FFmpeg coverage


Directory: ../../../ffmpeg/
File: src/libavfilter/af_volume.c
Date: 2025-04-25 22:50:00
Exec Total Coverage
Lines: 119 202 58.9%
Functions: 9 14 64.3%
Branches: 47 106 44.3%

Line Branch Exec Source
1 /*
2 * Copyright (c) 2011 Stefano Sabatini
3 * Copyright (c) 2012 Justin Ruggles <justin.ruggles@gmail.com>
4 *
5 * This file is part of FFmpeg.
6 *
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
11 *
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
16 *
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 */
21
22 /**
23 * @file
24 * audio volume filter
25 */
26
27 #include "libavutil/channel_layout.h"
28 #include "libavutil/common.h"
29 #include "libavutil/eval.h"
30 #include "libavutil/ffmath.h"
31 #include "libavutil/float_dsp.h"
32 #include "libavutil/mem.h"
33 #include "libavutil/opt.h"
34 #include "libavutil/replaygain.h"
35
36 #include "audio.h"
37 #include "avfilter.h"
38 #include "filters.h"
39 #include "formats.h"
40 #include "af_volume.h"
41
42 static const char * const precision_str[] = {
43 "fixed", "float", "double"
44 };
45
46 static const char *const var_names[] = {
47 "n", ///< frame number (starting at zero)
48 "nb_channels", ///< number of channels
49 "nb_consumed_samples", ///< number of samples consumed by the filter
50 "nb_samples", ///< number of samples in the current frame
51 "pts", ///< frame presentation timestamp
52 "sample_rate", ///< sample rate
53 "startpts", ///< PTS at start of stream
54 "startt", ///< time at start of stream
55 "t", ///< time in the file of the frame
56 "tb", ///< timebase
57 "volume", ///< last set value
58 NULL
59 };
60
61 #define OFFSET(x) offsetof(VolumeContext, x)
62 #define A AV_OPT_FLAG_AUDIO_PARAM
63 #define F AV_OPT_FLAG_FILTERING_PARAM
64 #define T AV_OPT_FLAG_RUNTIME_PARAM
65
66 static const AVOption volume_options[] = {
67 { "volume", "set volume adjustment expression",
68 OFFSET(volume_expr), AV_OPT_TYPE_STRING, { .str = "1.0" }, .flags = A|F|T },
69 { "precision", "select mathematical precision",
70 OFFSET(precision), AV_OPT_TYPE_INT, { .i64 = PRECISION_FLOAT }, PRECISION_FIXED, PRECISION_DOUBLE, A|F, .unit = "precision" },
71 { "fixed", "select 8-bit fixed-point", 0, AV_OPT_TYPE_CONST, { .i64 = PRECISION_FIXED }, INT_MIN, INT_MAX, A|F, .unit = "precision" },
72 { "float", "select 32-bit floating-point", 0, AV_OPT_TYPE_CONST, { .i64 = PRECISION_FLOAT }, INT_MIN, INT_MAX, A|F, .unit = "precision" },
73 { "double", "select 64-bit floating-point", 0, AV_OPT_TYPE_CONST, { .i64 = PRECISION_DOUBLE }, INT_MIN, INT_MAX, A|F, .unit = "precision" },
74 { "eval", "specify when to evaluate expressions", OFFSET(eval_mode), AV_OPT_TYPE_INT, {.i64 = EVAL_MODE_ONCE}, 0, EVAL_MODE_NB-1, .flags = A|F, .unit = "eval" },
