FFmpeg/libavfilter/avf_ahistogram.c
Nicolas George 2f76476549 lavfi: regroup formats lists in a single structure.
It will allow to refernce it as a whole without clunky macros.

Most of the changes have been automatically made with sed:

sed -i '
  s/-> *in_formats/->incfg.formats/g;
  s/-> *out_formats/->outcfg.formats/g;
  s/-> *in_channel_layouts/->incfg.channel_layouts/g;
  s/-> *out_channel_layouts/->outcfg.channel_layouts/g;
  s/-> *in_samplerates/->incfg.samplerates/g;
  s/-> *out_samplerates/->outcfg.samplerates/g;
  ' src/libavfilter/*(.)
2020-09-08 14:02:40 +02:00

438 lines
15 KiB
C

/*
* Copyright (c) 2015 Paul B Mahol
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/avassert.h"
#include "libavutil/opt.h"
#include "libavutil/parseutils.h"
#include "avfilter.h"
#include "filters.h"
#include "formats.h"
#include "audio.h"
#include "video.h"
#include "internal.h"
enum DisplayScale { LINEAR, SQRT, CBRT, LOG, RLOG, NB_SCALES };
enum AmplitudeScale { ALINEAR, ALOG, NB_ASCALES };
enum SlideMode { REPLACE, SCROLL, NB_SLIDES };
enum DisplayMode { SINGLE, SEPARATE, NB_DMODES };
enum HistogramMode { ACCUMULATE, CURRENT, NB_HMODES };
typedef struct AudioHistogramContext {
const AVClass *class;
AVFrame *out;
int w, h;
AVRational frame_rate;
uint64_t *achistogram;
uint64_t *shistogram;
int ascale;
int scale;
float phisto;
int histogram_h;
int apos;
int ypos;
int slide;
int dmode;
int dchannels;
int count;
int frame_count;
float *combine_buffer;
AVFrame *in[101];
int first;
int nb_samples;
} AudioHistogramContext;
#define OFFSET(x) offsetof(AudioHistogramContext, x)
#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
static const AVOption ahistogram_options[] = {
{ "dmode", "set method to display channels", OFFSET(dmode), AV_OPT_TYPE_INT, {.i64=SINGLE}, 0, NB_DMODES-1, FLAGS, "dmode" },
{ "single", "all channels use single histogram", 0, AV_OPT_TYPE_CONST, {.i64=SINGLE}, 0, 0, FLAGS, "dmode" },
{ "separate", "each channel have own histogram", 0, AV_OPT_TYPE_CONST, {.i64=SEPARATE}, 0, 0, FLAGS, "dmode" },
{ "rate", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
{ "r", "set video rate", OFFSET(frame_rate), AV_OPT_TYPE_VIDEO_RATE, {.str="25"}, 0, INT_MAX, FLAGS },
{ "size", "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="hd720"}, 0, 0, FLAGS },
{ "s", "set video size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, {.str="hd720"}, 0, 0, FLAGS },
{ "scale", "set display scale", OFFSET(scale), AV_OPT_TYPE_INT, {.i64=LOG}, LINEAR, NB_SCALES-1, FLAGS, "scale" },
{ "log", "logarithmic", 0, AV_OPT_TYPE_CONST, {.i64=LOG}, 0, 0, FLAGS, "scale" },
{ "sqrt", "square root", 0, AV_OPT_TYPE_CONST, {.i64=SQRT}, 0, 0, FLAGS, "scale" },
{ "cbrt", "cubic root", 0, AV_OPT_TYPE_CONST, {.i64=CBRT}, 0, 0, FLAGS, "scale" },
{ "lin", "linear", 0, AV_OPT_TYPE_CONST, {.i64=LINEAR}, 0, 0, FLAGS, "scale" },
{ "rlog", "reverse logarithmic", 0, AV_OPT_TYPE_CONST, {.i64=RLOG}, 0, 0, FLAGS, "scale" },
{ "ascale", "set amplitude scale", OFFSET(ascale), AV_OPT_TYPE_INT, {.i64=ALOG}, LINEAR, NB_ASCALES-1, FLAGS, "ascale" },
