aboutsummaryrefslogtreecommitdiffstats
path: root/libavfilter/af_silenceremove.c
blob: d565c75f07e8433880c061b5b888f4b958c54426 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
/*
 * Copyright (c) 2001 Heikki Leinonen
 * Copyright (c) 2001 Chris Bagwell
 * Copyright (c) 2003 Donnie Smith
 * Copyright (c) 2014 Paul B Mahol
 *
 * 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 <float.h> /* DBL_MAX */

#include "libavutil/avassert.h"
#include "libavutil/opt.h"
#include "audio.h"
#include "filters.h"
#include "avfilter.h"
#include "internal.h"

enum SilenceDetect {
    D_AVG,
    D_RMS,
    D_PEAK,
    D_MEDIAN,
    D_PTP,
    D_DEV,
    D_NB
};

enum TimestampMode {
    TS_WRITE,
    TS_COPY,
    TS_NB
};

enum ThresholdMode {
    T_ANY,
    T_ALL,
};

typedef struct SilenceRemoveContext {
    const AVClass *class;

    int start_mode;
    int start_periods;
    int64_t start_duration;
    int64_t start_duration_opt;
    double start_threshold;
    int64_t start_silence;
    int64_t start_silence_opt;

    int stop_mode;
    int stop_periods;
    int64_t stop_duration;
    int64_t stop_duration_opt;
    double stop_threshold;
    int64_t stop_silence;
    int64_t stop_silence_opt;

    int64_t window_duration_opt;

    int timestamp_mode;

    int start_found_periods;
    int stop_found_periods;

    int start_sample_count;
    int start_silence_count;

    int stop_sample_count;
    int stop_silence_count;

    AVFrame *start_window;
    AVFrame *stop_window;

    int *start_front;
    int *start_back;

    int *stop_front;
    int *stop_back;

    int64_t window_duration;
    int cache_size;

    int start_window_pos;
    int start_window_size;

    int stop_window_pos;
    int stop_window_size;

    double *start_cache;
    double *stop_cache;

    AVFrame *start_queuef;
    int start_queue_pos;
    int start_queue_size;

    AVFrame *stop_queuef;
    int stop_queue_pos;
    int stop_queue_size;

    int restart;
    int found_nonsilence;
    int64_t next_pts;

    int detection;

    float (*compute_flt)(float *c, float s, float ws, int size, int *front, int *back);
    double (*compute_dbl)(double *c, double s, double ws, int size, int *front, int *back);
} SilenceRemoveContext;

#define OFFSET(x) offsetof(SilenceRemoveContext, x)
#define AF AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_AUDIO_PARAM
#define AFR AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_AUDIO_PARAM|AV_OPT_FLAG_RUNTIME_PARAM

static const AVOption silenceremove_options[] = {
    { "start_periods",   "set periods of silence parts to skip from start",    OFFSET(start_periods),       AV_OPT_TYPE_INT,      {.i64=0},     0,      9000, AF },
    { "start_duration",  "set start duration of non-silence part",             OFFSET(start_duration_opt),  AV_OPT_TYPE_DURATION, {.i64=0},     0, INT32_MAX, AF },
    { "start_threshold", "set threshold for start silence detection",          OFFSET(start_threshold),     AV_OPT_TYPE_DOUBLE,   {.dbl=0},     0,   DBL_MAX, AFR },
    { "start_silence",   "set start duration of silence part to keep",         OFFSET(start_silence_opt),   AV_OPT_TYPE_DURATION, {.i64=0},     0, INT32_MAX, AF },
    { "start_mode",      "set which channel will trigger trimming from start", OFFSET(start_mode),          AV_OPT_TYPE_INT,      {.i64=T_ANY}, T_ANY, T_ALL, AFR, "mode" },
    {   "any",           0,                                                    0,                           AV_OPT_TYPE_CONST,    {.i64=T_ANY}, 0,         0, AFR, "mode" },
    {   "all",           0,                                                    0,                           AV_OPT_TYPE_CONST,    {.i64=T_ALL}, 0,         0, AFR, "mode" },
    { "stop_periods",    "set periods of silence parts to skip from end",      OFFSET(stop_periods),        AV_OPT_TYPE_INT,      {.i64=0}, -9000,      9000, AF },
    { "stop_duration",   "set stop duration of silence part",                  OFFSET(stop_duration_opt),   AV_OPT_TYPE_DURATION, {.i64=0},     0, INT32_MAX, AF },
    { "stop_threshold",  "set threshold for stop silence detection",           OFFSET(stop_threshold),      AV_OPT_TYPE_DOUBLE,   {.dbl=0},     0,   DBL_MAX, AFR },
    { "stop_silence",    "set stop duration of silence part to keep",          OFFSET(stop_silence_opt),    AV_OPT_TYPE_DURATION, {.i64=0},     0, INT32_MAX, AF },
    { "stop_mode",       "set which channel will trigger trimming from end",   OFFSET(stop_mode),           AV_OPT_TYPE_INT,      {.i64=T_ALL}, T_ANY, T_ALL, AFR, "mode" },
    { "detection",       "set how silence is detected",                        OFFSET(detection),           AV_OPT_TYPE_INT,      {.i64=D_RMS}, 0,    D_NB-1, AF, "detection" },
    {   "avg",           "use mean absolute values of samples",                0,                           AV_OPT_TYPE_CONST,    {.i64=D_AVG}, 0,         0, AF, "detection" },
    {   "rms",           "use root mean squared values of samples",            0,                           AV_OPT_TYPE_CONST,    {.i64=D_RMS}, 0,         0, AF, "detection" },
    {   "peak",          "use max absolute values of samples",                 0,                           AV_OPT_TYPE_CONST,    {.i64=D_PEAK},0,         0, AF, "detection" },
    {   "median",        "use median of absolute values of samples",           0,                           AV_OPT_TYPE_CONST,    {.i64=D_MEDIAN},0,       0, AF, "detection" },
    {   "ptp",           "use absolute of max peak to min peak difference",    0,                           AV_OPT_TYPE_CONST,    {.i64=D_PTP}, 0,         0, AF, "detection" },
    {   "dev",           "use standard deviation from values of samples",      0,                           AV_OPT_TYPE_CONST,    {.i64=D_DEV}, 0,         0, AF, "detection" },
    { "window",          "set duration of window for silence detection",       OFFSET(window_duration_opt), AV_OPT_TYPE_DURATION, {.i64=20000}, 0, 100000000, AF },
    { "timestamp",       "set how every output frame timestamp is processed",  OFFSET(timestamp_mode),      AV_OPT_TYPE_INT,      {.i64=TS_WRITE}, 0, TS_NB-1, AF, "timestamp" },
    {   "write",         "full timestamps rewrite, keep only the start time",  0,                           AV_OPT_TYPE_CONST,    {.i64=TS_WRITE}, 0,       0, AF, "timestamp" },
    {   "copy",          "non-dropped frames are left with same timestamp",    0,                           AV_OPT_TYPE_CONST,    {.i64=TS_COPY},  0,       0, AF, "timestamp" },
    { NULL }
};

AVFILTER_DEFINE_CLASS(silenceremove);

#define DEPTH 32
#include "silenceremove_template.c"

#undef DEPTH
#define DEPTH 64
#include "silenceremove_template.c"

static av_cold int init(AVFilterContext *ctx)
{
    SilenceRemoveContext *s = ctx->priv;

    if (s->stop_periods < 0) {
        s->stop_periods = -s->stop_periods;
        s->restart = 1;
    }

    return 0;
}

static void clear_windows(SilenceRemoveContext *s)
{
    av_samples_set_silence(s->start_window->extended_data, 0,
                           s->start_window->nb_samples,
                           s->start_window->ch_layout.nb_channels,
                           s->start_window->format);
    av_samples_set_silence(s->stop_window->extended_data, 0,
                           s->stop_window->nb_samples,
                           s->stop_window->ch_layout.nb_channels,
                           s->stop_window->format);

    s->start_window_pos = 0;
    s->start_window_size = 0;
    s->stop_window_pos = 0;
    s->stop_window_size = 0;
    s->start_queue_pos = 0;
    s->start_queue_size = 0;
    s->stop_queue_pos = 0;
    s->stop_queue_size = 0;
}

static int config_input(AVFilterLink *inlink)
{
    AVFilterContext *ctx = inlink->dst;
    SilenceRemoveContext *s = ctx->priv;

    s->next_pts = AV_NOPTS_VALUE;
    s->window_duration = av_rescale(s->window_duration_opt, inlink->sample_rate,
                                   AV_TIME_BASE);
    s->window_duration = FFMAX(1, s->window_duration);

    s->start_duration = av_rescale(s->start_duration_opt, inlink->sample_rate,
                                   AV_TIME_BASE);
    s->start_silence  = av_rescale(s->start_silence_opt, inlink->sample_rate,
                                   AV_TIME_BASE);
    s->stop_duration = av_rescale(s->stop_duration_opt, inlink->sample_rate,
                                   AV_TIME_BASE);
    s->stop_silence  = av_rescale(s->stop_silence_opt, inlink->sample_rate,
                                   AV_TIME_BASE);

    s->start_found_periods = 0;
    s->stop_found_periods  = 0;

    return 0;
}

static int config_output(AVFilterLink *outlink)
{
    AVFilterContext *ctx = outlink->src;
    SilenceRemoveContext *s = ctx->priv;

