FFmpeg coverage

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/avassert.h"
22			#include "libavutil/mem.h"
23			#include "libavutil/opt.h"
24			#include "avfilter.h"
25			#include "filters.h"
26			#include "formats.h"
27			#include "video.h"
28
29			enum DisplayScale { LINEAR, SQRT, CBRT, LOG, RLOG, NB_SCALES };
30			enum AmplitudeScale { ALINEAR, ALOG, NB_ASCALES };
31			enum SlideMode { REPLACE, SCROLL, NB_SLIDES };
32			enum DisplayMode { SINGLE, SEPARATE, NB_DMODES };
33			enum HistogramMode { ABS, SIGN, NB_HMODES };
34
35			typedef struct AudioHistogramContext {
36			const AVClass *class;
37			AVFrame *out;
38			int w, h;
39			AVRational frame_rate;
40			uint64_t *achistogram;
41			uint64_t *shistogram;
42			int ascale;
43			int scale;
44			float phisto;
45			int histogram_h;
46			int apos;
47			int ypos;
48			int slide;
49			int dmode;
50			int hmode;
51			int dchannels;
52			int count;
53			int frame_count;
54			float *combine_buffer;
55			AVFrame *in[101];
56			int first;
57			int nb_samples;
58
59			int (*get_bin)(float in, int w);
60			} AudioHistogramContext;
61
62			#define OFFSET(x) offsetof(AudioHistogramContext, x)
63			#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
64
65			static const AVOption ahistogram_options[] = {
66			{ "dmode", "set method to display channels", OFFSET(dmode), AV_OPT_TYPE_INT, {.i64=SINGLE}, 0, NB_DMODES-1, FLAGS, .unit = "dmode" },
67			{ "single", "all channels use single histogram", 0, AV_OPT_TYPE_CONST, {.i64=SINGLE}, 0, 0, FLAGS, .unit = "dmode" },
68			{ "separate", "each channel have own histogram", 0, AV_OPT_TYPE_CONST, {.i64=SEPARATE}, 0, 0, FLAGS, .unit = "dmode" },
69			{ "rate", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
70			{ "r", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
71			{ "size", "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="hd720"}, 0, 0, FLAGS },
72			{ "s", "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="hd720"}, 0, 0, FLAGS },
73			{ "scale", "set display scale", OFFSET(scale), AV_OPT_TYPE_INT, {.i64=LOG}, LINEAR, NB_SCALES-1, FLAGS, .unit = "scale" },
74			{ "log", "logarithmic", 0, AV_OPT_TYPE_CONST, {.i64=LOG}, 0, 0, FLAGS, .unit = "scale" },
75			{ "sqrt", "square root", 0, AV_OPT_TYPE_CONST, {.i64=SQRT}, 0, 0, FLAGS, .unit = "scale" },
76			{ "cbrt", "cubic root", 0, AV_OPT_TYPE_CONST, {.i64=CBRT}, 0, 0, FLAGS, .unit = "scale" },
77			{ "lin", "linear", 0, AV_OPT_TYPE_CONST, {.i64=LINEAR}, 0, 0, FLAGS, .unit = "scale" },
78			{ "rlog", "reverse logarithmic", 0, AV_OPT_TYPE_CONST, {.i64=RLOG}, 0, 0, FLAGS, .unit = "scale" },
79			{ "ascale", "set amplitude scale", OFFSET(ascale), AV_OPT_TYPE_INT, {.i64=ALOG}, LINEAR, NB_ASCALES-1, FLAGS, .unit = "ascale" },
80			{ "log", "logarithmic", 0, AV_OPT_TYPE_CONST, {.i64=ALOG}, 0, 0, FLAGS, .unit = "ascale" },
81			{ "lin", "linear", 0, AV_OPT_TYPE_CONST, {.i64=ALINEAR}, 0, 0, FLAGS, .unit = "ascale" },
82			{ "acount", "how much frames to accumulate", OFFSET(count), AV_OPT_TYPE_INT, {.i64=1}, -1, 100, FLAGS },
83			{ "rheight", "set histogram ratio of window height", OFFSET(phisto), AV_OPT_TYPE_FLOAT, {.dbl=0.10}, 0, 1, FLAGS },
84			{ "slide", "set sonogram sliding", OFFSET(slide), AV_OPT_TYPE_INT, {.i64=REPLACE}, 0, NB_SLIDES-1, FLAGS, .unit = "slide" },