{ "log", "logarithmic", 0, AV_OPT_TYPE_CONST, {.i64=ALOG}, 0, 0, FLAGS, "ascale" },
{ "lin", "linear", 0, AV_OPT_TYPE_CONST, {.i64=ALINEAR}, 0, 0, FLAGS, "ascale" },
{ "acount", "how much frames to accumulate", OFFSET(count), AV_OPT_TYPE_INT, {.i64=1}, -1, 100, FLAGS },
{ "rheight", "set histogram ratio of window height", OFFSET(phisto), AV_OPT_TYPE_FLOAT, {.dbl=0.10}, 0, 1, FLAGS },
{ "slide", "set sonogram sliding", OFFSET(slide), AV_OPT_TYPE_INT, {.i64=REPLACE}, 0, NB_SLIDES-1, FLAGS, "slide" },
{ "replace", "replace old rows with new", 0, AV_OPT_TYPE_CONST, {.i64=REPLACE}, 0, 0, FLAGS, "slide" },
{ "scroll", "scroll from top to bottom", 0, AV_OPT_TYPE_CONST, {.i64=SCROLL}, 0, 0, FLAGS, "slide" },
{ NULL }
};
AVFILTER_DEFINE_CLASS(ahistogram);
static int query_formats(AVFilterContext *ctx)
{
AVFilterFormats *formats = NULL;
AVFilterChannelLayouts *layouts = NULL;
AVFilterLink *inlink = ctx->inputs[0];
AVFilterLink *outlink = ctx->outputs[0];
static const enum AVSampleFormat sample_fmts[] = { AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_NONE };
static const enum AVPixelFormat pix_fmts[] = { AV_PIX_FMT_YUVA444P, AV_PIX_FMT_NONE };
int ret = AVERROR(EINVAL);
formats = ff_make_format_list(sample_fmts);
if ((ret = ff_formats_ref (formats, &inlink->outcfg.formats )) < 0 ||
(layouts = ff_all_channel_counts()) == NULL ||
(ret = ff_channel_layouts_ref (layouts, &inlink->outcfg.channel_layouts)) < 0)
return ret;
formats = ff_all_samplerates();
if ((ret = ff_formats_ref(formats, &inlink->outcfg.samplerates)) < 0)
return ret;
formats = ff_make_format_list(pix_fmts);
if ((ret = ff_formats_ref(formats, &outlink->incfg.formats)) < 0)
return ret;
return 0;
}
static int config_input(AVFilterLink *inlink)
{
AVFilterContext *ctx = inlink->dst;
AudioHistogramContext *s = ctx->priv;
s->nb_samples = FFMAX(1, av_rescale(inlink->sample_rate, s->frame_rate.den, s->frame_rate.num));
s->dchannels = s->dmode == SINGLE ? 1 : inlink->channels;
s->shistogram = av_calloc(s->w, s->dchannels * sizeof(*s->shistogram));
if (!s->shistogram)
return AVERROR(ENOMEM);
s->achistogram = av_calloc(s->w, s->dchannels * sizeof(*s->achistogram));
if (!s->achistogram)
return AVERROR(ENOMEM);
return 0;
}
static int config_output(AVFilterLink *outlink)
{
AudioHistogramContext *s = outlink->src->priv;
outlink->w = s->w;
outlink->h = s->h;
outlink->sample_aspect_ratio = (AVRational){1,1};
outlink->frame_rate = s->frame_rate;
s->histogram_h = s->h * s->phisto;
s->ypos = s->h * s->phisto;
if (s->dmode == SEPARATE) {
s->combine_buffer = av_malloc_array(outlink->w * 3, sizeof(*s->combine_buffer));
if (!s->combine_buffer)
return AVERROR(ENOMEM);
}
return 0;
}
static int filter_frame(AVFilterLink *inlink, AVFrame *in)
{
AVFilterContext *ctx = inlink->dst;
AVFilterLink *outlink = ctx->outputs[0];
AudioHistogramContext *s = ctx->priv;
const int H = s->histogram_h;
const int w = s->w;
int c, y, n, p, bin;
uint64_t acmax = 1;
AVFrame *clone;
if (!s->out || s->out->width != outlink->w ||
s->out->height != outlink->h) {
av_frame_free(&s->out);
