FFmpeg coverage


Directory: ../../../ffmpeg/
File: src/libavfilter/avf_showvolume.c
Date: 2024-11-20 23:03:26
Exec Total Coverage
Lines: 0 236 0.0%
Functions: 0 14 0.0%
Branches: 0 159 0.0%

Line Branch Exec Source
1 /*
2 * Copyright (c) 2015 Paul B Mahol
3 *
4 * This file is part of FFmpeg.
5 *
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.
10 *
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.
15 *
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
19 */
20
21 #include "libavutil/channel_layout.h"
22 #include "libavutil/eval.h"
23 #include "libavutil/intreadwrite.h"
24 #include "libavutil/mem.h"
25 #include "libavutil/opt.h"
26 #include "libavutil/xga_font_data.h"
27 #include "avfilter.h"
28 #include "filters.h"
29 #include "formats.h"
30 #include "video.h"
31
32 static const char *const var_names[] = { "VOLUME", "CHANNEL", "PEAK", NULL };
33 enum { VAR_VOLUME, VAR_CHANNEL, VAR_PEAK, VAR_VARS_NB };
34 enum DisplayScale { LINEAR, LOG, NB_DISPLAY_SCALE };
35
36 typedef struct ShowVolumeContext {
37 const AVClass *class;
38 int w, h;
39 int b;
40 double f;
41 AVRational frame_rate;
42 char *color;
43 int orientation;
44 int step;
45 float bgopacity;
46 int mode;
47
48 int nb_samples;
49 AVFrame *out;
50 AVExpr *c_expr;
51 int draw_text;
52 int draw_volume;
53 double *values;
54 uint32_t *color_lut;
55 float *max;
56 int display_scale;
57
58 double draw_persistent_duration; /* in second */
59 uint8_t persistant_max_rgba[4];
60 int persistent_max_frames; /* number of frames to check max value */
61 float *max_persistent; /* max value for draw_persistent_max for each channel */
62 int *nb_frames_max_display; /* number of frame for each channel, for displaying the max value */
63
64 void (*meter)(float *src, int nb_samples, float *max);
65 } ShowVolumeContext;
66
67 #define OFFSET(x) offsetof(ShowVolumeContext, x)
68 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
69
70 static const AVOption showvolume_options[] = {
71 { "rate", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
72 { "r", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
73 { "b", "set border width", OFFSET(b), AV_OPT_TYPE_INT, {.i64=1}, 0, 5, FLAGS },
74 { "w", "set channel width", OFFSET(w), AV_OPT_TYPE_INT, {.i64=400}, 80, 8192, FLAGS },
75 { "h", "set channel height", OFFSET(h), AV_OPT_TYPE_INT, {.i64=20}, 1, 900, FLAGS },
76 { "f", "set fade", OFFSET(f), AV_OPT_TYPE_DOUBLE, {.dbl=0.95}, 0, 1, FLAGS },
77 { "c", "set volume color expression", OFFSET(color), AV_OPT_TYPE_STRING, {.str="PEAK*255+floor((1-PEAK)*255)*256+0xff000000"}, 0, 0, FLAGS },
78 { "t", "display channel names", OFFSET(draw_text), AV_OPT_TYPE_BOOL, {.i64=1}, 0, 1, FLAGS },
79 { "v", "display volume value", OFFSET(draw_volume), AV_OPT_TYPE_BOOL, {.i64=1}, 0, 1, FLAGS },
80 { "dm", "duration for max value display", OFFSET(draw_persistent_duration), AV_OPT_TYPE_DOUBLE, {.dbl=0.}, 0, 9000, FLAGS},
81 { "dmc","set color of the max value line", OFFSET(persistant_max_rgba), AV_OPT_TYPE_COLOR, {.str = "orange"}, 0, 0, FLAGS },
82 { "o", "set orientation", OFFSET(orientation), AV_OPT_TYPE_INT, {.i64=0}, 0, 1, FLAGS, .unit = "orientation" },
83 { "h", "horizontal", 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, FLAGS, .unit = "orientation" },
84 { "v", "vertical", 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, FLAGS, .unit = "orientation" },
85 { "s", "set step size", OFFSET(step), AV_OPT_TYPE_INT, {.i64=0}, 0, 5, FLAGS },
86 { "p", "set background opacity", OFFSET(bgopacity), AV_OPT_TYPE_FLOAT, {.dbl=0}, 0, 1, FLAGS },
87 { "m", "set mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64=0}, 0, 1, FLAGS, .unit = "mode" },
88 { "p", "peak", 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, FLAGS, .unit = "mode" },
89 { "r", "rms", 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, FLAGS, .unit = "mode" },
90 { "ds", "set display scale", OFFSET(display_scale), AV_OPT_TYPE_INT, {.i64=LINEAR}, LINEAR, NB_DISPLAY_SCALE - 1, FLAGS, .unit = "display_scale" },
91 { "lin", "linear", 0, AV_OPT_TYPE_CONST, {.i64=LINEAR}, 0, 0, FLAGS, .unit = "display_scale" },
92 { "log", "log", 0, AV_OPT_TYPE_CONST, {.i64=LOG}, 0, 0, FLAGS, .unit = "display_scale" },
93 { NULL }
94 };
95
96 AVFILTER_DEFINE_CLASS(showvolume);
97
98 static av_cold int init(AVFilterContext *ctx)
99 {
100 ShowVolumeContext *s = ctx->priv;
101 int ret;
102
103 if (s->color) {
104 ret = av_expr_parse(&s->c_expr, s->color, var_names,
105 NULL, NULL, NULL, NULL, 0, ctx);
106 if (ret < 0)
107 return ret;
108 }
109
110 return 0;
111 }
112
113 static int query_formats(const AVFilterContext *ctx,
114 AVFilterFormatsConfig **cfg_in,
115 AVFilterFormatsConfig **cfg_out)
116 {
117 AVFilterFormats *formats = NULL;
118 static const enum AVSampleFormat sample_fmts[] = { AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_NONE };
119 static const enum AVPixelFormat pix_fmts[] = { AV_PIX_FMT_RGBA, AV_PIX_FMT_NONE };
120 int ret;
121
122 formats = ff_make_format_list(sample_fmts);
123 if ((ret = ff_formats_ref(formats, &cfg_in[0]->formats)) < 0)
124 return ret;
125
126 formats = ff_make_format_list(pix_fmts);
127 if ((ret = ff_formats_ref(formats, &cfg_out[0]->formats)) < 0)
128 return ret;
129
130 return 0;
131 }
132
133 static void find_peak(float *src, int nb_samples, float *peak)
134 {
135 float max = 0.f;
136
137 max = 0;
138 for (int i = 0; i < nb_samples; i++)
139 max = fmaxf(max, fabsf(src[i]));
140 *peak = max;
141 }
142
143 static void find_rms(float *src, int nb_samples, float *rms)
144 {
145 float sum = 0.f;
146
147 for (int i = 0; i < nb_samples; i++)
148 sum += src[i] * src[i];
149 *rms = sqrtf(sum / nb_samples);
150 }
151
152 static int config_input(AVFilterLink *inlink)
153 {
154 AVFilterContext *ctx = inlink->dst;
155 ShowVolumeContext *s = ctx->priv;
156
157 s->nb_samples = FFMAX(1, av_rescale(inlink->sample_rate, s->frame_rate.den, s->frame_rate.num));
158 s->values = av_calloc(inlink->ch_layout.nb_channels * VAR_VARS_NB, sizeof(double));
159 if (!s->values)
160 return AVERROR(ENOMEM);
161
162 s->color_lut = av_calloc(s->w, sizeof(*s->color_lut) * inlink->ch_layout.nb_channels);
163 if (!s->color_lut)
164 return AVERROR(ENOMEM);
165
166 s->max = av_calloc(inlink->ch_layout.nb_channels, sizeof(*s->max));
167 if (!s->max)
168 return AVERROR(ENOMEM);
169
170 switch (s->mode) {
171 case 0: s->meter = find_peak; break;
172 case 1: s->meter = find_rms; break;
173 default: return AVERROR_BUG;
174 }
175
176 if (s->draw_persistent_duration > 0.) {
177 s->persistent_max_frames = (int) FFMAX(av_q2d(s->frame_rate) * s->draw_persistent_duration, 1.);
178 s->max_persistent = av_calloc(inlink->ch_layout.nb_channels * s->persistent_max_frames, sizeof(*s->max_persistent));
179 s->nb_frames_max_display = av_calloc(inlink->ch_layout.nb_channels * s->persistent_max_frames, sizeof(*s->nb_frames_max_display));
180 if (!s->max_persistent ||
181 !s->nb_frames_max_display)
182 return AVERROR(ENOMEM);
183 }
184 return 0;
185 }
186
187 static int config_output(AVFilterLink *outlink)
188 {
189 FilterLink *l = ff_filter_link(outlink);
190 ShowVolumeContext *s = outlink->src->priv;
191 AVFilterLink *inlink = outlink->src->inputs[0];
192 int ch;
193
194 if (s->orientation) {
195 outlink->h = s->w;
196 outlink->w = s->h * inlink->ch_layout.nb_channels + (inlink->ch_layout.nb_channels - 1) * s->b;
197 } else {
198 outlink->w = s->w;
199 outlink->h = s->h * inlink->ch_layout.nb_channels + (inlink->ch_layout.nb_channels - 1) * s->b;
200 }
201
202 outlink->sample_aspect_ratio = (AVRational){1,1};
203 l->frame_rate = s->frame_rate;
204 outlink->time_base = av_inv_q(l->frame_rate);
205
206 for (ch = 0; ch < inlink->ch_layout.nb_channels; ch++) {
207 int i;
208
209 for (i = 0; i < s->w; i++) {
210 float max = i / (float)(s->w - 1);
211
212 s->values[ch * VAR_VARS_NB + VAR_PEAK] = max;
213 s->values[ch * VAR_VARS_NB + VAR_VOLUME] = 20.0 * log10(max);
214 s->values[ch * VAR_VARS_NB + VAR_CHANNEL] = ch;
215 s->color_lut[ch * s->w + i] = av_expr_eval(s->c_expr, &s->values[ch * VAR_VARS_NB], NULL);
216 }
217 }
218
219 return 0;
220 }
221
222 static void drawtext(AVFrame *pic, int x, int y, const char *txt, int o)
223 {
224 const uint8_t *font;
225 int font_height;
226 int i;
227
228 font = avpriv_cga_font, font_height = 8;
229
230 for (i = 0; txt[i]; i++) {
231 int char_y, mask;
232
233 if (o) { /* vertical orientation */
234 for (char_y = font_height - 1; char_y >= 0; char_y--) {
235 uint8_t *p = pic->data[0] + (y + i * 10) * pic->linesize[0] + x * 4;
236 for (mask = 0x80; mask; mask >>= 1) {
237 if (font[txt[i] * font_height + font_height - 1 - char_y] & mask)
238 AV_WN32(&p[char_y * 4], ~AV_RN32(&p[char_y * 4]));
239 p += pic->linesize[0];
240 }
241 }
242 } else { /* horizontal orientation */
243 uint8_t *p = pic->data[0] + y * pic->linesize[0] + (x + i * 8) * 4;
244 for (char_y = 0; char_y < font_height; char_y++) {
245 for (mask = 0x80; mask; mask >>= 1) {
246 if (font[txt[i] * font_height + char_y] & mask)
247 AV_WN32(p, ~AV_RN32(p));
248 p += 4;
249 }
250 p += pic->linesize[0] - 8 * 4;
251 }
252 }
253 }
254 }
255
256 static void clear_picture(ShowVolumeContext *s, AVFilterLink *outlink)
257 {
258 int i, j;
259 const uint32_t bg = (uint32_t)(s->bgopacity * 255) << 24;
260
261 for (i = 0; i < outlink->h; i++) {
262 uint32_t *dst = (uint32_t *)(s->out->data[0] + i * s->out->linesize[0]);
263 for (j = 0; j < outlink->w; j++)
264 AV_WN32A(dst + j, bg);
265 }
266 }
267
268 static inline int calc_max_draw(ShowVolumeContext *s, AVFilterLink *outlink, float max)
269 {
270 float max_val;
271 if (s->display_scale == LINEAR) {
272 max_val = max;
273 } else { /* log */
274 max_val = av_clipf(0.21 * log10(max) + 1, 0, 1);
275 }
276 if (s->orientation) { /* vertical */
277 return outlink->h - outlink->h * max_val;
278 } else { /* horizontal */
279 return s->w * max_val;
280 }
281 }
282
283 static inline void calc_persistent_max(ShowVolumeContext *s, float max, int channel)
284 {
285 /* update max value for persistent max display */
286 if ((max >= s->max_persistent[channel]) || (s->nb_frames_max_display[channel] >= s->persistent_max_frames)) { /* update max value for display */
287 s->max_persistent[channel] = max;
288 s->nb_frames_max_display[channel] = 0;
289 } else {
290 s->nb_frames_max_display[channel] += 1; /* incremente display frame count */
291 }
292 }
293
294 static inline void draw_max_line(ShowVolumeContext *s, int max_draw, int channel)
295 {
296 int k;
297 if (s->orientation) { /* vertical */
298 uint8_t *dst = s->out->data[0] + max_draw * s->out->linesize[0] + channel * (s->b + s->h) * 4;
299 for (k = 0; k < s->h; k++) {
300 memcpy(dst + k * 4, s->persistant_max_rgba, sizeof(s->persistant_max_rgba));
301 }
302 } else { /* horizontal */
303 for (k = 0; k < s->h; k++) {
304 uint8_t *dst = s->out->data[0] + (channel * s->h + channel * s->b + k) * s->out->linesize[0];
305 memcpy(dst + max_draw * 4, s->persistant_max_rgba, sizeof(s->persistant_max_rgba));
306 }
307 }
308 }
309
310 static int filter_frame(AVFilterLink *inlink, AVFrame *insamples)
311 {
312 AVFilterContext *ctx = inlink->dst;
313 AVFilterLink *outlink = ctx->outputs[0];
314 ShowVolumeContext *s = ctx->priv;
315 const int step = s->step;
316 int c, j, k, max_draw, ret;
317 char channel_name[64];
318 AVFrame *out;
319
320 if (!s->out || s->out->width != outlink->w ||
321 s->out->height != outlink->h) {
322 av_frame_free(&s->out);
323 s->out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
324 if (!s->out) {
325 av_frame_free(&insamples);
326 return AVERROR(ENOMEM);
327 }
328 clear_picture(s, outlink);
329 }
330 s->out->pts = av_rescale_q(insamples->pts, inlink->time_base, outlink->time_base);
331 s->out->duration = 1;
332
333 if ((s->f < 1.) && (s->f > 0.)) {
334 for (j = 0; j < outlink->h; j++) {
335 uint8_t *dst = s->out->data[0] + j * s->out->linesize[0];
336 const uint32_t alpha = s->bgopacity * 255;
337
338 for (k = 0; k < outlink->w; k++) {
339 dst[k * 4 + 0] = FFMAX(dst[k * 4 + 0] * s->f, 0);
340 dst[k * 4 + 1] = FFMAX(dst[k * 4 + 1] * s->f, 0);
341 dst[k * 4 + 2] = FFMAX(dst[k * 4 + 2] * s->f, 0);
342 dst[k * 4 + 3] = FFMAX(dst[k * 4 + 3] * s->f, alpha);
343 }
344 }
345 } else if (s->f == 0.) {
346 clear_picture(s, outlink);
347 }
348
349 if (s->orientation) { /* vertical */
350 for (c = 0; c < inlink->ch_layout.nb_channels; c++) {
351 float *src = (float *)insamples->extended_data[c];
352 uint32_t *lut = s->color_lut + s->w * c;
353 float max;
354
355 s->meter(src, insamples->nb_samples, &s->max[c]);
356 max = s->max[c];
357
358 s->values[c * VAR_VARS_NB + VAR_VOLUME] = 20.0 * log10(max);
359 max = av_clipf(max, 0, 1);
360 max_draw = calc_max_draw(s, outlink, max);
361
362 for (j = s->w - 1; j >= max_draw; j--) {
363 uint8_t *dst = s->out->data[0] + j * s->out->linesize[0] + c * (s->b + s->h) * 4;
364 for (k = 0; k < s->h; k++) {
365 AV_WN32A(&dst[k * 4], lut[s->w - j - 1]);
366 }
367 if (j & step)
368 j -= step;
369 }
370
371 if (s->draw_persistent_duration > 0.) {
372 calc_persistent_max(s, max, c);
373 max_draw = FFMAX(0, calc_max_draw(s, outlink, s->max_persistent[c]) - 1);
374 draw_max_line(s, max_draw, c);
375 }
376 }
377 } else { /* horizontal */
378 for (c = 0; c < inlink->ch_layout.nb_channels; c++) {
379 float *src = (float *)insamples->extended_data[c];
380 uint32_t *lut = s->color_lut + s->w * c;
381 float max;
382
383 s->meter(src, insamples->nb_samples, &s->max[c]);
384 max = s->max[c];
385
386 s->values[c * VAR_VARS_NB + VAR_VOLUME] = 20.0 * log10(max);
387 max = av_clipf(max, 0, 1);
388 max_draw = calc_max_draw(s, outlink, max);
389
390 for (j = 0; j < s->h; j++) {
391 uint8_t *dst = s->out->data[0] + (c * s->h + c * s->b + j) * s->out->linesize[0];
392
393 for (k = 0; k < max_draw; k++) {
394 AV_WN32A(dst + k * 4, lut[k]);
395 if (k & step)
396 k += step;
397 }
398 }
399
400 if (s->draw_persistent_duration > 0.) {
401 calc_persistent_max(s, max, c);