75 { "once", "eval volume expression once", 0, AV_OPT_TYPE_CONST, {.i64=EVAL_MODE_ONCE}, .flags = A|F, .unit = "eval" },
76 { "frame", "eval volume expression per-frame", 0, AV_OPT_TYPE_CONST, {.i64=EVAL_MODE_FRAME}, .flags = A|F, .unit = "eval" },
77 { "replaygain", "Apply replaygain side data when present",
78 OFFSET(replaygain), AV_OPT_TYPE_INT, { .i64 = REPLAYGAIN_DROP }, REPLAYGAIN_DROP, REPLAYGAIN_ALBUM, A|F, .unit = "replaygain" },
79 { "drop", "replaygain side data is dropped", 0, AV_OPT_TYPE_CONST, { .i64 = REPLAYGAIN_DROP }, 0, 0, A|F, .unit = "replaygain" },
80 { "ignore", "replaygain side data is ignored", 0, AV_OPT_TYPE_CONST, { .i64 = REPLAYGAIN_IGNORE }, 0, 0, A|F, .unit = "replaygain" },
81 { "track", "track gain is preferred", 0, AV_OPT_TYPE_CONST, { .i64 = REPLAYGAIN_TRACK }, 0, 0, A|F, .unit = "replaygain" },
82 { "album", "album gain is preferred", 0, AV_OPT_TYPE_CONST, { .i64 = REPLAYGAIN_ALBUM }, 0, 0, A|F, .unit = "replaygain" },
83 { "replaygain_preamp", "Apply replaygain pre-amplification",
84 OFFSET(replaygain_preamp), AV_OPT_TYPE_DOUBLE, { .dbl = 0.0 }, -15.0, 15.0, A|F },
85 { "replaygain_noclip", "Apply replaygain clipping prevention",
86 OFFSET(replaygain_noclip), AV_OPT_TYPE_BOOL, { .i64 = 1 }, 0, 1, A|F },
87 { NULL }
88 };
89
90 AVFILTER_DEFINE_CLASS(volume);
91
92 6 static int set_expr(AVExpr **pexpr, const char *expr, void *log_ctx)
93 {
94 int ret;
95 6 AVExpr *old = NULL;
96
97
1/2
✗ Branch 0 not taken.
✓ Branch 1 taken 6 times.
6 if (*pexpr)
98 old = *pexpr;
99 6 ret = av_expr_parse(pexpr, expr, var_names,
100 NULL, NULL, NULL, NULL, 0, log_ctx);
101
1/2
✗ Branch 0 not taken.
✓ Branch 1 taken 6 times.
6 if (ret < 0) {
102 av_log(log_ctx, AV_LOG_ERROR,
103 "Error when evaluating the volume expression '%s'\n", expr);
104 *pexpr = old;
105 return ret;
106 }
107
108 6 av_expr_free(old);
109 6 return 0;
110 }
111
112 6 static av_cold int init(AVFilterContext *ctx)
113 {
114 6 VolumeContext *vol = ctx->priv;
115
116 6 vol->fdsp = avpriv_float_dsp_alloc(0);
117
1/2
✗ Branch 0 not taken.
✓ Branch 1 taken 6 times.
6 if (!vol->fdsp)
118 return AVERROR(ENOMEM);
119
120 6 return set_expr(&vol->volume_pexpr, vol->volume_expr, ctx);
121 }
122
123 6 static av_cold void uninit(AVFilterContext *ctx)
124 {
125 6 VolumeContext *vol = ctx->priv;
126 6 av_expr_free(vol->volume_pexpr);
127 6 av_freep(&vol->fdsp);
128 6 }
129
130 3 static int query_formats(const AVFilterContext *ctx,
131 AVFilterFormatsConfig **cfg_in,
132 AVFilterFormatsConfig **cfg_out)
133 {
134 3 const VolumeContext *vol = ctx->priv;
135 static const enum AVSampleFormat sample_fmts[][7] = {
136 [PRECISION_FIXED] = {
137 AV_SAMPLE_FMT_U8,
138 AV_SAMPLE_FMT_U8P,
139 AV_SAMPLE_FMT_S16,
