[FFmpeg-devel] [PATCH] avfilter: add audio dynamic equalizer filter

Paul B Mahol onemda at gmail.com
Sun Nov 28 15:21:08 EET 2021


Signed-off-by: Paul B Mahol <onemda at gmail.com>
---
 doc/filters.texi                   |  75 +++++++
 libavfilter/Makefile               |   1 +
 libavfilter/af_adynamicequalizer.c | 310 +++++++++++++++++++++++++++++
 libavfilter/allfilters.c           |   1 +
 4 files changed, 387 insertions(+)
 create mode 100644 libavfilter/af_adynamicequalizer.c

diff --git a/doc/filters.texi b/doc/filters.texi
index 7852948d2f..cfddefff30 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -843,6 +843,81 @@ Compute derivative/integral of audio stream.
 
 Applying both filters one after another produces original audio.
 
+ at section adynamicequalizer
+
+Apply dynamic equalization to input audio stream.
+
+A description of the accepted options follows.
+
+ at table @option
+ at item threshold
+Set the detection threshold used to trigger equalization.
+Threshold detection is using bandpass filter.
+Default value is 0. Allowed range is from 0 to 50.
+
+ at item dfrequency
+Set the detection frequency in Hz used for bandpass filter used to trigger equalization.
+Default value is 1000 Hz. Allowed range is between 2 and 1000000 Hz.
+
+ at item dqfactor
+Set the detection resonance factor for bandpass filter used to trigger equalization.
+Default value is 1. Allowed range is from 0.001 to 1000.
+
+ at item sfrequency
+Set the frequency of lowpass filter to smooth equalization filter gains.
+Default value is 100 Hz. Allowed range is between 2 and 1000000 Hz.
+
+ at item sqfactor
+Set the resonance factor of lowpass filter to smooth equalization filter gains.
+Default value is 1. Allowed range is from 0.001 to 1000.
+
+ at item tfrequency
+Set the target frequency of equalization filter.
+Default value is 1000 Hz. Allowed range is between 2 and 1000000 Hz.
+
+ at item tqfactor
+Set the target resonance factor for target equalization filter.
+Default value is 1. Allowed range is from 0.001 to 1000.
+
+ at item attack
+Set the amount of milliseconds the signal from detection has to rise above
+the detection threshold before equalization starts.
+Default is 10. Allowed range is between 1 and 2000.
+
+ at item release
+Set the amount of milliseconds the signal from detection has to fall below the
+detection threshold before equalization ends.
+Default is 80. Allowed range is between 1 and 2000.
+
+ at item knee
+Curve the sharp knee around the detection threshold to calculate
+equalization gain more softly.
+Default is 2. Allowed range is between 1 and 8.
+
+ at item ratio
+Set the ratio by which the equalization gain is raised.
+Default is 1. Allowed range is between 1 and 20.
+
+ at item range
+Set the max allowed cut/boost amount. Default is 0.06125.
+Allowed range is from 0.00000001 to 1.
+
+ at item mode
+Set the mode of filter operation, can be one of the following:
+
+ at table @samp
+ at item cut
+Cut frequencies above detection threshold.
+ at item boost
+Boost frequencies bellow detection threshold.
+ at end table
+Default mode is @samp{cut}.
+ at end table
+
+ at subsection Commands
+
+This filter supports the all above options as @ref{commands}.
+
 @section adynamicsmooth
 
