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/avstring.h"
22 #include "libavutil/channel_layout.h"
23 #include "libavutil/eval.h"
24 #include "libavutil/intreadwrite.h"
25 #include "libavutil/opt.h"
26 #include "libavutil/parseutils.h"
27 #include "libavutil/xga_font_data.h"
28 #include "avfilter.h"
29 #include "filters.h"
30 #include "formats.h"
31 #include "audio.h"
32 #include "video.h"
33 #include "internal.h"
34
35 static const char *const var_names[] = { "VOLUME", "CHANNEL", "PEAK", NULL };
36 enum { VAR_VOLUME, VAR_CHANNEL, VAR_PEAK, VAR_VARS_NB };
37 enum DisplayScale { LINEAR, LOG, NB_DISPLAY_SCALE };
38
39 typedef struct ShowVolumeContext {
40 const AVClass *class;
41 int w, h;
42 int b;
43 double f;
44 AVRational frame_rate;
45 char *color;
46 int orientation;
47 int step;
48 float bgopacity;
49 int mode;
50
51 int nb_samples;
52 AVFrame *out;
53 AVExpr *c_expr;
54 int draw_text;
55 int draw_volume;
56 double *values;
57 uint32_t *color_lut;
58 float *max;
59 float rms_factor;
60 int display_scale;
61
62 double draw_persistent_duration; /* in second */
63 uint8_t persistant_max_rgba[4];
64 int persistent_max_frames; /* number of frames to check max value */
65 float *max_persistent; /* max value for draw_persistent_max for each channel */
66 int *nb_frames_max_display; /* number of frame for each channel, for displaying the max value */
67
68 void (*meter)(float *src, int nb_samples, float *max, float factor);
69 } ShowVolumeContext;
70
71 #define OFFSET(x) offsetof(ShowVolumeContext, x)
72 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
73
74 static const AVOption showvolume_options[] = {
75 { "rate", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
76 { "r", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
77 { "b", "set border width", OFFSET(b), AV_OPT_TYPE_INT, {.i64=1}, 0, 5, FLAGS },
78 { "w", "set channel width", OFFSET(w), AV_OPT_TYPE_INT, {.i64=400}, 80, 8192, FLAGS },
79 { "h", "set channel height", OFFSET(h), AV_OPT_TYPE_INT, {.i64=20}, 1, 900, FLAGS },
80 { "f", "set fade", OFFSET(f), AV_OPT_TYPE_DOUBLE, {.dbl=0.95}, 0, 1, FLAGS },
81 { "c", "set volume color expression", OFFSET(color), AV_OPT_TYPE_STRING, {.str="PEAK*255+floor((1-PEAK)*255)*256+0xff000000"}, 0, 0, FLAGS },
82 { "t", "display channel names", OFFSET(draw_text), AV_OPT_TYPE_BOOL, {.i64=1}, 0, 1, FLAGS },
83 { "v", "display volume value", OFFSET(draw_volume), AV_OPT_TYPE_BOOL, {.i64=1}, 0, 1, FLAGS },
84 { "dm", "duration for max value display", OFFSET(draw_persistent_duration), AV_OPT_TYPE_DOUBLE, {.dbl=0.}, 0, 9000, FLAGS},
85 { "dmc","set color of the max value line", OFFSET(persistant_max_rgba), AV_OPT_TYPE_COLOR, {.str = "orange"}, 0, 0, FLAGS },
86 { "o", "set orientation", OFFSET(orientation), AV_OPT_TYPE_INT, {.i64=0}, 0, 1, FLAGS, "orientation" },
87 { "h", "horizontal", 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, FLAGS, "orientation" },
88 { "v", "vertical", 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, FLAGS, "orientation" },
89 { "s", "set step size", OFFSET(step), AV_OPT_TYPE_INT, {.i64=0}, 0, 5, FLAGS },
90 { "p", "set background opacity", OFFSET(bgopacity), AV_OPT_TYPE_FLOAT, {.dbl=0}, 0, 1, FLAGS },
91 { "m", "set mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64=0}, 0, 1, FLAGS, "mode" },
92 { "p", "peak", 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, FLAGS, "mode" },
93 { "r", "rms", 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, FLAGS, "mode" },