140 AV_SAMPLE_FMT_S16P,
141 AV_SAMPLE_FMT_S32,
142 AV_SAMPLE_FMT_S32P,
143 AV_SAMPLE_FMT_NONE
144 },
145 [PRECISION_FLOAT] = {
146 AV_SAMPLE_FMT_FLT,
147 AV_SAMPLE_FMT_FLTP,
148 AV_SAMPLE_FMT_NONE
149 },
150 [PRECISION_DOUBLE] = {
151 AV_SAMPLE_FMT_DBL,
152 AV_SAMPLE_FMT_DBLP,
153 AV_SAMPLE_FMT_NONE
154 }
155 };
156 int ret;
157
158 3 ret = ff_set_common_formats_from_list2(ctx, cfg_in, cfg_out, sample_fmts[vol->precision]);
159
1/2
✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
3 if (ret < 0)
160 return ret;
161
162 3 return 0;
163 }
164
165 static inline void scale_samples_u8(uint8_t *dst, const uint8_t *src,
166 int nb_samples, int volume)
167 {
168 int i;
169 for (i = 0; i < nb_samples; i++)
170 dst[i] = av_clip_uint8(((((int64_t)src[i] - 128) * volume + 128) >> 8) + 128);
171 }
172
173 static inline void scale_samples_u8_small(uint8_t *dst, const uint8_t *src,
174 int nb_samples, int volume)
175 {
176 int i;
177 for (i = 0; i < nb_samples; i++)
178 dst[i] = av_clip_uint8((((src[i] - 128) * volume + 128) >> 8) + 128);
179 }
180
181 static inline void scale_samples_s16(uint8_t *dst, const uint8_t *src,
182 int nb_samples, int volume)
183 {
184 int i;
185 int16_t *smp_dst = (int16_t *)dst;
186 const int16_t *smp_src = (const int16_t *)src;
187 for (i = 0; i < nb_samples; i++)
188 smp_dst[i] = av_clip_int16(((int64_t)smp_src[i] * volume + 128) >> 8);
189 }
190
191 65 static inline void scale_samples_s16_small(uint8_t *dst, const uint8_t *src,
192 int nb_samples, int volume)
193 {
194 int i;
195 65 int16_t *smp_dst = (int16_t *)dst;
196 65 const int16_t *smp_src = (const int16_t *)src;
197
2/2
✓ Branch 0 taken 529200 times.
✓ Branch 1 taken 65 times.
529265 for (i = 0; i < nb_samples; i++)
198 529200 smp_dst[i] = av_clip_int16((smp_src[i] * volume + 128) >> 8);
199 65 }
200
201 static inline void scale_samples_s32(uint8_t *dst, const uint8_t *src,
202 int nb_samples, int volume)
203 {
204 int i;
205 int32_t *smp_dst = (int32_t *)dst;
206 const int32_t *smp_src = (const int32_t *)src;
207 for (i = 0; i < nb_samples; i++)
208 smp_dst[i] = av_clipl_int32((((int64_t)smp_src[i] * volume + 128) >> 8));
209 }
210
211 3 static av_cold void volume_init(VolumeContext *vol)
212 {
213 3 vol->samples_align = 1;
214
215
3/6
✗ Branch 1 not taken.
✓ Branch 2 taken 1 times.
✓ Branch 3 taken 1 times.
✓ Branch 4 taken 1 times.
✗ Branch 5 not taken.
✗ Branch 6 not taken.
3 switch (av_get_packed_sample_fmt(vol->sample_fmt)) {
216 case AV_SAMPLE_FMT_U8:
217 if (vol->volume_i < 0x1000000)
218 vol->scale_samples = scale_samples_u8_small;
219 else
220 vol->scale_samples = scale_samples_u8;
221 break;
222 1 case AV_SAMPLE_FMT_S16:
223
1/2
✓ Branch 0 taken 1 times.
✗ Branch 1 not taken.