 Apply dynamic smoothing to input audio stream.
diff --git a/libavfilter/Makefile b/libavfilter/Makefile
index c8082c4a2f..d40be4b252 100644
--- a/libavfilter/Makefile
+++ b/libavfilter/Makefile
@@ -44,6 +44,7 @@ OBJS-$(CONFIG_ADECORRELATE_FILTER)           += af_adecorrelate.o
 OBJS-$(CONFIG_ADELAY_FILTER)                 += af_adelay.o
 OBJS-$(CONFIG_ADENORM_FILTER)                += af_adenorm.o
 OBJS-$(CONFIG_ADERIVATIVE_FILTER)            += af_aderivative.o
+OBJS-$(CONFIG_ADYNAMICEQUALIZER_FILTER)      += af_adynamicequalizer.o
 OBJS-$(CONFIG_ADYNAMICSMOOTH_FILTER)         += af_adynamicsmooth.o
 OBJS-$(CONFIG_AECHO_FILTER)                  += af_aecho.o
 OBJS-$(CONFIG_AEMPHASIS_FILTER)              += af_aemphasis.o
diff --git a/libavfilter/af_adynamicequalizer.c b/libavfilter/af_adynamicequalizer.c
new file mode 100644
index 0000000000..da3d96d7cf
--- /dev/null
+++ b/libavfilter/af_adynamicequalizer.c
@@ -0,0 +1,310 @@
+/*
+ * 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/ffmath.h"
+#include "libavutil/opt.h"
+#include "avfilter.h"
+#include "audio.h"
+#include "formats.h"
+#include "hermite.h"
+
+typedef struct AudioDynamicEqualizerContext {
+    const AVClass *class;
+
+    double threshold;
+    double lin_threshold;
+    double dfrequency;
+    double dqfactor;
+    double tfrequency;
+    double tqfactor;
+    double sfrequency;
+    double sqfactor;
+    double attack;
+    double release;
+    double knee;
+    double ratio;
+    double range;
+    double knee_sqrt;
+    double attack_coeff;
+    double release_coeff;
+    double lin_knee_start;
+    double lin_knee_stop;
+    double knee_start;
+    double knee_stop;
+    double compressed_knee_start;
+    int mode;
+
+    AVFrame *state;
+} AudioDynamicEqualizerContext;
+
+static int config_input(AVFilterLink *inlink)
+{
+    AVFilterContext *ctx = inlink->dst;
+    AudioDynamicEqualizerContext *s = ctx->priv;
+
+    s->state = ff_get_audio_buffer(inlink, 7);
+    if (!s->state)
+        return AVERROR(ENOMEM);
+
+    return 0;
+}
+
+static double get_gain(AVFilterContext *ctx, double in,
+                       double sample_rate, int ch)
+{
+    AudioDynamicEqualizerContext *s = ctx->priv;
+    double *state = (double *)s->state->extended_data[ch];
+    const double ratio = s->ratio;
+    const double knee = s->knee;
+    const double attack_coeff  = s->attack_coeff;
+    const double release_coeff = s->release_coeff;
+    const double lin_knee_start = s->lin_knee_start;
+    const double threshold = s->threshold;
+    const double knee_start = s->knee_start;
+    const double knee_stop = s->knee_stop;
+    const double compressed_knee_start = s->compressed_knee_start;
+    const double abs_sample = fabs(in);
+    const int mode = s->mode;
+    double lin_slope = state[2];
+    double range = s->range;
+    double gain = 0.;
+    double slope;
+    int detected;
+
+    lin_slope += (abs_sample - lin_slope) * (abs_sample > lin_slope ? attack_coeff : release_coeff);
+
+    detected = lin_slope > lin_knee_start;
+
+    state[2] = lin_slope;
+    if (lin_slope <= 0. || !detected)
+        return 1.;
+
+    slope = log(lin_slope);
+    gain = (slope - threshold) * ratio + threshold;
+
+    if (knee >= 1.)
+        gain = hermite_interpolation(slope, knee_stop, knee_start,
+                                     knee_stop, compressed_knee_start,
+                                     1., ratio);
+    if (!mode)
+        gain = -gain;
+
+    gain = exp(gain - slope);
+    gain = mode ? av_clipd(gain, 1., 1. / range) : av_clipd(gain, range, 1.);
+
+    return gain;
+}
+
+static double get_svf(double in, double *m, double *a, double *b)
+{
+    const double v0 = in;
+    const double v3 = v0 - b[1];
+    const double v1 = a[0] * b[0] + a[1] * v3;
+    const double v2 = b[1] + a[1] * b[0] + a[2] * v3;
+
+    b[0] = 2. * v1 - b[0];
+    b[1] = 2. * v2 - b[1];
+
+    return m[0] * v0 + m[1] * v1 + m[2] * v2;
+}
+
+typedef struct ThreadData {