s->out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
if (!s->out) {
av_frame_free(&in);
return AVERROR(ENOMEM);
}
for (n = H; n < s->h; n++) {
memset(s->out->data[0] + n * s->out->linesize[0], 0, w);
memset(s->out->data[1] + n * s->out->linesize[0], 127, w);
memset(s->out->data[2] + n * s->out->linesize[0], 127, w);
memset(s->out->data[3] + n * s->out->linesize[0], 0, w);
}
}
if (s->dmode == SEPARATE) {
for (y = 0; y < w; y++) {
s->combine_buffer[3 * y ] = 0;
s->combine_buffer[3 * y + 1] = 127.5;
s->combine_buffer[3 * y + 2] = 127.5;
}
}
for (n = 0; n < H; n++) {
memset(s->out->data[0] + n * s->out->linesize[0], 0, w);
memset(s->out->data[1] + n * s->out->linesize[0], 127, w);
memset(s->out->data[2] + n * s->out->linesize[0], 127, w);
memset(s->out->data[3] + n * s->out->linesize[0], 0, w);
}
s->out->pts = in->pts;
s->first = s->frame_count;
switch (s->ascale) {
case ALINEAR:
for (c = 0; c < inlink->channels; c++) {
const float *src = (const float *)in->extended_data[c];
uint64_t *achistogram = &s->achistogram[(s->dmode == SINGLE ? 0: c) * w];
for (n = 0; n < in->nb_samples; n++) {
bin = lrint(av_clipf(fabsf(src[n]), 0, 1) * (w - 1));
achistogram[bin]++;
}
if (s->in[s->first] && s->count >= 0) {
uint64_t *shistogram = &s->shistogram[(s->dmode == SINGLE ? 0: c) * w];
const float *src2 = (const float *)s->in[s->first]->extended_data[c];
for (n = 0; n < in->nb_samples; n++) {
bin = lrint(av_clipf(fabsf(src2[n]), 0, 1) * (w - 1));
shistogram[bin]++;
}
}
}
break;
case ALOG:
for (c = 0; c < inlink->channels; c++) {
const float *src = (const float *)in->extended_data[c];
uint64_t *achistogram = &s->achistogram[(s->dmode == SINGLE ? 0: c) * w];
for (n = 0; n < in->nb_samples; n++) {
bin = lrint(av_clipf(1 + log10(fabsf(src[n])) / 6, 0, 1) * (w - 1));
achistogram[bin]++;
}
if (s->in[s->first] && s->count >= 0) {
uint64_t *shistogram = &s->shistogram[(s->dmode == SINGLE ? 0: c) * w];
const float *src2 = (const float *)s->in[s->first]->extended_data[c];
for (n = 0; n < in->nb_samples; n++) {
bin = lrint(av_clipf(1 + log10(fabsf(src2[n])) / 6, 0, 1) * (w - 1));
shistogram[bin]++;
}
}
}
break;
}
av_frame_free(&s->in[s->frame_count]);
s->in[s->frame_count] = in;
s->frame_count++;
if (s->frame_count > s->count)
s->frame_count = 0;
for (n = 0; n < w * s->dchannels; n++) {
acmax = FFMAX(s->achistogram[n] - s->shistogram[n], acmax);
}
for (c = 0; c < s->dchannels; c++) {
uint64_t *shistogram = &s->shistogram[c * w];
uint64_t *achistogram = &s->achistogram[c * w];
float yf, uf, vf;
if (s->dmode == SEPARATE) {
yf = 256.0f / s->dchannels;
uf = yf * M_PI;
vf = yf * M_PI;
uf *= 0.5 * sin((2 * M_PI * c) / s->dchannels);
vf *= 0.5 * cos((2 * M_PI * c) / s->dchannels);
}
for (n = 0; n < w; n++) {
double a, aa;
int h;
a = achistogram[n] - shistogram[n];
switch (s->scale) {
case LINEAR:
aa = a / (double)acmax;
break;
case SQRT:
aa = sqrt(a) / sqrt(acmax);
break;
case CBRT:
aa = cbrt(a) / cbrt(acmax);
break;
case LOG:
aa = log2(a + 1) / log2(acmax + 1);
break;
case RLOG:
aa = 1. - log2(a + 1) / log2(acmax + 1);
if (aa == 1.)