1 if (vol->volume_i < 0x10000)
224 1 vol->scale_samples = scale_samples_s16_small;
225 else
226 vol->scale_samples = scale_samples_s16;
227 1 break;
228 1 case AV_SAMPLE_FMT_S32:
229 1 vol->scale_samples = scale_samples_s32;
230 1 break;
231 1 case AV_SAMPLE_FMT_FLT:
232 1 vol->samples_align = 4;
233 1 break;
234 case AV_SAMPLE_FMT_DBL:
235 vol->samples_align = 8;
236 break;
237 }
238
239 #if ARCH_X86
240 3 ff_volume_init_x86(vol);
241 #endif
242 3 }
243
244 3 static int set_volume(AVFilterContext *ctx)
245 {
246 3 VolumeContext *vol = ctx->priv;
247
248 3 vol->volume = av_expr_eval(vol->volume_pexpr, vol->var_values, NULL);
249
1/2
✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
3 if (isnan(vol->volume)) {
250 if (vol->eval_mode == EVAL_MODE_ONCE) {
251 av_log(ctx, AV_LOG_ERROR, "Invalid value NaN for volume\n");
252 return AVERROR(EINVAL);
253 } else {
254 av_log(ctx, AV_LOG_WARNING, "Invalid value NaN for volume, setting to 0\n");
255 vol->volume = 0;
256 }
257 }
258 3 vol->var_values[VAR_VOLUME] = vol->volume;
259
260 3 av_log(ctx, AV_LOG_VERBOSE, "n:%f t:%f pts:%f precision:%s ",
261 vol->var_values[VAR_N], vol->var_values[VAR_T], vol->var_values[VAR_PTS],
262 3 precision_str[vol->precision]);
263
264
2/2
✓ Branch 0 taken 2 times.
✓ Branch 1 taken 1 times.
3 if (vol->precision == PRECISION_FIXED) {
265 2 vol->volume_i = (int)(vol->volume * 256 + 0.5);
266 2 vol->volume = vol->volume_i / 256.0;
267 2 av_log(ctx, AV_LOG_VERBOSE, "volume_i:%d/255 ", vol->volume_i);
268 }
269 3 av_log(ctx, AV_LOG_VERBOSE, "volume:%f volume_dB:%f\n",
270 3 vol->volume, 20.0*log10(vol->volume));
271
272 3 volume_init(vol);
273 3 return 0;
274 }
275
276 3 static int config_output(AVFilterLink *outlink)
277 {
278 3 AVFilterContext *ctx = outlink->src;
279 3 VolumeContext *vol = ctx->priv;
280 3 AVFilterLink *inlink = ctx->inputs[0];
281
282 3 vol->sample_fmt = inlink->format;
283 3 vol->channels = inlink->ch_layout.nb_channels;
284
2/2
✓ Branch 1 taken 2 times.
✓ Branch 2 taken 1 times.
3 vol->planes = av_sample_fmt_is_planar(inlink->format) ? vol->channels : 1;
285
286 3 vol->var_values[VAR_N] =
287 3 vol->var_values[VAR_NB_CONSUMED_SAMPLES] =
288 3 vol->var_values[VAR_NB_SAMPLES] =
289 3 vol->var_values[VAR_PTS] =
290 3 vol->var_values[VAR_STARTPTS] =
291 3 vol->var_values[VAR_STARTT] =
292 3 vol->var_values[VAR_T] =
293 3 vol->var_values[VAR_VOLUME] = NAN;
294
295 3 vol->var_values[VAR_NB_CHANNELS] = inlink->ch_layout.nb_channels;
296 3 vol->var_values[VAR_TB] = av_q2d(inlink->time_base);
297 3 vol->var_values[VAR_SAMPLE_RATE] = inlink->sample_rate;
298
299 3 av_log(inlink->src, AV_LOG_VERBOSE, "tb:%f sample_rate:%f nb_channels:%f\n",
300 vol->var_values[VAR_TB],
301 vol->var_values[VAR_SAMPLE_RATE],
302 vol->var_values[VAR_NB_CHANNELS]);
303
304 3 return set_volume(ctx);
305 }
306
307 static int process_command(AVFilterContext *ctx, const char *cmd, const char *args,
308 char *res, int res_len, int flags)
309 {
310 VolumeContext *vol = ctx->priv;
311 int ret = AVERROR(ENOSYS);
312
313 if (!strcmp(cmd, "volume")) {
314 if ((ret = set_expr(&vol->volume_pexpr, args, ctx)) < 0)
315 return ret;
316 if (vol->eval_mode == EVAL_MODE_ONCE)
317 set_volume(ctx);
318 }
319
320 return ret;
321 }
322
323 199 static int filter_frame(AVFilterLink *inlink, AVFrame *buf)
324 {
325 199 FilterLink *inl = ff_filter_link(inlink);
326 199 AVFilterContext *ctx = inlink->dst;
327 199 VolumeContext *vol = inlink->dst->priv;
328 199 AVFilterLink *outlink = inlink->dst->outputs[0];
329 199 int nb_samples = buf->nb_samples;
330 AVFrame *out_buf;
331 199 AVFrameSideData *sd = av_frame_get_side_data(buf, AV_FRAME_DATA_REPLAYGAIN);
332 int ret;
333
334
1/4
✗ Branch 0 not taken.
