ffmpeg/libavfilter/vf_readeia608.c

565 lines
17 KiB
C
Raw Normal View History

/*
* Copyright (c) 2017 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
*/
/**
* @file
* Filter for reading closed captioning data (EIA-608).
* See also https://en.wikipedia.org/wiki/EIA-608
*/
#include <string.h>
#include "libavutil/internal.h"
#include "libavutil/opt.h"
#include "libavutil/pixdesc.h"
#include "libavutil/timestamp.h"
#include "avfilter.h"
#include "formats.h"
#include "internal.h"
#include "video.h"
#define LAG 25
#define CLOCK_BITSIZE_MIN 0.2f
#define CLOCK_BITSIZE_MAX 1.5f
#define SYNC_BITSIZE_MIN 12.f
#define SYNC_BITSIZE_MAX 15.f
typedef struct LineItem {
int input;
int output;
float unfiltered;
float filtered;
float average;
float deviation;
} LineItem;
typedef struct CodeItem {
uint8_t bit;
int size;
} CodeItem;
typedef struct ScanItem {
int nb_line;
int found;
int white;
int black;
uint64_t *histogram;
uint8_t byte[2];
CodeItem *code;
LineItem *line;
} ScanItem;
typedef struct ReadEIA608Context {
const AVClass *class;
int start, end;
float spw;
int chp;
int lp;
int depth;
int max;
int nb_allocated;
ScanItem *scan;
void (*read_line[2])(AVFrame *in, int nb_line,
LineItem *line, int lp, int w);
} ReadEIA608Context;
#define OFFSET(x) offsetof(ReadEIA608Context, x)
#define FLAGS AV_OPT_FLAG_VIDEO_PARAM|AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_RUNTIME_PARAM
static const AVOption readeia608_options[] = {
{ "scan_min", "set from which line to scan for codes", OFFSET(start), AV_OPT_TYPE_INT, {.i64=0}, 0, INT_MAX, FLAGS },
{ "scan_max", "set to which line to scan for codes", OFFSET(end), AV_OPT_TYPE_INT, {.i64=29}, 0, INT_MAX, FLAGS },
{ "spw", "set ratio of width reserved for sync code detection", OFFSET(spw), AV_OPT_TYPE_FLOAT, {.dbl=.27}, 0.1, 0.7, FLAGS },
{ "chp", "check and apply parity bit", OFFSET(chp), AV_OPT_TYPE_BOOL, {.i64= 0}, 0, 1, FLAGS },
{ "lp", "lowpass line prior to processing", OFFSET(lp), AV_OPT_TYPE_BOOL, {.i64= 1}, 0, 1, FLAGS },
{ NULL }
};
AVFILTER_DEFINE_CLASS(readeia608);
static int query_formats(AVFilterContext *ctx)
{
static const enum AVPixelFormat pixel_fmts[] = {
AV_PIX_FMT_GRAY8, AV_PIX_FMT_GRAY9,
AV_PIX_FMT_GRAY10, AV_PIX_FMT_GRAY12, AV_PIX_FMT_GRAY14,
AV_PIX_FMT_GRAY16,
AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUV411P,
AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUV422P,
AV_PIX_FMT_YUV440P, AV_PIX_FMT_YUV444P,
AV_PIX_FMT_YUVJ420P, AV_PIX_FMT_YUVJ422P,
AV_PIX_FMT_YUVJ440P, AV_PIX_FMT_YUVJ444P,
AV_PIX_FMT_YUVJ411P,
AV_PIX_FMT_YUV420P9, AV_PIX_FMT_YUV422P9, AV_PIX_FMT_YUV444P9,
AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10,
AV_PIX_FMT_YUV440P10,
AV_PIX_FMT_YUV444P12, AV_PIX_FMT_YUV422P12, AV_PIX_FMT_YUV420P12,
AV_PIX_FMT_YUV440P12,
AV_PIX_FMT_YUV444P14, AV_PIX_FMT_YUV422P14, AV_PIX_FMT_YUV420P14,
AV_PIX_FMT_YUV420P16, AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16,
AV_PIX_FMT_YUVA420P, AV_PIX_FMT_YUVA422P, AV_PIX_FMT_YUVA444P,
AV_PIX_FMT_YUVA444P9, AV_PIX_FMT_YUVA444P10, AV_PIX_FMT_YUVA444P12, AV_PIX_FMT_YUVA444P16,
AV_PIX_FMT_YUVA422P9, AV_PIX_FMT_YUVA422P10, AV_PIX_FMT_YUVA422P12, AV_PIX_FMT_YUVA422P16,
AV_PIX_FMT_YUVA420P9, AV_PIX_FMT_YUVA420P10, AV_PIX_FMT_YUVA420P16,
AV_PIX_FMT_NONE
};
return ff_set_common_formats_from_list(ctx, pixel_fmts);
}
static int config_filter(AVFilterContext *ctx, int start, int end)
{
ReadEIA608Context *s = ctx->priv;
AVFilterLink *inlink = ctx->inputs[0];
int size = inlink->w + LAG;
if (end >= inlink->h) {
av_log(ctx, AV_LOG_WARNING, "Last line to scan too large, clipping.\n");
end = inlink->h - 1;
}
if (start > end) {