94 { "ds", "set display scale", OFFSET(display_scale), AV_OPT_TYPE_INT, {.i64=LINEAR}, LINEAR, NB_DISPLAY_SCALE - 1, FLAGS, "display_scale" },
95 { "lin", "linear", 0, AV_OPT_TYPE_CONST, {.i64=LINEAR}, 0, 0, FLAGS, "display_scale" },
96 { "log", "log", 0, AV_OPT_TYPE_CONST, {.i64=LOG}, 0, 0, FLAGS, "display_scale" },
97 { NULL }
98 };
99
100 AVFILTER_DEFINE_CLASS(showvolume);
101
init(AVFilterContext * ctx)102 static av_cold int init(AVFilterContext *ctx)
103 {
104 ShowVolumeContext *s = ctx->priv;
105 int ret;
106
107 if (s->color) {
108 ret = av_expr_parse(&s->c_expr, s->color, var_names,
109 NULL, NULL, NULL, NULL, 0, ctx);
110 if (ret < 0)
111 return ret;
112 }
113
114 return 0;
115 }
116
query_formats(AVFilterContext * ctx)117 static int query_formats(AVFilterContext *ctx)
118 {
119 AVFilterFormats *formats = NULL;
120 AVFilterChannelLayouts *layouts = NULL;
121 AVFilterLink *inlink = ctx->inputs[0];
122 AVFilterLink *outlink = ctx->outputs[0];
123 static const enum AVSampleFormat sample_fmts[] = { AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_NONE };
124 static const enum AVPixelFormat pix_fmts[] = { AV_PIX_FMT_RGBA, AV_PIX_FMT_NONE };
125 int ret;
126
127 formats = ff_make_format_list(sample_fmts);
128 if ((ret = ff_formats_ref(formats, &inlink->outcfg.formats)) < 0)
129 return ret;
130
131 layouts = ff_all_channel_counts();
132 if ((ret = ff_channel_layouts_ref(layouts, &inlink->outcfg.channel_layouts)) < 0)
133 return ret;
134
135 formats = ff_all_samplerates();
136 if ((ret = ff_formats_ref(formats, &inlink->outcfg.samplerates)) < 0)
137 return ret;
138
139 formats = ff_make_format_list(pix_fmts);
140 if ((ret = ff_formats_ref(formats, &outlink->incfg.formats)) < 0)
141 return ret;
142
143 return 0;
144 }
145
find_peak(float * src,int nb_samples,float * peak,float factor)146 static void find_peak(float *src, int nb_samples, float *peak, float factor)
147 {
148 int i;
149
150 *peak = 0;
151 for (i = 0; i < nb_samples; i++)
152 *peak = FFMAX(*peak, FFABS(src[i]));
153 }
154
find_rms(float * src,int nb_samples,float * rms,float factor)155 static void find_rms(float *src, int nb_samples, float *rms, float factor)
156 {
157 int i;
158
159 for (i = 0; i < nb_samples; i++)
160 *rms += factor * (src[i] * src[i] - *rms);
161 }
162
config_input(AVFilterLink * inlink)163 static int config_input(AVFilterLink *inlink)
164 {
165 AVFilterContext *ctx = inlink->dst;
166 ShowVolumeContext *s = ctx->priv;
167
168 s->nb_samples = FFMAX(1, av_rescale(inlink->sample_rate, s->frame_rate.den, s->frame_rate.num));
169 s->values = av_calloc(inlink->channels * VAR_VARS_NB, sizeof(double));
170 if (!s->values)
171 return AVERROR(ENOMEM);
172
173 s->color_lut = av_calloc(s->w, sizeof(*s->color_lut) * inlink->channels);
174 if (!s->color_lut)
175 return AVERROR(ENOMEM);
176
177 s->max = av_calloc(inlink->channels, sizeof(*s->max));
178 if (!s->max)
179 return AVERROR(ENOMEM);
180
181 s->rms_factor = 10000. / inlink->sample_rate;
182
183 switch (s->mode) {
184 case 0: s->meter = find_peak; break;
185 case 1: s->meter = find_rms; break;
186 default: return AVERROR_BUG;
187 }
188
189 if (s->draw_persistent_duration > 0.) {
190 s->persistent_max_frames = (int) FFMAX(av_q2d(s->frame_rate) * s->draw_persistent_duration, 1.);
191 s->max_persistent = av_calloc(inlink->channels * s->persistent_max_frames, sizeof(*s->max_persistent));