85			{ "replace", "replace old rows with new", 0, AV_OPT_TYPE_CONST, {.i64=REPLACE}, 0, 0, FLAGS, .unit = "slide" },
86			{ "scroll", "scroll from top to bottom", 0, AV_OPT_TYPE_CONST, {.i64=SCROLL}, 0, 0, FLAGS, .unit = "slide" },
87			{ "hmode", "set histograms mode", OFFSET(hmode), AV_OPT_TYPE_INT, {.i64=ABS}, 0, NB_HMODES-1, FLAGS, .unit = "hmode" },
88			{ "abs", "use absolute samples", 0, AV_OPT_TYPE_CONST, {.i64=ABS}, 0, 0, FLAGS, .unit = "hmode" },
89			{ "sign", "use unchanged samples", 0, AV_OPT_TYPE_CONST, {.i64=SIGN},0, 0, FLAGS, .unit = "hmode" },
90			{ NULL }
91			};
92
93			AVFILTER_DEFINE_CLASS(ahistogram);
94
95		✗	static int query_formats(const AVFilterContext *ctx,
96			AVFilterFormatsConfig **cfg_in,
97			AVFilterFormatsConfig **cfg_out)
98			{
99		✗	AVFilterFormats *formats = NULL;
100			static const enum AVSampleFormat sample_fmts[] = { AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_NONE };
101			static const enum AVPixelFormat pix_fmts[] = { AV_PIX_FMT_YUVA444P, AV_PIX_FMT_NONE };
102		✗	int ret = AVERROR(EINVAL);
103
104		✗	formats = ff_make_format_list(sample_fmts);
105		✗	if ((ret = ff_formats_ref(formats, &cfg_in[0]->formats)) < 0)
106		✗	return ret;
107
108		✗	formats = ff_make_format_list(pix_fmts);
109		✗	if ((ret = ff_formats_ref(formats, &cfg_out[0]->formats)) < 0)
110		✗	return ret;
111
112		✗	return 0;
113			}
114
115		✗	static int config_input(AVFilterLink *inlink)
116			{
117		✗	AVFilterContext *ctx = inlink->dst;
118		✗	AudioHistogramContext *s = ctx->priv;
119
120		✗	s->nb_samples = FFMAX(1, av_rescale(inlink->sample_rate, s->frame_rate.den, s->frame_rate.num));
121		✗	s->dchannels = s->dmode == SINGLE ? 1 : inlink->ch_layout.nb_channels;
122		✗	s->shistogram = av_calloc(s->w, s->dchannels * sizeof(*s->shistogram));
123		✗	if (!s->shistogram)
124		✗	return AVERROR(ENOMEM);
125
126		✗	s->achistogram = av_calloc(s->w, s->dchannels * sizeof(*s->achistogram));
127		✗	if (!s->achistogram)
128		✗	return AVERROR(ENOMEM);
129
130		✗	return 0;
131			}
132
133		✗	static int get_lin_bin_abs(float in, int w)
134			{
135		✗	return lrintf(av_clipf(fabsf(in), 0.f, 1.f) * (w - 1));
136			}
137
138		✗	static int get_lin_bin_sign(float in, int w)
139			{
140		✗	return lrintf((1.f + av_clipf(in, -1.f, 1.f)) * 0.5f * (w - 1));
141			}
142
143		✗	static int get_log_bin_abs(float in, int w)
144			{
145		✗	return lrintf(av_clipf(1.f + log10f(fabsf(in)) / 6.f, 0.f, 1.f) * (w - 1));
146			}
147
148		✗	static int get_log_bin_sign(float in, int w)
149			{
150		✗	return (w / 2) + FFSIGN(in) * lrintf(av_clipf(1.f + log10f(fabsf(in)) / 6.f, 0.f, 1.f) * (w / 2));
151			}
152
153		✗	static int config_output(AVFilterLink *outlink)
154			{
155		✗	AudioHistogramContext *s = outlink->src->priv;
156		✗	FilterLink *l = ff_filter_link(outlink);
157
158		✗	outlink->w = s->w;
159		✗	outlink->h = s->h;
160		✗	outlink->sample_aspect_ratio = (AVRational){1,1};
161		✗	l->frame_rate = s->frame_rate;
162		✗	outlink->time_base = av_inv_q(l->frame_rate);
163
164		✗	s->histogram_h = s->h * s->phisto;
165		✗	s->ypos = s->h * s->phisto;
166
167		✗	switch (s->ascale) {
168		✗	case ALINEAR:
169		✗	switch (s->hmode) {
170		✗	case ABS: s->get_bin = get_lin_bin_abs; break;
171		✗	case SIGN: s->get_bin = get_lin_bin_sign; break;