+    AVFrame *in, *out;
+} ThreadData;
+
+static int filter_channels(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
+{
+    AudioDynamicEqualizerContext *s = ctx->priv;
+    ThreadData *td = arg;
+    AVFrame *in = td->in;
+    AVFrame *out = td->out;
+    const double sample_rate = in->sample_rate;
+    const double range = s->range;
+    const double irange = 1. / range;
+    const double dfrequency = fmin(s->dfrequency, sample_rate * 0.5);
+    const double tfrequency = fmin(s->tfrequency, sample_rate * 0.5);
+    const double sfrequency = fmin(s->sfrequency, sample_rate * 0.5);
+    const double dqfactor = s->dqfactor;
+    const double tqfactor = s->tqfactor;
+    const double sqfactor = s->tqfactor;
+    const double fg = tan(M_PI * tfrequency / sample_rate);
+    const double dg = tan(M_PI * dfrequency / sample_rate);
+    const double sg = tan(M_PI * sfrequency / sample_rate);
+    const int start = (in->channels * jobnr) / nb_jobs;
+    const int end = (in->channels * (jobnr+1)) / nb_jobs;
+    const int mode = s->mode;
+    double da[3], dm[3];
+
+    {
+        double k = 1. / dqfactor;
+
+        da[0] = 1. / (1. + dg * (dg + k));
+        da[1] = dg * da[0];
+        da[2] = dg * da[1];
+
+        dm[0] = 0.;
+        dm[1] = 1.;
+        dm[2] = 0.;
+    }
+
+    for (int ch = start; ch < end; ch++) {
+        const double *src = (const double *)in->extended_data[ch];
+        double *dst = (double *)out->extended_data[ch];
+        double *state = (double *)s->state->extended_data[ch];
+
+        for (int n = 0; n < out->nb_samples; n++) {
+            double fa[3], fm[3], sa[3], sm[3];
+            double detect, gain, v;
+
+            detect = get_svf(src[n], dm, da, state);
+            gain = get_gain(ctx, detect, sample_rate, ch);
+
+            {
+                double k = 1. / sqfactor;
+
+                sa[0] = 1. / (1. + sg * (sg + k));
+                sa[1] = sg * sa[0];
+                sa[2] = sg * sa[1];
+
+                sm[0] = 0.;
+                sm[1] = 0.;
+                sm[2] = 1.;
+            }
+
+            gain = get_svf(gain, sm, sa, &state[5]);
+            gain = mode ? av_clipd(gain, 1., irange) : av_clipd(gain, range, 1.);
+
+            {
+                double k = 1. / (tqfactor * gain);
+
+                fa[0] = 1. / (1. + fg * (fg + k));
+                fa[1] = fg * fa[0];
+                fa[2] = fg * fa[1];
+
+                fm[0] = 1.;
+                fm[1] = k * (gain * gain - 1.);
+                fm[2] = 0.;
+            }
+
+            v = get_svf(src[n], fm, fa, &state[3]);
+            dst[n] = ctx->is_disabled ? src[n] : v;
+        }
+    }
+
+    return 0;
+}
+
+static int filter_frame(AVFilterLink *inlink, AVFrame *in)
+{
+    AVFilterContext *ctx = inlink->dst;
+    AudioDynamicEqualizerContext *s = ctx->priv;
+    AVFilterLink *outlink = ctx->outputs[0];
+    ThreadData td;
+    AVFrame *out;
+
+    if (av_frame_is_writable(in)) {
+        out = in;
+    } else {
+        out = ff_get_audio_buffer(outlink, in->nb_samples);
+        if (!out) {
+            av_frame_free(&in);
+            return AVERROR(ENOMEM);
+        }
+        av_frame_copy_props(out, in);
+    }
+
+    s->attack_coeff  = FFMIN(1., 1. / (s->attack * in->sample_rate / 4000.));
+    s->release_coeff = FFMIN(1., 1. / (s->release * in->sample_rate / 4000.));
+    s->knee_sqrt = sqrt(s->knee);
+    s->lin_knee_stop = s->lin_threshold * s->knee_sqrt;
+    s->lin_knee_start = s->lin_threshold / s->knee_sqrt;
+    s->knee_start = log(s->lin_knee_start);
+    s->knee_stop = log(s->lin_knee_stop);
+    s->threshold = log(s->lin_threshold);
+    s->compressed_knee_start = (s->knee_start - s->threshold) / s->ratio + s->threshold;
+
+    td.in = in;
+    td.out = out;
+    ff_filter_execute(ctx, filter_channels, &td, NULL,
+                     FFMIN(outlink->channels, ff_filter_get_nb_threads(ctx)));
+
+    if (out != in)