    switch (s->detection) {
    case D_AVG:
    case D_RMS:
        s->cache_size = 1;
        break;
    case D_DEV:
        s->cache_size = 2;
        break;
    case D_MEDIAN:
    case D_PEAK:
    case D_PTP:
        s->cache_size = s->window_duration;
        break;
    }

    s->start_window = ff_get_audio_buffer(outlink, s->window_duration);
    s->stop_window = ff_get_audio_buffer(outlink, s->window_duration);
    s->start_cache = av_calloc(outlink->ch_layout.nb_channels, s->cache_size * sizeof(*s->start_cache));
    s->stop_cache = av_calloc(outlink->ch_layout.nb_channels, s->cache_size * sizeof(*s->stop_cache));
    if (!s->start_window || !s->stop_window || !s->start_cache || !s->stop_cache)
        return AVERROR(ENOMEM);

    s->start_queuef = ff_get_audio_buffer(outlink, s->start_silence + 1);
    s->stop_queuef = ff_get_audio_buffer(outlink, s->stop_silence + 1);
    if (!s->start_queuef || !s->stop_queuef)
        return AVERROR(ENOMEM);

    s->start_front = av_calloc(outlink->ch_layout.nb_channels, sizeof(*s->start_front));
    s->start_back = av_calloc(outlink->ch_layout.nb_channels, sizeof(*s->start_back));
    s->stop_front = av_calloc(outlink->ch_layout.nb_channels, sizeof(*s->stop_front));
    s->stop_back = av_calloc(outlink->ch_layout.nb_channels, sizeof(*s->stop_back));
    if (!s->start_front || !s->start_back || !s->stop_front || !s->stop_back)
        return AVERROR(ENOMEM);

    clear_windows(s);

    switch (s->detection) {
    case D_AVG:
        s->compute_flt = compute_avg_flt;
        s->compute_dbl = compute_avg_dbl;
        break;
    case D_DEV:
        s->compute_flt = compute_dev_flt;
        s->compute_dbl = compute_dev_dbl;
        break;
    case D_PTP:
        s->compute_flt = compute_ptp_flt;
        s->compute_dbl = compute_ptp_dbl;
        break;
    case D_MEDIAN:
        s->compute_flt = compute_median_flt;
        s->compute_dbl = compute_median_dbl;
        break;
    case D_PEAK:
        s->compute_flt = compute_peak_flt;
        s->compute_dbl = compute_peak_dbl;
        break;
    case D_RMS:
        s->compute_flt = compute_rms_flt;
        s->compute_dbl = compute_rms_dbl;
        break;
    }

    return 0;
}

static int filter_frame(AVFilterLink *outlink, AVFrame *in)
{
    const int nb_channels = outlink->ch_layout.nb_channels;
    AVFilterContext *ctx = outlink->src;
    SilenceRemoveContext *s = ctx->priv;
    int max_out_nb_samples;
    int out_nb_samples = 0;
    int in_nb_samples;
    const double *srcd;
    const float *srcf;
    AVFrame *out;
    double *dstd;
    float *dstf;

    if (s->next_pts == AV_NOPTS_VALUE)
        s->next_pts = in->pts;

    in_nb_samples = in->nb_samples;
    max_out_nb_samples = in->nb_samples +
                         s->start_silence +
                         s->stop_silence;
    if (max_out_nb_samples <= 0) {
        av_frame_free(&in);
        ff_filter_set_ready(ctx, 100);
        return 0;
    }

    out = ff_get_audio_buffer(outlink, max_out_nb_samples);
    if (!out) {
        av_frame_free(&in);
        return AVERROR(ENOMEM);
    }

    if (s->timestamp_mode == TS_WRITE)
        out->pts = s->next_pts;
    else
        out->pts = in->pts;