172		✗	default:
173		✗	return AVERROR_BUG;
174			}
175		✗	break;
176		✗	case ALOG:
177		✗	switch (s->hmode) {
178		✗	case ABS: s->get_bin = get_log_bin_abs; break;
179		✗	case SIGN: s->get_bin = get_log_bin_sign; break;
180		✗	default:
181		✗	return AVERROR_BUG;
182			}
183		✗	break;
184		✗	default:
185		✗	return AVERROR_BUG;
186			}
187
188		✗	if (s->dmode == SEPARATE) {
189		✗	s->combine_buffer = av_malloc_array(outlink->w * 3, sizeof(*s->combine_buffer));
190		✗	if (!s->combine_buffer)
191		✗	return AVERROR(ENOMEM);
192			}
193
194		✗	return 0;
195			}
196
197		✗	static int filter_frame(AVFilterLink *inlink, AVFrame *in)
198			{
199		✗	AVFilterContext *ctx = inlink->dst;
200		✗	AVFilterLink *outlink = ctx->outputs[0];
201		✗	AudioHistogramContext *s = ctx->priv;
202		✗	const int nb_samples = in->nb_samples;
203		✗	const int H = s->histogram_h;
204		✗	const int w = s->w;
205			int c, y, n, p, bin, ret;
206		✗	uint64_t acmax = 1;
207			AVFrame *clone;
208
209		✗	if (!s->out || s->out->width != outlink->w ||
210		✗	s->out->height != outlink->h) {
211		✗	av_frame_free(&s->out);
212		✗	s->out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
213		✗	if (!s->out) {
214		✗	av_frame_free(&in);
215		✗	return AVERROR(ENOMEM);
216			}
217		✗	for (n = H; n < s->h; n++) {
218		✗	memset(s->out->data[0] + n * s->out->linesize[0], 0, w);
219		✗	memset(s->out->data[1] + n * s->out->linesize[0], 127, w);
220		✗	memset(s->out->data[2] + n * s->out->linesize[0], 127, w);
221		✗	memset(s->out->data[3] + n * s->out->linesize[0], 0, w);
222			}
223			}
224
225		✗	ret = ff_inlink_make_frame_writable(outlink, &s->out);
226		✗	if (ret < 0) {
227		✗	av_frame_free(&in);
228		✗	return ret;
229			}
230
231		✗	if (s->dmode == SEPARATE) {
232		✗	for (y = 0; y < w; y++) {
233		✗	s->combine_buffer[3 * y ] = 0;
234		✗	s->combine_buffer[3 * y + 1] = 127.5;
235		✗	s->combine_buffer[3 * y + 2] = 127.5;
236			}
237			}
238
239		✗	for (n = 0; n < H; n++) {
240		✗	memset(s->out->data[0] + n * s->out->linesize[0], 0, w);
241		✗	memset(s->out->data[1] + n * s->out->linesize[0], 127, w);
242		✗	memset(s->out->data[2] + n * s->out->linesize[0], 127, w);
243		✗	memset(s->out->data[3] + n * s->out->linesize[0], 0, w);
244			}
245		✗	s->out->pts = av_rescale_q(in->pts, inlink->time_base, outlink->time_base);
246		✗	s->out->duration = 1;
247
248		✗	s->first = s->frame_count;
249
250		✗	switch (s->ascale) {
251		✗	case ALINEAR:
252		✗	for (c = 0; c < inlink->ch_layout.nb_channels; c++) {
253		✗	const float *src = (const float *)in->extended_data[c];
254		✗	uint64_t *achistogram = &s->achistogram[(s->dmode == SINGLE ? 0: c) * w];
255
256		✗	for (n = 0; n < nb_samples; n++) {
257		✗	bin = s->get_bin(src[n], w);
258
259		✗	achistogram[bin]++;
260			}
261
262		✗	if (s->in[s->first] && s->count >= 0) {
263		✗	uint64_t *shistogram = &s->shistogram[(s->dmode == SINGLE ? 0: c) * w];
264		✗	const float *src2 = (const float *)s->in[s->first]->extended_data[c];
265
266		✗	for (n = 0; n < nb_samples; n++) {
267		✗	bin = s->get_bin(src2[n], w);
268
269		✗	shistogram[bin]++;
270			}
271			}
272			}
273		✗	break;
274		✗	case ALOG:
275		✗	for (c = 0; c < inlink->ch_layout.nb_channels; c++) {
276		✗	const float *src = (const float *)in->extended_data[c];