402 max_draw = FFMAX(0, calc_max_draw(s, outlink, s->max_persistent[c]) - 1);
403 draw_max_line(s, max_draw, c);
404 }
405 }
406 }
407
408 av_frame_free(&insamples);
409 out = av_frame_clone(s->out);
410 if (!out)
411 return AVERROR(ENOMEM);
412 ret = ff_inlink_make_frame_writable(outlink, &out);
413 if (ret < 0) {
414 av_frame_free(&out);
415 return ret;
416 }
417
418 /* draw channel names */
419 for (c = 0; c < inlink->ch_layout.nb_channels && s->h >= 10 && s->draw_text; c++) {
420 if (s->orientation) { /* vertical */
421 int ret = av_channel_name(channel_name, sizeof(channel_name), av_channel_layout_channel_from_index(&inlink->ch_layout, c));
422 if (ret < 0)
423 continue;
424 drawtext(out, c * (s->h + s->b) + (s->h - 10) / 2, outlink->h - 35, channel_name, 1);
425 } else { /* horizontal */
426 int ret = av_channel_name(channel_name, sizeof(channel_name), av_channel_layout_channel_from_index(&inlink->ch_layout, c));
427 if (ret < 0)
428 continue;
429 drawtext(out, 2, c * (s->h + s->b) + (s->h - 8) / 2, channel_name, 0);
430 }
431 }
432
433 /* draw volume level */
434 for (c = 0; c < inlink->ch_layout.nb_channels && s->h >= 8 && s->draw_volume; c++) {
435 char buf[16];
436
437 if (s->orientation) { /* vertical */
438 snprintf(buf, sizeof(buf), "%.2f", s->values[c * VAR_VARS_NB + VAR_VOLUME]);
439 drawtext(out, c * (s->h + s->b) + (s->h - 8) / 2, 2, buf, 1);
440 } else { /* horizontal */
441 snprintf(buf, sizeof(buf), "%.2f", s->values[c * VAR_VARS_NB + VAR_VOLUME]);
442 drawtext(out, FFMAX(0, s->w - 8 * (int)strlen(buf)), c * (s->h + s->b) + (s->h - 8) / 2, buf, 0);
443 }
444 }
445
446 return ff_filter_frame(outlink, out);
447 }
448
449 static int activate(AVFilterContext *ctx)
450 {
451 AVFilterLink *inlink = ctx->inputs[0];
452 AVFilterLink *outlink = ctx->outputs[0];
453 ShowVolumeContext *s = ctx->priv;
454 AVFrame *in = NULL;
455 int ret;
456
457 FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
458
459 ret = ff_inlink_consume_samples(inlink, s->nb_samples, s->nb_samples, &in);
460 if (ret < 0)
461 return ret;
462 if (ret > 0)
463 return filter_frame(inlink, in);
464
465 if (ff_inlink_queued_samples(inlink) >= s->nb_samples) {
466 ff_filter_set_ready(ctx, 10);
467 return 0;
468 }
469
470 FF_FILTER_FORWARD_STATUS(inlink, outlink);
471 FF_FILTER_FORWARD_WANTED(outlink, inlink);
472
473 return FFERROR_NOT_READY;
474 }
475
476 static av_cold void uninit(AVFilterContext *ctx)
477 {
478 ShowVolumeContext *s = ctx->priv;
479
480 av_frame_free(&s->out);
481 av_expr_free(s->c_expr);
482 av_freep(&s->values);
483 av_freep(&s->color_lut);
484 av_freep(&s->max);
485 av_freep(&s->max_persistent);
486 av_freep(&s->nb_frames_max_display);
487 }
488
489 static const AVFilterPad showvolume_inputs[] = {
490 {
491 .name = "default",
492 .type = AVMEDIA_TYPE_AUDIO,
493 .config_props = config_input,
494 },
495 };
496
497 static const AVFilterPad showvolume_outputs[] = {
498 {
499 .name = "default",
500 .type = AVMEDIA_TYPE_VIDEO,
501 .config_props = config_output,
502 },
503 };
504
505 const AVFilter ff_avf_showvolume = {
506 .name = "showvolume",
507 .description = NULL_IF_CONFIG_SMALL("Convert input audio volume to video output."),
508 .init = init,
509 .activate = activate,
510 .uninit = uninit,
511 .priv_size = sizeof(ShowVolumeContext),
512 FILTER_INPUTS(showvolume_inputs),
513 FILTER_OUTPUTS(showvolume_outputs),
514 FILTER_QUERY_FUNC2(query_formats),
515 .priv_class = &showvolume_class,
516 };
517