2 * Copyright (c) 2012 Clément Bœsch <u pkh me>
4 * This file is part of FFmpeg.
6 * FFmpeg is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
11 * FFmpeg is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with FFmpeg; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23 * Audio silence detector
26 #include <float.h> /* DBL_MAX */
28 #include "libavutil/opt.h"
29 #include "libavutil/timestamp.h"
37 double noise; ///< noise amplitude ratio
38 double duration; ///< minimum duration of silence until notification
39 int64_t nb_null_samples; ///< current number of continuous zero samples
40 int64_t start; ///< if silence is detected, this value contains the time of the first zero sample
41 int last_sample_rate; ///< last sample rate to check for sample rate changes
42 } SilenceDetectContext;
44 #define OFFSET(x) offsetof(SilenceDetectContext, x)
45 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_AUDIO_PARAM
46 static const AVOption silencedetect_options[] = {
47 { "n", "set noise tolerance", OFFSET(noise), AV_OPT_TYPE_DOUBLE, {.dbl=0.001}, 0, DBL_MAX, FLAGS },
48 { "noise", "set noise tolerance", OFFSET(noise), AV_OPT_TYPE_DOUBLE, {.dbl=0.001}, 0, DBL_MAX, FLAGS },
49 { "d", "set minimum duration in seconds", OFFSET(duration), AV_OPT_TYPE_DOUBLE, {.dbl=2.}, 0, 24*60*60, FLAGS },
50 { "duration", "set minimum duration in seconds", OFFSET(duration), AV_OPT_TYPE_DOUBLE, {.dbl=2.}, 0, 24*60*60, FLAGS },
54 AVFILTER_DEFINE_CLASS(silencedetect);
56 static char *get_metadata_val(AVFrame *insamples, const char *key)
58 AVDictionaryEntry *e = av_dict_get(insamples->metadata, key, NULL, 0);
59 return e && e->value ? e->value : NULL;
62 static int filter_frame(AVFilterLink *inlink, AVFrame *insamples)
65 SilenceDetectContext *silence = inlink->dst->priv;
66 const int nb_channels = inlink->channels;
67 const int srate = inlink->sample_rate;
68 const int nb_samples = insamples->nb_samples * nb_channels;
69 const int64_t nb_samples_notify = srate * silence->duration * nb_channels;
71 // scale number of null samples to the new sample rate
72 if (silence->last_sample_rate && silence->last_sample_rate != srate)
73 silence->nb_null_samples =
74 srate * silence->nb_null_samples / silence->last_sample_rate;
75 silence->last_sample_rate = srate;
77 // TODO: support more sample formats
78 // TODO: document metadata
79 if (insamples->format == AV_SAMPLE_FMT_DBL) {
80 double *p = (double *)insamples->data[0];
82 for (i = 0; i < nb_samples; i++, p++) {
83 if (*p < silence->noise && *p > -silence->noise) {
84 if (!silence->start) {
85 silence->nb_null_samples++;
86 if (silence->nb_null_samples >= nb_samples_notify) {
87 silence->start = insamples->pts - (int64_t)(silence->duration / av_q2d(inlink->time_base) + .5);
88 av_dict_set(&insamples->metadata, "lavfi.silence_start",
89 av_ts2timestr(silence->start, &inlink->time_base), 0);
90 av_log(silence, AV_LOG_INFO, "silence_start: %s\n",
91 get_metadata_val(insamples, "lavfi.silence_start"));
96 av_dict_set(&insamples->metadata, "lavfi.silence_end",
97 av_ts2timestr(insamples->pts, &inlink->time_base), 0);
98 av_dict_set(&insamples->metadata, "lavfi.silence_duration",
99 av_ts2timestr(insamples->pts - silence->start, &inlink->time_base), 0);
100 av_log(silence, AV_LOG_INFO,
101 "silence_end: %s | silence_duration: %s\n",
102 get_metadata_val(insamples, "lavfi.silence_end"),
103 get_metadata_val(insamples, "lavfi.silence_duration"));
105 silence->nb_null_samples = silence->start = 0;
110 return ff_filter_frame(inlink->dst->outputs[0], insamples);
113 static int query_formats(AVFilterContext *ctx)
115 AVFilterFormats *formats = NULL;
116 AVFilterChannelLayouts *layouts = NULL;
117 static const enum AVSampleFormat sample_fmts[] = {
122 layouts = ff_all_channel_layouts();
124 return AVERROR(ENOMEM);
125 ff_set_common_channel_layouts(ctx, layouts);
127 formats = ff_make_format_list(sample_fmts);
129 return AVERROR(ENOMEM);
130 ff_set_common_formats(ctx, formats);
132 formats = ff_all_samplerates();
134 return AVERROR(ENOMEM);
135 ff_set_common_samplerates(ctx, formats);
140 static const AVFilterPad silencedetect_inputs[] = {
143 .type = AVMEDIA_TYPE_AUDIO,
144 .filter_frame = filter_frame,
149 static const AVFilterPad silencedetect_outputs[] = {
152 .type = AVMEDIA_TYPE_AUDIO,
157 AVFilter avfilter_af_silencedetect = {
158 .name = "silencedetect",
159 .description = NULL_IF_CONFIG_SMALL("Detect silence."),
160 .priv_size = sizeof(SilenceDetectContext),
161 .query_formats = query_formats,
162 .inputs = silencedetect_inputs,
163 .outputs = silencedetect_outputs,
164 .priv_class = &silencedetect_class,