277		✗	uint64_t *achistogram = &s->achistogram[(s->dmode == SINGLE ? 0: c) * w];
278
279		✗	for (n = 0; n < nb_samples; n++) {
280		✗	bin = s->get_bin(src[n], w);
281
282		✗	achistogram[bin]++;
283			}
284
285		✗	if (s->in[s->first] && s->count >= 0) {
286		✗	uint64_t *shistogram = &s->shistogram[(s->dmode == SINGLE ? 0: c) * w];
287		✗	const float *src2 = (const float *)s->in[s->first]->extended_data[c];
288
289		✗	for (n = 0; n < nb_samples; n++) {
290		✗	bin = s->get_bin(src2[n], w);
291
292		✗	shistogram[bin]++;
293			}
294			}
295			}
296		✗	break;
297			}
298
299		✗	av_frame_free(&s->in[s->frame_count]);
300		✗	s->in[s->frame_count] = in;
301		✗	s->frame_count++;
302		✗	if (s->frame_count > s->count)
303		✗	s->frame_count = 0;
304
305		✗	for (n = 0; n < w * s->dchannels; n++) {
306		✗	acmax = FFMAX(s->achistogram[n] - s->shistogram[n], acmax);
307			}
308
309		✗	for (c = 0; c < s->dchannels; c++) {
310		✗	uint64_t *shistogram = &s->shistogram[c * w];
311		✗	uint64_t *achistogram = &s->achistogram[c * w];
312			float yf, uf, vf;
313
314		✗	if (s->dmode == SEPARATE) {
315		✗	yf = 255.0f / s->dchannels;
316		✗	uf = yf * M_PI;
317		✗	vf = yf * M_PI;
318		✗	uf *= 0.5 * sin((2 * M_PI * c) / s->dchannels);
319		✗	vf *= 0.5 * cos((2 * M_PI * c) / s->dchannels);
320			}
321
322		✗	for (n = 0; n < w; n++) {
323			double a, aa;
324			int h;
325
326		✗	a = achistogram[n] - shistogram[n];
327
328		✗	switch (s->scale) {
329		✗	case LINEAR:
330		✗	aa = a / (double)acmax;
331		✗	break;
332		✗	case SQRT:
333		✗	aa = sqrt(a) / sqrt(acmax);
334		✗	break;
335		✗	case CBRT:
336		✗	aa = cbrt(a) / cbrt(acmax);
337		✗	break;
338		✗	case LOG:
339		✗	aa = log2(a + 1) / log2(acmax + 1);
340		✗	break;
341		✗	case RLOG:
342		✗	aa = 1. - log2(a + 1) / log2(acmax + 1);
343		✗	if (aa == 1.)
344		✗	aa = 0;
345		✗	break;
346		✗	default:
347		✗	av_assert0(0);
348			}
349
350		✗	h = aa * (H - 1);
351
352		✗	if (s->dmode == SINGLE) {
353		✗	int start = H - h, end = H;
354		✗	const int linesizey = s->out->linesize[0];
355		✗	const int linesizea = s->out->linesize[3];
356		✗	uint8_t *dsty = s->out->data[0] + start * linesizey;
357		✗	uint8_t *dsta = s->out->data[3] + start * linesizea;
358
359		✗	for (y = start; y < end; y++, dsty += linesizey, dsta += linesizea) {
360		✗	dsty[n] = 255;
361		✗	dsta[n] = 255;
362			}
363
364		✗	if (s->h - H > 0) {
365		✗	h = aa * 255;
366
367		✗	s->out->data[0][s->ypos * s->out->linesize[0] + n] = av_clip_uint8(h);
368		✗	s->out->data[1][s->ypos * s->out->linesize[1] + n] = 127;
369		✗	s->out->data[2][s->ypos * s->out->linesize[2] + n] = 127;
370		✗	s->out->data[3][s->ypos * s->out->linesize[3] + n] = 255;
371			}
372		✗	} else if (s->dmode == SEPARATE) {
373		✗	int start = H - h, end = H;
374		✗	float *out = &s->combine_buffer[3 * n];
375		✗	const int linesizey = s->out->linesize[0];
376		✗	const int linesizeu = s->out->linesize[1];
377		✗	const int linesizev = s->out->linesize[2];
378		✗	const int linesizea = s->out->linesize[3];
379		✗	uint8_t *dsty = s->out->data[0] + start * linesizey;
380		✗	uint8_t *dstu = s->out->data[1] + start * linesizeu;
381		✗	uint8_t *dstv = s->out->data[2] + start * linesizev;
382		✗	uint8_t *dsta = s->out->data[3] + start * linesizea;
383			int old;
384
385		✗	old = dsty[n];
386		✗	for (y = start; y < end; y++) {