✓ Branch 1 taken 199 times.
✗ Branch 2 not taken.
✗ Branch 3 not taken.
199 if (sd && vol->replaygain != REPLAYGAIN_IGNORE) {
335 if (vol->replaygain != REPLAYGAIN_DROP) {
336 AVReplayGain *replaygain = (AVReplayGain*)sd->data;
337 int32_t gain = 100000;
338 uint32_t peak = 100000;
339 float g, p;
340
341 if (vol->replaygain == REPLAYGAIN_TRACK &&
342 replaygain->track_gain != INT32_MIN) {
343 gain = replaygain->track_gain;
344
345 if (replaygain->track_peak != 0)
346 peak = replaygain->track_peak;
347 } else if (replaygain->album_gain != INT32_MIN) {
348 gain = replaygain->album_gain;
349
350 if (replaygain->album_peak != 0)
351 peak = replaygain->album_peak;
352 } else {
353 av_log(inlink->dst, AV_LOG_WARNING, "Both ReplayGain gain "
354 "values are unknown.\n");
355 }
356 g = gain / 100000.0f;
357 p = peak / 100000.0f;
358
359 av_log(inlink->dst, AV_LOG_VERBOSE,
360 "Using gain %f dB from replaygain side data.\n", g);
361
362 vol->volume = ff_exp10((g + vol->replaygain_preamp) / 20);
363 if (vol->replaygain_noclip)
364 vol->volume = FFMIN(vol->volume, 1.0 / p);
365 vol->volume_i = (int)(vol->volume * 256 + 0.5);
366
367 volume_init(vol);
368 }
369 av_frame_remove_side_data(buf, AV_FRAME_DATA_REPLAYGAIN);
370 }
371
372
2/2
✓ Branch 0 taken 3 times.
✓ Branch 1 taken 196 times.
199 if (isnan(vol->var_values[VAR_STARTPTS])) {
373
1/2
✓ Branch 0 taken 3 times.
✗ Branch 1 not taken.
3 vol->var_values[VAR_STARTPTS] = TS2D(buf->pts);
374
1/2
✓ Branch 0 taken 3 times.
✗ Branch 1 not taken.
3 vol->var_values[VAR_STARTT ] = TS2T(buf->pts, inlink->time_base);
375 }
376
1/2
✓ Branch 0 taken 199 times.
✗ Branch 1 not taken.
199 vol->var_values[VAR_PTS] = TS2D(buf->pts);
377
1/2
✓ Branch 0 taken 199 times.
✗ Branch 1 not taken.
199 vol->var_values[VAR_T ] = TS2T(buf->pts, inlink->time_base);
378 199 vol->var_values[VAR_N ] = inl->frame_count_out;
379
380
1/2
✗ Branch 0 not taken.
✓ Branch 1 taken 199 times.
199 if (vol->eval_mode == EVAL_MODE_FRAME)
381 set_volume(ctx);
382
383
2/4
✓ Branch 0 taken 199 times.
✗ Branch 1 not taken.
✗ Branch 2 not taken.
✓ Branch 3 taken 199 times.
199 if (vol->volume == 1.0 || vol->volume_i == 256) {
384 out_buf = buf;
385 goto end;
386 }
387
388 /* do volume scaling in-place if input buffer is writable */
389
1/2
✓ Branch 1 taken 199 times.
✗ Branch 2 not taken.
199 if (av_frame_is_writable(buf)
390
4/4
✓ Branch 0 taken 158 times.
✓ Branch 1 taken 41 times.
✓ Branch 2 taken 65 times.
✓ Branch 3 taken 93 times.
