0)
+        return ret;
+
+    formats = ff_all_samplerates();
+    if (!formats)
+        return AVERROR(ENOMEM);
+    return ff_set_common_samplerates(ctx, formats);
+}
+
+#define acompressor_options options
+AVFILTER_DEFINE_CLASS(acompressor);
+
+static const AVFilterPad acompressor_inputs[] = {
+    {
+        .name           = "default",
+        .type           = AVMEDIA_TYPE_AUDIO,
+        .filter_frame   = acompressor_filter_frame,
+        .needs_writable = 1,
+    },
+    { NULL }
+};
+
+static const AVFilterPad acompressor_outputs[] = {
+    {
+        .name          = "default",
+        .type          = AVMEDIA_TYPE_AUDIO,
+        .config_props  = compressor_config_output,
+    },
+    { NULL }
+};
+
+AVFilter ff_af_acompressor = {
+    .name           = "acompressor",
+    .description    = NULL_IF_CONFIG_SMALL("Audio compressor."),
+    .priv_size      = sizeof(SidechainCompressContext),
+    .priv_class     = &acompressor_class,
+    .init           = init,
+    .query_formats  = acompressor_query_formats,
+    .inputs         = acompressor_inputs,
+    .outputs        = acompressor_outputs,
+};
+#endif  /* CONFIG_ACOMPRESSOR_FILTER */

libavfilter/allfilters.c

@@ -45,6 +45,7 @@ void avfilter_register_all(void)
         return;
     initialized = 1;
 
+    REGISTER_FILTER(ACOMPRESSOR,    acompressor,    af);
     REGISTER_FILTER(ACROSSFADE,     acrossfade,     af);
     REGISTER_FILTER(ADELAY,         adelay,         af);
     REGISTER_FILTER(AECHO,          aecho,          af);

libavfilter/version.h

@@ -30,7 +30,7 @@
 #include "libavutil/version.h"
 
 #define LIBAVFILTER_VERSION_MAJOR   6
-#define LIBAVFILTER_VERSION_MINOR  16
+#define LIBAVFILTER_VERSION_MINOR  17
 #define LIBAVFILTER_VERSION_MICRO 100
 
 #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \