2 * Copyright (c) 2015 Paul B Mahol
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
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"
34 static const char *const var_names[] = { "VOLUME", "CHANNEL", "PEAK", NULL };
35 enum { VAR_VOLUME, VAR_CHANNEL, VAR_PEAK, VAR_VARS_NB };
36 enum DisplayScale { LINEAR, LOG, NB_DISPLAY_SCALE };
38 typedef struct ShowVolumeContext {
43 AVRational frame_rate;
60 double draw_persistent_duration; /* in second */
61 uint8_t persistant_max_rgba[4];
62 int persistent_max_frames; /* number of frames to check max value */
63 float *max_persistent; /* max value for draw_persistent_max for each channel */
64 int *nb_frames_max_display; /* number of frame for each channel, for displaying the max value */
66 void (*meter)(float *src, int nb_samples, float *max, float factor);
69 #define OFFSET(x) offsetof(ShowVolumeContext, x)
70 #define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
72 static const AVOption showvolume_options[] = {
73 { "rate", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
74 { "r", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
75 { "b", "set border width", OFFSET(b), AV_OPT_TYPE_INT, {.i64=1}, 0, 5, FLAGS },
76 { "w", "set channel width", OFFSET(w), AV_OPT_TYPE_INT, {.i64=400}, 80, 8192, FLAGS },
77 { "h", "set channel height", OFFSET(h), AV_OPT_TYPE_INT, {.i64=20}, 1, 900, FLAGS },
78 { "f", "set fade", OFFSET(f), AV_OPT_TYPE_DOUBLE, {.dbl=0.95}, 0, 1, FLAGS },
79 { "c", "set volume color expression", OFFSET(color), AV_OPT_TYPE_STRING, {.str="PEAK*255+floor((1-PEAK)*255)*256+0xff000000"}, 0, 0, FLAGS },
80 { "t", "display channel names", OFFSET(draw_text), AV_OPT_TYPE_BOOL, {.i64=1}, 0, 1, FLAGS },
81 { "v", "display volume value", OFFSET(draw_volume), AV_OPT_TYPE_BOOL, {.i64=1}, 0, 1, FLAGS },
82 { "dm", "duration for max value display", OFFSET(draw_persistent_duration), AV_OPT_TYPE_DOUBLE, {.dbl=0.}, 0, 9000, FLAGS},
83 { "dmc","set color of the max value line", OFFSET(persistant_max_rgba), AV_OPT_TYPE_COLOR, {.str = "orange"}, CHAR_MIN, CHAR_MAX, FLAGS },
84 { "o", "set orientation", OFFSET(orientation), AV_OPT_TYPE_INT, {.i64=0}, 0, 1, FLAGS, "orientation" },
85 { "h", "horizontal", 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, FLAGS, "orientation" },
86 { "v", "vertical", 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, FLAGS, "orientation" },
87 { "s", "set step size", OFFSET(step), AV_OPT_TYPE_INT, {.i64=0}, 0, 5, FLAGS },
88 { "p", "set background opacity", OFFSET(bgopacity), AV_OPT_TYPE_FLOAT, {.dbl=0}, 0, 1, FLAGS },
89 { "m", "set mode", OFFSET(mode), AV_OPT_TYPE_INT, {.i64=0}, 0, 1, FLAGS, "mode" },
90 { "p", "peak", 0, AV_OPT_TYPE_CONST, {.i64=0}, 0, 0, FLAGS, "mode" },
91 { "r", "rms", 0, AV_OPT_TYPE_CONST, {.i64=1}, 0, 0, FLAGS, "mode" },
92 { "ds", "set display scale", OFFSET(display_scale), AV_OPT_TYPE_INT, {.i64=LINEAR}, LINEAR, NB_DISPLAY_SCALE - 1, FLAGS, "display_scale" },
93 { "lin", "linear", 0, AV_OPT_TYPE_CONST, {.i64=LINEAR}, 0, 0, FLAGS, "display_scale" },
94 { "log", "log", 0, AV_OPT_TYPE_CONST, {.i64=LOG}, 0, 0, FLAGS, "display_scale" },
98 AVFILTER_DEFINE_CLASS(showvolume);
100 static av_cold int init(AVFilterContext *ctx)
102 ShowVolumeContext *s = ctx->priv;
106 ret = av_expr_parse(&s->c_expr, s->color, var_names,
107 NULL, NULL, NULL, NULL, 0, ctx);
115 static int query_formats(AVFilterContext *ctx)
117 AVFilterFormats *formats = NULL;
118 AVFilterChannelLayouts *layouts = NULL;
119 AVFilterLink *inlink = ctx->inputs[0];
120 AVFilterLink *outlink = ctx->outputs[0];
121 static const enum AVSampleFormat sample_fmts[] = { AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_NONE };
122 static const enum AVPixelFormat pix_fmts[] = { AV_PIX_FMT_RGBA, AV_PIX_FMT_NONE };
125 formats = ff_make_format_list(sample_fmts);
126 if ((ret = ff_formats_ref(formats, &inlink->out_formats)) < 0)
129 layouts = ff_all_channel_counts();
130 if ((ret = ff_channel_layouts_ref(layouts, &inlink->out_channel_layouts)) < 0)
133 formats = ff_all_samplerates();
134 if ((ret = ff_formats_ref(formats, &inlink->out_samplerates)) < 0)
137 formats = ff_make_format_list(pix_fmts);
138 if ((ret = ff_formats_ref(formats, &outlink->in_formats)) < 0)
144 static void find_peak(float *src, int nb_samples, float *peak, float factor)
149 for (i = 0; i < nb_samples; i++)
150 *peak = FFMAX(*peak, FFABS(src[i]));
153 static void find_rms(float *src, int nb_samples, float *rms, float factor)
157 for (i = 0; i < nb_samples; i++)
158 *rms += factor * (src[i] * src[i] - *rms);
161 static int config_input(AVFilterLink *inlink)
163 AVFilterContext *ctx = inlink->dst;
164 ShowVolumeContext *s = ctx->priv;
167 nb_samples = FFMAX(1024, ((double)inlink->sample_rate / av_q2d(s->frame_rate)) + 0.5);
168 inlink->partial_buf_size =
169 inlink->min_samples =
170 inlink->max_samples = nb_samples;
171 s->values = av_calloc(inlink->channels * VAR_VARS_NB, sizeof(double));
173 return AVERROR(ENOMEM);
175 s->color_lut = av_calloc(s->w, sizeof(*s->color_lut) * inlink->channels);
177 return AVERROR(ENOMEM);
179 s->max = av_calloc(inlink->channels, sizeof(*s->max));
181 return AVERROR(ENOMEM);
183 s->rms_factor = 10000. / inlink->sample_rate;
186 case 0: s->meter = find_peak; break;
187 case 1: s->meter = find_rms; break;
188 default: return AVERROR_BUG;
191 if (s->draw_persistent_duration > 0.) {
192 s->persistent_max_frames = (int) FFMAX(av_q2d(s->frame_rate) * s->draw_persistent_duration, 1.);
193 s->max_persistent = av_calloc(inlink->channels * s->persistent_max_frames, sizeof(*s->max_persistent));
194 s->nb_frames_max_display = av_calloc(inlink->channels * s->persistent_max_frames, sizeof(*s->nb_frames_max_display));
199 static int config_output(AVFilterLink *outlink)
201 ShowVolumeContext *s = outlink->src->priv;
202 AVFilterLink *inlink = outlink->src->inputs[0];
205 if (s->orientation) {
207 outlink->w = s->h * inlink->channels + (inlink->channels - 1) * s->b;
210 outlink->h = s->h * inlink->channels + (inlink->channels - 1) * s->b;
213 outlink->sample_aspect_ratio = (AVRational){1,1};
214 outlink->frame_rate = s->frame_rate;
216 for (ch = 0; ch < inlink->channels; ch++) {
219 for (i = 0; i < s->w; i++) {
220 float max = i / (float)(s->w - 1);
222 s->values[ch * VAR_VARS_NB + VAR_PEAK] = max;
223 s->values[ch * VAR_VARS_NB + VAR_VOLUME] = 20.0 * log10(max);
224 s->values[ch * VAR_VARS_NB + VAR_CHANNEL] = ch;
225 s->color_lut[ch * s->w + i] = av_expr_eval(s->c_expr, &s->values[ch * VAR_VARS_NB], NULL);
232 static void drawtext(AVFrame *pic, int x, int y, const char *txt, int o)
238 font = avpriv_cga_font, font_height = 8;
240 for (i = 0; txt[i]; i++) {
243 if (o) { /* vertical orientation */
244 for (char_y = font_height - 1; char_y >= 0; char_y--) {
245 uint8_t *p = pic->data[0] + (y + i * 10) * pic->linesize[0] + x * 4;
246 for (mask = 0x80; mask; mask >>= 1) {
247 if (font[txt[i] * font_height + font_height - 1 - char_y] & mask)
248 AV_WN32(&p[char_y * 4], ~AV_RN32(&p[char_y * 4]));
249 p += pic->linesize[0];
252 } else { /* horizontal orientation */
253 uint8_t *p = pic->data[0] + y * pic->linesize[0] + (x + i * 8) * 4;
254 for (char_y = 0; char_y < font_height; char_y++) {
255 for (mask = 0x80; mask; mask >>= 1) {
256 if (font[txt[i] * font_height + char_y] & mask)
257 AV_WN32(p, ~AV_RN32(p));
260 p += pic->linesize[0] - 8 * 4;
266 static void clear_picture(ShowVolumeContext *s, AVFilterLink *outlink)
269 const uint32_t bg = (uint32_t)(s->bgopacity * 255) << 24;
271 for (i = 0; i < outlink->h; i++) {
272 uint32_t *dst = (uint32_t *)(s->out->data[0] + i * s->out->linesize[0]);
273 for (j = 0; j < outlink->w; j++)
274 AV_WN32A(dst + j, bg);
278 static inline int calc_max_draw(ShowVolumeContext *s, AVFilterLink *outlink, float max)
281 if (s->display_scale == LINEAR) {
284 max_val = av_clipf(0.21 * log10(max) + 1, 0, 1);
286 if (s->orientation) { /* vertical */
287 return outlink->h - outlink->h * max_val;
288 } else { /* horizontal */
289 return s->w * max_val;
293 static inline void calc_persistent_max(ShowVolumeContext *s, float max, int channel)
295 /* update max value for persistent max display */
296 if ((max >= s->max_persistent[channel]) || (s->nb_frames_max_display[channel] >= s->persistent_max_frames)) { /* update max value for display */
297 s->max_persistent[channel] = max;
298 s->nb_frames_max_display[channel] = 0;
300 s->nb_frames_max_display[channel] += 1; /* incremente display frame count */
304 static inline void draw_max_line(ShowVolumeContext *s, int max_draw, int channel)
307 if (s->orientation) { /* vertical */
308 uint8_t *dst = s->out->data[0] + max_draw * s->out->linesize[0] + channel * (s->b + s->h) * 4;
309 for (k = 0; k < s->h; k++) {
310 memcpy(dst + k * 4, s->persistant_max_rgba, sizeof(s->persistant_max_rgba));
312 } else { /* horizontal */
313 for (k = 0; k < s->h; k++) {
314 uint8_t *dst = s->out->data[0] + (channel * s->h + channel * s->b + k) * s->out->linesize[0];
315 memcpy(dst + max_draw * 4, s->persistant_max_rgba, sizeof(s->persistant_max_rgba));
320 static int filter_frame(AVFilterLink *inlink, AVFrame *insamples)
322 AVFilterContext *ctx = inlink->dst;
323 AVFilterLink *outlink = ctx->outputs[0];
324 ShowVolumeContext *s = ctx->priv;
325 const int step = s->step;
326 int c, j, k, max_draw;
329 if (!s->out || s->out->width != outlink->w ||
330 s->out->height != outlink->h) {
331 av_frame_free(&s->out);
332 s->out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
334 av_frame_free(&insamples);
335 return AVERROR(ENOMEM);
337 clear_picture(s, outlink);
339 s->out->pts = insamples->pts;
341 if ((s->f < 1.) && (s->f > 0.)) {
342 for (j = 0; j < outlink->h; j++) {
343 uint8_t *dst = s->out->data[0] + j * s->out->linesize[0];
344 const uint32_t alpha = s->bgopacity * 255;
346 for (k = 0; k < outlink->w; k++) {
347 dst[k * 4 + 0] = FFMAX(dst[k * 4 + 0] * s->f, 0);
348 dst[k * 4 + 1] = FFMAX(dst[k * 4 + 1] * s->f, 0);
349 dst[k * 4 + 2] = FFMAX(dst[k * 4 + 2] * s->f, 0);
350 dst[k * 4 + 3] = FFMAX(dst[k * 4 + 3] * s->f, alpha);
353 } else if (s->f == 0.) {
354 clear_picture(s, outlink);
357 if (s->orientation) { /* vertical */
358 for (c = 0; c < inlink->channels; c++) {
359 float *src = (float *)insamples->extended_data[c];
360 uint32_t *lut = s->color_lut + s->w * c;
363 s->meter(src, insamples->nb_samples, &s->max[c], s->rms_factor);
366 s->values[c * VAR_VARS_NB + VAR_VOLUME] = 20.0 * log10(max);
367 max = av_clipf(max, 0, 1);
368 max_draw = calc_max_draw(s, outlink, max);
370 for (j = max_draw; j < s->w; j++) {
371 uint8_t *dst = s->out->data[0] + j * s->out->linesize[0] + c * (s->b + s->h) * 4;
372 for (k = 0; k < s->h; k++) {
373 AV_WN32A(&dst[k * 4], lut[s->w - j - 1]);
379 if (s->h >= 8 && s->draw_text) {
380 const char *channel_name = av_get_channel_name(av_channel_layout_extract_channel(insamples->channel_layout, c));
383 drawtext(s->out, c * (s->h + s->b) + (s->h - 10) / 2, outlink->h - 35, channel_name, 1);
386 if (s->draw_persistent_duration > 0.) {
387 calc_persistent_max(s, max, c);
388 max_draw = FFMAX(0, calc_max_draw(s, outlink, s->max_persistent[c]) - 1);
389 draw_max_line(s, max_draw, c);
392 } else { /* horizontal */
393 for (c = 0; c < inlink->channels; c++) {
394 float *src = (float *)insamples->extended_data[c];
395 uint32_t *lut = s->color_lut + s->w * c;
398 s->meter(src, insamples->nb_samples, &s->max[c], s->rms_factor);
401 s->values[c * VAR_VARS_NB + VAR_VOLUME] = 20.0 * log10(max);
402 max = av_clipf(max, 0, 1);
403 max_draw = calc_max_draw(s, outlink, max);
405 for (j = 0; j < s->h; j++) {
406 uint8_t *dst = s->out->data[0] + (c * s->h + c * s->b + j) * s->out->linesize[0];
408 for (k = 0; k < max_draw; k++) {
409 AV_WN32A(dst + k * 4, lut[k]);
415 if (s->h >= 8 && s->draw_text) {
416 const char *channel_name = av_get_channel_name(av_channel_layout_extract_channel(insamples->channel_layout, c));
419 drawtext(s->out, 2, c * (s->h + s->b) + (s->h - 8) / 2, channel_name, 0);
422 if (s->draw_persistent_duration > 0.) {
423 calc_persistent_max(s, max, c);
424 max_draw = FFMAX(0, calc_max_draw(s, outlink, s->max_persistent[c]) - 1);
425 draw_max_line(s, max_draw, c);
430 av_frame_free(&insamples);
431 out = av_frame_clone(s->out);
433 return AVERROR(ENOMEM);
434 av_frame_make_writable(out);
436 /* draw volume level */
437 for (c = 0; c < inlink->channels && s->h >= 8 && s->draw_volume; c++) {
440 if (s->orientation) { /* vertical */
441 snprintf(buf, sizeof(buf), "%.2f", s->values[c * VAR_VARS_NB + VAR_VOLUME]);
442 drawtext(out, c * (s->h + s->b) + (s->h - 8) / 2, 2, buf, 1);
443 } else { /* horizontal */
444 snprintf(buf, sizeof(buf), "%.2f", s->values[c * VAR_VARS_NB + VAR_VOLUME]);
445 drawtext(out, FFMAX(0, s->w - 8 * (int)strlen(buf)), c * (s->h + s->b) + (s->h - 8) / 2, buf, 0);
449 return ff_filter_frame(outlink, out);
452 static av_cold void uninit(AVFilterContext *ctx)
454 ShowVolumeContext *s = ctx->priv;
456 av_frame_free(&s->out);
457 av_expr_free(s->c_expr);
458 av_freep(&s->values);
459 av_freep(&s->color_lut);
463 static const AVFilterPad showvolume_inputs[] = {
466 .type = AVMEDIA_TYPE_AUDIO,
467 .config_props = config_input,
468 .filter_frame = filter_frame,
473 static const AVFilterPad showvolume_outputs[] = {
476 .type = AVMEDIA_TYPE_VIDEO,
477 .config_props = config_output,
482 AVFilter ff_avf_showvolume = {
483 .name = "showvolume",
484 .description = NULL_IF_CONFIG_SMALL("Convert input audio volume to video output."),
487 .query_formats = query_formats,
488 .priv_size = sizeof(ShowVolumeContext),
489 .inputs = showvolume_inputs,
490 .outputs = showvolume_outputs,
491 .priv_class = &showvolume_class,