aa = 0;
break;
default:
av_assert0(0);
}
h = aa * (H - 1);
if (s->dmode == SINGLE) {
for (y = H - h; y < H; y++) {
s->out->data[0][y * s->out->linesize[0] + n] = 255;
s->out->data[3][y * s->out->linesize[0] + n] = 255;
}
if (s->h - H > 0) {
h = aa * 255;
s->out->data[0][s->ypos * s->out->linesize[0] + n] = h;
s->out->data[1][s->ypos * s->out->linesize[1] + n] = 127;
s->out->data[2][s->ypos * s->out->linesize[2] + n] = 127;
s->out->data[3][s->ypos * s->out->linesize[3] + n] = 255;
}
} else if (s->dmode == SEPARATE) {
float *out = &s->combine_buffer[3 * n];
int old;
old = s->out->data[0][(H - h) * s->out->linesize[0] + n];
for (y = H - h; y < H; y++) {
if (s->out->data[0][y * s->out->linesize[0] + n] != old)
break;
old = s->out->data[0][y * s->out->linesize[0] + n];
s->out->data[0][y * s->out->linesize[0] + n] = yf;
s->out->data[1][y * s->out->linesize[1] + n] = 128+uf;
s->out->data[2][y * s->out->linesize[2] + n] = 128+vf;
s->out->data[3][y * s->out->linesize[3] + n] = 255;
}
out[0] += aa * yf;
out[1] += aa * uf;
out[2] += aa * vf;
}
}
}
if (s->h - H > 0) {
if (s->dmode == SEPARATE) {
for (n = 0; n < w; n++) {
float *cb = &s->combine_buffer[3 * n];
s->out->data[0][s->ypos * s->out->linesize[0] + n] = cb[0];
s->out->data[1][s->ypos * s->out->linesize[1] + n] = cb[1];
s->out->data[2][s->ypos * s->out->linesize[2] + n] = cb[2];
s->out->data[3][s->ypos * s->out->linesize[3] + n] = 255;
}
}
if (s->slide == SCROLL) {
for (p = 0; p < 4; p++) {
for (y = s->h; y >= H + 1; y--) {
memmove(s->out->data[p] + (y ) * s->out->linesize[p],
s->out->data[p] + (y-1) * s->out->linesize[p], w);
}
}
}
s->ypos++;
if (s->slide == SCROLL || s->ypos >= s->h)
s->ypos = H;
}
clone = av_frame_clone(s->out);
if (!clone)
return AVERROR(ENOMEM);
return ff_filter_frame(outlink, clone);
}
static int activate(AVFilterContext *ctx)
{
AVFilterLink *inlink = ctx->inputs[0];
AVFilterLink *outlink = ctx->outputs[0];
AudioHistogramContext *s = ctx->priv;
AVFrame *in;
int ret;
FF_FILTER_FORWARD_STATUS_BACK(outlink, inlink);
ret = ff_inlink_consume_samples(inlink, s->nb_samples, s->nb_samples, &in);
if (ret < 0)
return ret;
if (ret > 0)
return filter_frame(inlink, in);
FF_FILTER_FORWARD_STATUS(inlink, outlink);
FF_FILTER_FORWARD_WANTED(outlink, inlink);
return FFERROR_NOT_READY;
}
static av_cold void uninit(AVFilterContext *ctx)
{
AudioHistogramContext *s = ctx->priv;
int i;
av_frame_free(&s->out);
av_freep(&s->shistogram);
av_freep(&s->achistogram);
av_freep(&s->combine_buffer);
for (i = 0; i < 101; i++)
av_frame_free(&s->in[i]);
}
static const AVFilterPad ahistogram_inputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_AUDIO,
.config_props = config_input,
},
{ NULL }
};
static const AVFilterPad ahistogram_outputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_VIDEO,
.config_props = config_output,
},
{ NULL }
};
AVFilter ff_avf_ahistogram = {
.name = "ahistogram",
.description = NULL_IF_CONFIG_SMALL("Convert input audio to histogram video output."),
.uninit = uninit,
.query_formats = query_formats,
.priv_size = sizeof(AudioHistogramContext),
.activate = activate,
.inputs = ahistogram_inputs,
.outputs = ahistogram_outputs,
.priv_class = &ahistogram_class,
};