387		✗	if (dsty[n] != old)
388		✗	break;
389		✗	old = dsty[n];
390		✗	dsty[n] = av_clip_uint8(yf);
391		✗	dstu[n] = av_clip_uint8(128.f+uf);
392		✗	dstv[n] = av_clip_uint8(128.f+vf);
393		✗	dsta[n] = 255;
394
395		✗	dsty += linesizey;
396		✗	dstu += linesizeu;
397		✗	dstv += linesizev;
398		✗	dsta += linesizea;
399			}
400
401		✗	out[0] += aa * yf;
402		✗	out[1] += aa * uf;
403		✗	out[2] += aa * vf;
404			}
405			}
406			}
407
408		✗	if (s->h - H > 0) {
409		✗	if (s->dmode == SEPARATE) {
410		✗	for (n = 0; n < w; n++) {
411		✗	float *cb = &s->combine_buffer[3 * n];
412
413		✗	s->out->data[0][s->ypos * s->out->linesize[0] + n] = cb[0];
414		✗	s->out->data[1][s->ypos * s->out->linesize[1] + n] = cb[1];
415		✗	s->out->data[2][s->ypos * s->out->linesize[2] + n] = cb[2];
416		✗	s->out->data[3][s->ypos * s->out->linesize[3] + n] = 255;
417			}
418			}
419
420		✗	if (s->slide == SCROLL) {
421		✗	for (p = 0; p < 4; p++) {
422		✗	for (y = s->h - 1; y >= H + 1; y--) {
423		✗	memmove(s->out->data[p] + (y ) * s->out->linesize[p],
424		✗	s->out->data[p] + (y-1) * s->out->linesize[p], w);
425			}
426			}
427			}
428
429		✗	s->ypos++;
430		✗	if (s->slide == SCROLL || s->ypos >= s->h)
431		✗	s->ypos = H;
432			}
433
434		✗	clone = av_frame_clone(s->out);
435		✗	if (!clone)
436		✗	return AVERROR(ENOMEM);
437
438		✗	return ff_filter_frame(outlink, clone);
439			}
440
441		✗	static int activate(AVFilterContext *ctx)
442			{
443		✗	AVFilterLink *inlink = ctx->inputs[0];
444		✗	AVFilterLink *outlink = ctx->outputs[0];
445		✗	AudioHistogramContext *s = ctx->priv;
446			AVFrame *in;
447			int ret;
448
449		✗	FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
450
451		✗	ret = ff_inlink_consume_samples(inlink, s->nb_samples, s->nb_samples, &in);
452		✗	if (ret < 0)
453		✗	return ret;
454		✗	if (ret > 0)
455		✗	return filter_frame(inlink, in);
456
457		✗	if (ff_inlink_queued_samples(inlink) >= s->nb_samples) {
458		✗	ff_filter_set_ready(ctx, 10);
459		✗	return 0;
460			}
461
462		✗	FF_FILTER_FORWARD_STATUS(inlink, outlink);
463		✗	FF_FILTER_FORWARD_WANTED(outlink, inlink);
464
465		✗	return FFERROR_NOT_READY;
466			}
467
468		✗	static av_cold void uninit(AVFilterContext *ctx)
469			{
470		✗	AudioHistogramContext *s = ctx->priv;
471			int i;
472
473		✗	av_frame_free(&s->out);
474		✗	av_freep(&s->shistogram);
475		✗	av_freep(&s->achistogram);
476		✗	av_freep(&s->combine_buffer);
477		✗	for (i = 0; i < 101; i++)
478		✗	av_frame_free(&s->in[i]);
479		✗	}
480
481			static const AVFilterPad ahistogram_inputs[] = {
482			{
483			.name = "default",
484			.type = AVMEDIA_TYPE_AUDIO,
485			.config_props = config_input,
486			},
487			};
488
489			static const AVFilterPad ahistogram_outputs[] = {
490			{
491			.name = "default",
492			.type = AVMEDIA_TYPE_VIDEO,
493			.config_props = config_output,
494			},
495			};
496
497			const AVFilter ff_avf_ahistogram = {
498			.name = "ahistogram",
499			.description = NULL_IF_CONFIG_SMALL("Convert input audio to histogram video output."),
500			.uninit = uninit,
501			.priv_size = sizeof(AudioHistogramContext),
502			.activate = activate,
503			FILTER_INPUTS(ahistogram_inputs),
504			FILTER_OUTPUTS(ahistogram_outputs),
505			FILTER_QUERY_FUNC2(query_formats),
506			.priv_class = &ahistogram_class,
507			};
508