192 s->nb_frames_max_display = av_calloc(inlink->channels * s->persistent_max_frames, sizeof(*s->nb_frames_max_display));
193 }
194 return 0;
195 }
196
config_output(AVFilterLink * outlink)197 static int config_output(AVFilterLink *outlink)
198 {
199 ShowVolumeContext *s = outlink->src->priv;
200 AVFilterLink *inlink = outlink->src->inputs[0];
201 int ch;
202
203 if (s->orientation) {
204 outlink->h = s->w;
205 outlink->w = s->h * inlink->channels + (inlink->channels - 1) * s->b;
206 } else {
207 outlink->w = s->w;
208 outlink->h = s->h * inlink->channels + (inlink->channels - 1) * s->b;
209 }
210
211 outlink->sample_aspect_ratio = (AVRational){1,1};
212 outlink->frame_rate = s->frame_rate;
213
214 for (ch = 0; ch < inlink->channels; ch++) {
215 int i;
216
217 for (i = 0; i < s->w; i++) {
218 float max = i / (float)(s->w - 1);
219
220 s->values[ch * VAR_VARS_NB + VAR_PEAK] = max;
221 s->values[ch * VAR_VARS_NB + VAR_VOLUME] = 20.0 * log10(max);
222 s->values[ch * VAR_VARS_NB + VAR_CHANNEL] = ch;
223 s->color_lut[ch * s->w + i] = av_expr_eval(s->c_expr, &s->values[ch * VAR_VARS_NB], NULL);
224 }
225 }
226
227 return 0;
228 }
229
drawtext(AVFrame * pic,int x,int y,const char * txt,int o)230 static void drawtext(AVFrame *pic, int x, int y, const char *txt, int o)
231 {
232 const uint8_t *font;
233 int font_height;
234 int i;
235
236 font = avpriv_cga_font, font_height = 8;
237
238 for (i = 0; txt[i]; i++) {
239 int char_y, mask;
240
241 if (o) { /* vertical orientation */
242 for (char_y = font_height - 1; char_y >= 0; char_y--) {
243 uint8_t *p = pic->data[0] + (y + i * 10) * pic->linesize[0] + x * 4;
244 for (mask = 0x80; mask; mask >>= 1) {
245 if (font[txt[i] * font_height + font_height - 1 - char_y] & mask)
246 AV_WN32(&p[char_y * 4], ~AV_RN32(&p[char_y * 4]));
247 p += pic->linesize[0];
248 }
249 }
250 } else { /* horizontal orientation */
251 uint8_t *p = pic->data[0] + y * pic->linesize[0] + (x + i * 8) * 4;
252 for (char_y = 0; char_y < font_height; char_y++) {
253 for (mask = 0x80; mask; mask >>= 1) {
254 if (font[txt[i] * font_height + char_y] & mask)
255 AV_WN32(p, ~AV_RN32(p));
256 p += 4;
257 }
258 p += pic->linesize[0] - 8 * 4;
259 }
260 }
261 }
262 }
263
clear_picture(ShowVolumeContext * s,AVFilterLink * outlink)264 static void clear_picture(ShowVolumeContext *s, AVFilterLink *outlink)
265 {
266 int i, j;
267 const uint32_t bg = (uint32_t)(s->bgopacity * 255) << 24;
268
269 for (i = 0; i < outlink->h; i++) {
270 uint32_t *dst = (uint32_t *)(s->out->data[0] + i * s->out->linesize[0]);
271 for (j = 0; j < outlink->w; j++)
272 AV_WN32A(dst + j, bg);
273 }
274 }
275
calc_max_draw(ShowVolumeContext * s,AVFilterLink * outlink,float max)276 static inline int calc_max_draw(ShowVolumeContext *s, AVFilterLink *outlink, float max)
277 {
278 float max_val;
279 if (s->display_scale == LINEAR) {
280 max_val = max;
281 } else { /* log */
282 max_val = av_clipf(0.21 * log10(max) + 1, 0, 1);
283 }
284 if (s->orientation) { /* vertical */
285 return outlink->h - outlink->h * max_val;
286 } else { /* horizontal */
287 return s->w * max_val;
288 }
289 }
290
calc_persistent_max(ShowVolumeContext * s,float max,int channel)291 static inline void calc_persistent_max(ShowVolumeContext *s, float max, int channel)
292 {
293 /* update max value for persistent max display */
294 if ((max >= s->max_persistent[channel]) || (s->nb_frames_max_display[channel] >= s->persistent_max_frames)) { /* update max value for display */
295 s->max_persistent[channel] = max;
296 s->nb_frames_max_display[channel] = 0;
297 } else {
298 s->nb_frames_max_display[channel] += 1; /* incremente display frame count */
299 }
300 }
301
draw_max_line(ShowVolumeContext * s,int max_draw,int channel)302 static inline void draw_max_line(ShowVolumeContext *s, int max_draw, int channel)
303 {
304 int k;
305 if (s->orientation) { /* vertical */
306 uint8_t *dst = s->out->data[0] + max_draw * s->out->linesize[0] + channel * (s->b + s->h) * 4;
307 for (k = 0; k < s->h; k++) {
308 memcpy(dst + k * 4, s->persistant_max_rgba, sizeof(s->persistant_max_rgba));
309 }
310 } else { /* horizontal */
311 for (k = 0; k < s->h; k++) {
312 uint8_t *dst = s->out->data[0] + (channel * s->h + channel * s->b + k) * s->out->linesize[0];
313 memcpy(dst + max_draw * 4, s->persistant_max_rgba, sizeof(s->persistant_max_rgba));
314 }
315 }
316 }
317
filter_frame(AVFilterLink * inlink,AVFrame * insamples)318 static int filter_frame(AVFilterLink *inlink, AVFrame *insamples)
319 {
320 AVFilterContext *ctx = inlink->dst;
321 AVFilterLink *outlink = ctx->outputs[0];
322 ShowVolumeContext *s = ctx->priv;
323 const int step = s->step;
324 int c, j, k, max_draw;
325 AVFrame *out;
326
327 if (!s->out || s->out->width != outlink->w ||
328 s->out->height != outlink->h) {
329 av_frame_free(&s->out);
330 s->out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
331 if (!s->out) {
332 av_frame_free(&insamples);
333 return AVERROR(ENOMEM);
334 }
335 clear_picture(s, outlink);
336 }
337 s->out->pts = insamples->pts;
338
339 if ((s->f < 1.) && (s->f > 0.)) {
340 for (j = 0; j < outlink->h; j++) {
341 uint8_t *dst = s->out->data[0] + j * s->out->linesize[0];
342 const uint32_t alpha = s->bgopacity * 255;
343
344 for (k = 0; k < outlink->w; k++) {
345 dst[k * 4 + 0] = FFMAX(dst[k * 4 + 0] * s->f, 0);
346 dst[k * 4 + 1] = FFMAX(dst[k * 4 + 1] * s->f, 0);
347 dst[k * 4 + 2] = FFMAX(dst[k * 4 + 2] * s->f, 0);
348 dst[k * 4 + 3] = FFMAX(dst[k * 4 + 3] * s->f, alpha);
349 }
350 }
351 } else if (s->f == 0.) {
352 clear_picture(s, outlink);
353 }
354
355 if (s->orientation) { /* vertical */
356 for (c = 0; c < inlink->channels; c++) {
357 float *src = (float *)insamples->extended_data[c];
358 uint32_t *lut = s->color_lut + s->w * c;
359 float max;
360
361 s->meter(src, insamples->nb_samples, &s->max[c], s->rms_factor);
362 max = s->max[c];
363
364 s->values[c * VAR_VARS_NB + VAR_VOLUME] = 20.0 * log10(max);
365 max = av_clipf(max, 0, 1);
366 max_draw = calc_max_draw(s, outlink, max);
367
368 for (j = max_draw; j < s->w; j++) {
369 uint8_t *dst = s->out->data[0] + j * s->out->linesize[0] + c * (s->b + s->h) * 4;
370 for (k = 0; k < s->h; k++) {
371 AV_WN32A(&dst[k * 4], lut[s->w - j - 1]);
372 if (j & step)
373 j += step;
374 }
375 }
376
377 if (s->h >= 8 && s->draw_text) {
378 const char *channel_name = av_get_channel_name(av_channel_layout_extract_channel(insamples->channel_layout, c));
379 if (!channel_name)
380 continue;
381 drawtext(s->out, c * (s->h + s->b) + (s->h - 10) / 2, outlink->h - 35, channel_name, 1);
382 }
383
384 if (s->draw_persistent_duration > 0.) {
385 calc_persistent_max(s, max, c);
386 max_draw = FFMAX(0, calc_max_draw(s, outlink, s->max_persistent[c]) - 1);
387 draw_max_line(s, max_draw, c);
388 }
389 }
390 } else { /* horizontal */
391 for (c = 0; c < inlink->channels; c++) {
392 float *src = (float *)insamples->extended_data[c];