199 && (vol->precision != PRECISION_FIXED || vol->volume_i > 0)) {
391 106 out_buf = buf;
392 } else {
393 93 out_buf = ff_get_audio_buffer(outlink, nb_samples);
394
1/2
✗ Branch 0 not taken.
✓ Branch 1 taken 93 times.
93 if (!out_buf) {
395 av_frame_free(&buf);
396 return AVERROR(ENOMEM);
397 }
398 93 ret = av_frame_copy_props(out_buf, buf);
399
1/2
✗ Branch 0 not taken.
✓ Branch 1 taken 93 times.
93 if (ret < 0) {
400 av_frame_free(&out_buf);
401 av_frame_free(&buf);
402 return ret;
403 }
404 }
405
406
4/4
✓ Branch 0 taken 158 times.
✓ Branch 1 taken 41 times.
✓ Branch 2 taken 65 times.
✓ Branch 3 taken 93 times.
199 if (vol->precision != PRECISION_FIXED || vol->volume_i > 0) {
407 int p, plane_samples;
408
409
2/2
✓ Branch 1 taken 41 times.
✓ Branch 2 taken 65 times.
106 if (av_sample_fmt_is_planar(buf->format))
410 41 plane_samples = FFALIGN(nb_samples, vol->samples_align);
411 else
412 65 plane_samples = FFALIGN(nb_samples * vol->channels, vol->samples_align);
413
414
2/2
✓ Branch 0 taken 65 times.
✓ Branch 1 taken 41 times.
106 if (vol->precision == PRECISION_FIXED) {
415
2/2
✓ Branch 0 taken 65 times.
✓ Branch 1 taken 65 times.
130 for (p = 0; p < vol->planes; p++) {
416 65 vol->scale_samples(out_buf->extended_data[p],
417 65 buf->extended_data[p], plane_samples,
418 vol->volume_i);
419 }
420
1/2
✓ Branch 1 taken 41 times.
✗ Branch 2 not taken.
41 } else if (av_get_packed_sample_fmt(vol->sample_fmt) == AV_SAMPLE_FMT_FLT) {
421
2/2
✓ Branch 0 taken 82 times.
✓ Branch 1 taken 41 times.
123 for (p = 0; p < vol->planes; p++) {
422 82 vol->fdsp->vector_fmul_scalar((float *)out_buf->extended_data[p],
423 82 (const float *)buf->extended_data[p],
424 82 vol->volume, plane_samples);
425 }
426 } else {
427 for (p = 0; p < vol->planes; p++) {
428 vol->fdsp->vector_dmul_scalar((double *)out_buf->extended_data[p],
429 (const double *)buf->extended_data[p],
430 vol->volume, plane_samples);
431 }
432 }
433 }
434
435
2/2
✓ Branch 0 taken 106 times.
✓ Branch 1 taken 93 times.
199 if (buf != out_buf)
436 93 av_frame_free(&buf);
437
438 106 end:
439 199 vol->var_values[VAR_NB_CONSUMED_SAMPLES] += out_buf->nb_samples;
440 199 return ff_filter_frame(outlink, out_buf);
441 }
442
443 static const AVFilterPad avfilter_af_volume_inputs[] = {
444 {
445 .name = "default",
446 .type = AVMEDIA_TYPE_AUDIO,
447 .filter_frame = filter_frame,
448 },
449 };
450
451 static const AVFilterPad avfilter_af_volume_outputs[] = {
452 {
453 .name = "default",
454 .type = AVMEDIA_TYPE_AUDIO,
455 .config_props = config_output,
456 },
457 };
458
459 const FFFilter ff_af_volume = {
460 .p.name = "volume",
461 .p.description = NULL_IF_CONFIG_SMALL("Change input volume."),
462 .p.priv_class = &volume_class,
463 .p.flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC,
464 .priv_size = sizeof(VolumeContext),
465 .init = init,
466 .uninit = uninit,
467 FILTER_INPUTS(avfilter_af_volume_inputs),
468 FILTER_OUTPUTS(avfilter_af_volume_outputs),
469 FILTER_QUERY_FUNC2(query_formats),
470 .process_command = process_command,
471 };
472