2 * Copyright (c) 2012 Clément Bœsch <ubitux@gmail.com>
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 "libavutil/opt.h"
27 #include "libavutil/timestamp.h"
32 char *noise_str; ///< noise option string
33 double noise; ///< noise amplitude ratio
34 int duration; ///< minimum duration of silence until notification
35 int64_t nb_null_samples; ///< current number of continuous zero samples
36 int64_t start; ///< if silence is detected, this value contains the time of the first zero sample
37 int last_sample_rate; ///< last sample rate to check for sample rate changes
38 } SilenceDetectContext;
40 #define OFFSET(x) offsetof(SilenceDetectContext, x)
41 static const AVOption silencedetect_options[] = {
42 { "n", "set noise tolerance", OFFSET(noise_str), AV_OPT_TYPE_STRING, {.str="-60dB"}, CHAR_MIN, CHAR_MAX },
43 { "noise", "set noise tolerance", OFFSET(noise_str), AV_OPT_TYPE_STRING, {.str="-60dB"}, CHAR_MIN, CHAR_MAX },
44 { "d", "set minimum duration in seconds", OFFSET(duration), AV_OPT_TYPE_INT, {.dbl=2}, 0, INT_MAX},
45 { "duration", "set minimum duration in seconds", OFFSET(duration), AV_OPT_TYPE_INT, {.dbl=2}, 0, INT_MAX},
49 static const char *silencedetect_get_name(void *ctx)
51 return "silencedetect";
54 static const AVClass silencedetect_class = {
55 .class_name = "SilenceDetectContext",
56 .item_name = silencedetect_get_name,
57 .option = silencedetect_options,
60 static av_cold int init(AVFilterContext *ctx, const char *args, void *opaque)
64 SilenceDetectContext *silence = ctx->priv;
66 silence->class = &silencedetect_class;
67 av_opt_set_defaults(silence);
69 if ((ret = av_set_options_string(silence, args, "=", ":")) < 0) {
70 av_log(ctx, AV_LOG_ERROR, "Error parsing options string: '%s'\n", args);
74 silence->noise = strtod(silence->noise_str, &tail);
75 if (!strcmp(tail, "dB")) {
76 silence->noise = pow(10, silence->noise/20);
78 av_log(ctx, AV_LOG_ERROR, "Invalid value '%s' for noise parameter.\n",
80 return AVERROR(EINVAL);
86 static void filter_samples(AVFilterLink *inlink, AVFilterBufferRef *insamples)
89 SilenceDetectContext *silence = inlink->dst->priv;
90 const int nb_channels = av_get_channel_layout_nb_channels(inlink->channel_layout);
91 const int srate = inlink->sample_rate;
92 const int nb_samples = insamples->audio->nb_samples * nb_channels;
93 const int64_t nb_samples_notify = srate * silence->duration * nb_channels;
95 // scale number of null samples to the new sample rate
96 if (silence->last_sample_rate && silence->last_sample_rate != srate)
97 silence->nb_null_samples =
98 srate * silence->nb_null_samples / silence->last_sample_rate;
99 silence->last_sample_rate = srate;
101 // TODO: support more sample formats
102 if (insamples->format == AV_SAMPLE_FMT_DBL) {
103 double *p = (double *)insamples->data[0];
105 for (i = 0; i < nb_samples; i++, p++) {
106 if (*p < silence->noise && *p > -silence->noise) {
107 if (!silence->start) {
108 silence->nb_null_samples++;
109 if (silence->nb_null_samples >= nb_samples_notify) {
110 silence->start = insamples->pts - silence->duration / av_q2d(inlink->time_base);
111 av_log(silence, AV_LOG_INFO,
112 "silence_start: %s\n", av_ts2timestr(silence->start, &inlink->time_base));
117 av_log(silence, AV_LOG_INFO,
118 "silence_end: %s | silence_duration: %s\n",
119 av_ts2timestr(insamples->pts, &inlink->time_base),
120 av_ts2timestr(insamples->pts - silence->start, &inlink->time_base));
121 silence->nb_null_samples = silence->start = 0;
126 avfilter_filter_samples(inlink->dst->outputs[0], insamples);
129 static int query_formats(AVFilterContext *ctx)
131 AVFilterFormats *formats = NULL;
132 enum AVSampleFormat sample_fmts[] = {
136 int packing_fmts[] = { AVFILTER_PACKED, -1 };
138 formats = avfilter_make_all_channel_layouts();
140 return AVERROR(ENOMEM);
141 avfilter_set_common_channel_layouts(ctx, formats);
143 formats = avfilter_make_format_list(sample_fmts);
145 return AVERROR(ENOMEM);
146 avfilter_set_common_sample_formats(ctx, formats);
148 formats = avfilter_make_format_list(packing_fmts);
150 return AVERROR(ENOMEM);
151 avfilter_set_common_packing_formats(ctx, formats);
156 AVFilter avfilter_af_silencedetect = {
157 .name = "silencedetect",
158 .description = NULL_IF_CONFIG_SMALL("Detect silence."),
159 .priv_size = sizeof(SilenceDetectContext),
161 .query_formats = query_formats,
163 .inputs = (const AVFilterPad[]) {
165 .type = AVMEDIA_TYPE_AUDIO,
166 .get_audio_buffer = avfilter_null_get_audio_buffer,
167 .filter_samples = filter_samples, },
170 .outputs = (const AVFilterPad[]) {
172 .type = AVMEDIA_TYPE_AUDIO, },