393 uint32_t *lut = s->color_lut + s->w * c;
394 float max;
395
396 s->meter(src, insamples->nb_samples, &s->max[c], s->rms_factor);
397 max = s->max[c];
398
399 s->values[c * VAR_VARS_NB + VAR_VOLUME] = 20.0 * log10(max);
400 max = av_clipf(max, 0, 1);
401 max_draw = calc_max_draw(s, outlink, max);
402
403 for (j = 0; j < s->h; j++) {
404 uint8_t *dst = s->out->data[0] + (c * s->h + c * s->b + j) * s->out->linesize[0];
405
406 for (k = 0; k < max_draw; k++) {
407 AV_WN32A(dst + k * 4, lut[k]);
408 if (k & step)
409 k += step;
410 }
411 }
412
413 if (s->h >= 8 && s->draw_text) {
414 const char *channel_name = av_get_channel_name(av_channel_layout_extract_channel(insamples->channel_layout, c));
415 if (!channel_name)
416 continue;
417 drawtext(s->out, 2, c * (s->h + s->b) + (s->h - 8) / 2, channel_name, 0);
418 }
419
420 if (s->draw_persistent_duration > 0.) {
421 calc_persistent_max(s, max, c);
422 max_draw = FFMAX(0, calc_max_draw(s, outlink, s->max_persistent[c]) - 1);
423 draw_max_line(s, max_draw, c);
424 }
425 }
426 }
427
428 av_frame_free(&insamples);
429 out = av_frame_clone(s->out);
430 if (!out)
431 return AVERROR(ENOMEM);
432 av_frame_make_writable(out);
433
434 /* draw volume level */
435 for (c = 0; c < inlink->channels && s->h >= 8 && s->draw_volume; c++) {
436 char buf[16];
437
438 if (s->orientation) { /* vertical */
439 snprintf(buf, sizeof(buf), "%.2f", s->values[c * VAR_VARS_NB + VAR_VOLUME]);
440 drawtext(out, c * (s->h + s->b) + (s->h - 8) / 2, 2, buf, 1);
441 } else { /* horizontal */
442 snprintf(buf, sizeof(buf), "%.2f", s->values[c * VAR_VARS_NB + VAR_VOLUME]);
443 drawtext(out, FFMAX(0, s->w - 8 * (int)strlen(buf)), c * (s->h + s->b) + (s->h - 8) / 2, buf, 0);
444 }
445 }
446
447 return ff_filter_frame(outlink, out);
448 }
449
activate(AVFilterContext * ctx)450 static int activate(AVFilterContext *ctx)
451 {
452 AVFilterLink *inlink = ctx->inputs[0];
453 AVFilterLink *outlink = ctx->outputs[0];
454 ShowVolumeContext *s = ctx->priv;
455 AVFrame *in = NULL;
456 int ret;
457
458 FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
459
460 ret = ff_inlink_consume_samples(inlink, s->nb_samples, s->nb_samples, &in);
461 if (ret < 0)
462 return ret;
463 if (ret > 0)
464 return filter_frame(inlink, in);
465
466 FF_FILTER_FORWARD_STATUS(inlink, outlink);
467 FF_FILTER_FORWARD_WANTED(outlink, inlink);
468
469 return FFERROR_NOT_READY;
470 }
471
uninit(AVFilterContext * ctx)472 static av_cold void uninit(AVFilterContext *ctx)
473 {
474 ShowVolumeContext *s = ctx->priv;
475
476 av_frame_free(&s->out);
477 av_expr_free(s->c_expr);
478 av_freep(&s->values);
479 av_freep(&s->color_lut);
480 av_freep(&s->max);
481 }
482
483 static const AVFilterPad showvolume_inputs[] = {
484 {
485 .name = "default",
486 .type = AVMEDIA_TYPE_AUDIO,
487 .config_props = config_input,
488 },
489 { NULL }
490 };
491
492 static const AVFilterPad showvolume_outputs[] = {
493 {
494 .name = "default",
495 .type = AVMEDIA_TYPE_VIDEO,
496 .config_props = config_output,
497 },
498 { NULL }
499 };
500
501 AVFilter ff_avf_showvolume = {
502 .name = "showvolume",
503 .description = NULL_IF_CONFIG_SMALL("Convert input audio volume to video output."),
504 .init = init,
505 .activate = activate,
506 .uninit = uninit,
507 .query_formats = query_formats,
508 .priv_size = sizeof(ShowVolumeContext),
509 .inputs = showvolume_inputs,
510 .outputs = showvolume_outputs,
511 .priv_class = &showvolume_class,
512 };
513