+        av_frame_free(&in);
+    return ff_filter_frame(outlink, out);
+}
+
+static av_cold void uninit(AVFilterContext *ctx)
+{
+    AudioDynamicEqualizerContext *s = ctx->priv;
+
+    av_frame_free(&s->state);
+}
+
+#define OFFSET(x) offsetof(AudioDynamicEqualizerContext, x)
+#define FLAGS AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_RUNTIME_PARAM
+
+static const AVOption adynamicequalizer_options[] = {
+    { "threshold",  "set detection threshold", OFFSET(lin_threshold),  AV_OPT_TYPE_DOUBLE, {.dbl=0},        0, 50,      FLAGS },
+    { "dfrequency", "set detection frequency", OFFSET(dfrequency),     AV_OPT_TYPE_DOUBLE, {.dbl=1000},     2, 1000000, FLAGS },
+    { "dqfactor",   "set detection Q factor",  OFFSET(dqfactor),       AV_OPT_TYPE_DOUBLE, {.dbl=1},    0.001, 1000,    FLAGS },
+    { "sfrequency", "set smooth frequency",    OFFSET(sfrequency),     AV_OPT_TYPE_DOUBLE, {.dbl=100},      2, 1000000, FLAGS },
+    { "sqfactor",   "set smooth Q factor",     OFFSET(sqfactor),       AV_OPT_TYPE_DOUBLE, {.dbl=1},    0.001, 1000,    FLAGS },
+    { "tfrequency", "set target frequency",    OFFSET(tfrequency),     AV_OPT_TYPE_DOUBLE, {.dbl=1000},     2, 1000000, FLAGS },
+    { "tqfactor",   "set target Q factor",     OFFSET(tqfactor),       AV_OPT_TYPE_DOUBLE, {.dbl=1},    0.001, 1000,    FLAGS },
+    { "attack",     "set attack",              OFFSET(attack),         AV_OPT_TYPE_DOUBLE, {.dbl=10},       1, 2000,    FLAGS },
+    { "release",    "set release",             OFFSET(release),        AV_OPT_TYPE_DOUBLE, {.dbl=80},       1, 2000,    FLAGS },
+    { "knee",       "set knee factor",         OFFSET(knee),           AV_OPT_TYPE_DOUBLE, {.dbl=2},        1, 8,       FLAGS },
+    { "ratio",      "set ratio factor",        OFFSET(ratio),          AV_OPT_TYPE_DOUBLE, {.dbl=1},        1, 20,      FLAGS },
+    { "range",      "set max gain",            OFFSET(range),          AV_OPT_TYPE_DOUBLE, {.dbl=0.06125},0.00000001, 1,FLAGS },
+    { "mode",       "set mode",                OFFSET(mode),           AV_OPT_TYPE_INT,    {.i64=0},        0, 1,       FLAGS, "mode" },
+    {   "cut",      0,                         0,                      AV_OPT_TYPE_CONST,  {.i64=0},        0, 0,       FLAGS, "mode" },
+    {   "boost",    0,                         0,                      AV_OPT_TYPE_CONST,  {.i64=1},        0, 0,       FLAGS, "mode" },
+    { NULL }
+};
+
+AVFILTER_DEFINE_CLASS(adynamicequalizer);
+
+static const AVFilterPad inputs[] = {
+    {
+        .name         = "default",
+        .type         = AVMEDIA_TYPE_AUDIO,
+        .filter_frame = filter_frame,
+        .config_props = config_input,
+    },
+};
+
+static const AVFilterPad outputs[] = {
+    {
+        .name = "default",
+        .type = AVMEDIA_TYPE_AUDIO,
+    },
+};
+
+const AVFilter ff_af_adynamicequalizer = {
+    .name            = "adynamicequalizer",
+    .description     = NULL_IF_CONFIG_SMALL("Apply Dynamic Equalization of input audio."),
+    .priv_size       = sizeof(AudioDynamicEqualizerContext),
+    .priv_class      = &adynamicequalizer_class,
+    .uninit          = uninit,
+    FILTER_INPUTS(inputs),
+    FILTER_OUTPUTS(outputs),
+    FILTER_SINGLE_SAMPLEFMT(AV_SAMPLE_FMT_DBLP),
+    .flags           = AVFILTER_FLAG_SUPPORT_TIMELINE_INTERNAL |
+                       AVFILTER_FLAG_SLICE_THREADS,
+    .process_command = ff_filter_process_command,
+};
diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
index c5c0e9b28b..7018337c85 100644
--- a/libavfilter/allfilters.c
+++ b/libavfilter/allfilters.c
@@ -37,6 +37,7 @@ extern const AVFilter ff_af_adecorrelate;
 extern const AVFilter ff_af_adelay;
 extern const AVFilter ff_af_adenorm;
 extern const AVFilter ff_af_aderivative;
+extern const AVFilter ff_af_adynamicequalizer;
 extern const AVFilter ff_af_adynamicsmooth;
 extern const AVFilter ff_af_aecho;
 extern const AVFilter ff_af_aemphasis;
-- 
2.33.0



More information about the ffmpeg-devel mailing list