    switch (outlink->format) {
    case AV_SAMPLE_FMT_FLT:
        srcf = (const float *)in->data[0];
        dstf = (float *)out->data[0];
        if (s->start_periods > 0 && s->stop_periods > 0) {
            const float *src = srcf;
            if (s->start_found_periods >= 0) {
                for (int n = 0; n < in_nb_samples; n++) {
                    filter_start_flt(ctx, src + n * nb_channels,
                                     dstf, &out_nb_samples,
                                     nb_channels);
                }
                in_nb_samples = out_nb_samples;
                out_nb_samples = 0;
                src = dstf;
            }
            for (int n = 0; n < in_nb_samples; n++) {
                filter_stop_flt(ctx, src + n * nb_channels,
                                dstf, &out_nb_samples,
                                nb_channels);
            }
        } else if (s->start_periods > 0) {
            for (int n = 0; n < in_nb_samples; n++) {
                filter_start_flt(ctx, srcf + n * nb_channels,
                                 dstf, &out_nb_samples,
                                 nb_channels);
            }
        } else if (s->stop_periods > 0) {
            for (int n = 0; n < in_nb_samples; n++) {
                filter_stop_flt(ctx, srcf + n * nb_channels,
                                dstf, &out_nb_samples,
                                nb_channels);
            }
        }
        break;
    case AV_SAMPLE_FMT_DBL:
        srcd = (const double *)in->data[0];
        dstd = (double *)out->data[0];
        if (s->start_periods > 0 && s->stop_periods > 0) {
            const double *src = srcd;
            if (s->start_found_periods >= 0) {
                for (int n = 0; n < in_nb_samples; n++) {
                    filter_start_dbl(ctx, src + n * nb_channels,
                                     dstd, &out_nb_samples,
                                     nb_channels);
                }
                in_nb_samples = out_nb_samples;
                out_nb_samples = 0;
                src = dstd;
            }
            for (int n = 0; n < in_nb_samples; n++) {
                filter_stop_dbl(ctx, src + n * nb_channels,
                                dstd, &out_nb_samples,
                                nb_channels);
            }
        } else if (s->start_periods > 0) {
            for (int n = 0; n < in_nb_samples; n++) {
                filter_start_dbl(ctx, srcd + n * nb_channels,
                                 dstd, &out_nb_samples,
                                 nb_channels);
            }
        } else if (s->stop_periods > 0) {
            for (int n = 0; n < in_nb_samples; n++) {
                filter_stop_dbl(ctx, srcd + n * nb_channels,
                                dstd, &out_nb_samples,
                                nb_channels);
            }
        }
        break;
    }

    av_frame_free(&in);
    if (out_nb_samples > 0) {
        s->next_pts += out_nb_samples;
        out->nb_samples = out_nb_samples;
        return ff_filter_frame(outlink, out);
    }

    av_frame_free(&out);
    ff_filter_set_ready(ctx, 100);

    return 0;
}

static int activate(AVFilterContext *ctx)
{
    AVFilterLink *outlink = ctx->outputs[0];
    AVFilterLink *inlink = ctx->inputs[0];
    SilenceRemoveContext *s = ctx->priv;
    AVFrame *in;
    int ret;

    FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);

    ret = ff_inlink_consume_frame(inlink, &in);
    if (ret < 0)
        return ret;
    if (ret > 0) {
        if (s->start_periods == 1 && s->stop_periods == 0 &&
            s->start_found_periods < 0) {
            if (s->timestamp_mode == TS_WRITE)
                in->pts = s->next_pts;
            s->next_pts += in->nb_samples;
            return ff_filter_frame(outlink, in);
        }
        if (s->start_periods == 0 && s->stop_periods == 0)
            return ff_filter_frame(outlink, in);
        return filter_frame(outlink, in);
    }

    FF_FILTER_FORWARD_STATUS(inlink, outlink);
    FF_FILTER_FORWARD_WANTED(outlink, inlink);

    return FFERROR_NOT_READY;
}

static av_cold void uninit(AVFilterContext *ctx)
{
    SilenceRemoveContext *s = ctx->priv;

    av_frame_free(&s->start_window);
    av_frame_free(&s->stop_window);
    av_frame_free(&s->start_queuef);
    av_frame_free(&s->stop_queuef);

    av_freep(&s->start_cache);
    av_freep(&s->stop_cache);
    av_freep(&s->start_front);
    av_freep(&s->start_back);
    av_freep(&s->stop_front);
    av_freep(&s->stop_back);
}

static const AVFilterPad silenceremove_inputs[] = {
    {
        .name         = "default",
        .type         = AVMEDIA_TYPE_AUDIO,
        .config_props = config_input,
    },
};

static const AVFilterPad silenceremove_outputs[] = {
    {
        .name         = "default",
        .type         = AVMEDIA_TYPE_AUDIO,
        .config_props = config_output,
    },
};

const AVFilter ff_af_silenceremove = {
    .name          = "silenceremove",
    .description   = NULL_IF_CONFIG_SMALL("Remove silence."),
    .priv_size     = sizeof(SilenceRemoveContext),
    .priv_class    = &silenceremove_class,
    .init          = init,
    .activate      = activate,
    .uninit        = uninit,
    FILTER_INPUTS(silenceremove_inputs),
    FILTER_OUTPUTS(silenceremove_outputs),
    FILTER_SAMPLEFMTS(AV_SAMPLE_FMT_FLT,
                      AV_SAMPLE_FMT_DBL),
    .process_command = ff_filter_process_command,
    .flags           = AVFILTER_FLAG_SUPPORT_TIMELINE_INTERNAL,
};