av_log(ctx, AV_LOG_ERROR, "Invalid range.\n");
return AVERROR(EINVAL);
}
if (s->nb_allocated < end - start + 1) {
const int diff = end - start + 1 - s->nb_allocated;
s->scan = av_realloc_f(s->scan, end - start + 1, sizeof(*s->scan));
if (!s->scan)
return AVERROR(ENOMEM);
memset(&s->scan[s->nb_allocated], 0, diff * sizeof(*s->scan));
s->nb_allocated = end - start + 1;
}
for (int i = 0; i < s->nb_allocated; i++) {
ScanItem *scan = &s->scan[i];
if (!scan->histogram)
scan->histogram = av_calloc(s->max + 1, sizeof(*scan->histogram));
if (!scan->line)
scan->line = av_calloc(size, sizeof(*scan->line));
if (!scan->code)
scan->code = av_calloc(size, sizeof(*scan->code));
if (!scan->line || !scan->code || !scan->histogram)
return AVERROR(ENOMEM);
}
s->start = start;
s->end = end;
return 0;
}
static void build_histogram(ReadEIA608Context *s, ScanItem *scan, const LineItem *line, int len)
{
memset(scan->histogram, 0, (s->max + 1) * sizeof(*scan->histogram));
for (int i = LAG; i < len + LAG; i++)
scan->histogram[line[i].input]++;
}
static void find_black_and_white(ReadEIA608Context *s, ScanItem *scan)
{
const int max = s->max;
int start = 0, end = 0, middle;
int black = 0, white = 0;
int cnt;
for (int i = 0; i <= max; i++) {
if (scan->histogram[i]) {
start = i;
break;
}
}
for (int i = max; i >= 0; i--) {
if (scan->histogram[i]) {
end = i;
break;
}
}
middle = start + (end - start) / 2;
cnt = 0;
for (int i = start; i <= middle; i++) {
if (scan->histogram[i] > cnt) {
cnt = scan->histogram[i];
black = i;
}
}
cnt = 0;
for (int i = end; i >= middle; i--) {
if (scan->histogram[i] > cnt) {
cnt = scan->histogram[i];
white = i;
}
}
scan->black = black;
scan->white = white;
}
static float meanf(const LineItem *line, int len)
{
float sum = 0.0, mean = 0.0;
for (int i = 0; i < len; i++)
sum += line[i].filtered;
mean = sum / len;
return mean;
}
static float stddevf(const LineItem *line, int len)
{
float m = meanf(line, len);
float standard_deviation = 0.f;
for (int i = 0; i < len; i++)
standard_deviation += (line[i].filtered - m) * (line[i].filtered - m);
return sqrtf(standard_deviation / (len - 1));
}
static void thresholding(ReadEIA608Context *s, ScanItem *scan, LineItem *line,
int lag, float threshold, float influence, int len)
{
for (int i = lag; i < len + lag; i++) {
line[i].unfiltered = line[i].input / 255.f;
line[i].filtered = line[i].unfiltered;
}
for (int i = 0; i < lag; i++) {
line[i].unfiltered = meanf(line, len * s->spw);
line[i].filtered = line[i].unfiltered;
}
line[lag - 1].average = meanf(line, lag);
line[lag - 1].deviation = stddevf(line, lag);
for (int i = lag; i < len + lag; i++) {
if (fabsf(line[i].unfiltered - line[i-1].average) > threshold * line[i-1].deviation) {
if (line[i].unfiltered > line[i-1].average) {
line[i].output = 255;
} else {
line[i].output = 0;
}
line[i].filtered = influence * line[i].unfiltered + (1.f - influence) * line[i-1].filtered;
} else {
int distance_from_black, distance_from_white;
distance_from_black = FFABS(line[i].input - scan->black);
distance_from_white = FFABS(line[i].input - scan->white);
line[i].output = distance_from_black <= distance_from_white ? 0 : 255;
}
line[i].average = meanf(line + i - lag, lag);
line[i].deviation = stddevf(line + i - lag, lag);
}
}
static int periods(const LineItem *line, CodeItem *code, int len)
{
int hold = line[LAG].output, cnt = 0;
int last = LAG;
memset(code, 0, len * sizeof(*code));
for (int i = LAG + 1; i < len + LAG; i++) {
if (line[i].output != hold) {
code[cnt].size = i - last;
code[cnt].bit = hold;
hold = line[i].output;
last = i;
cnt++;
}
}
code[cnt].size = LAG + len - last;
code[cnt].bit = hold;
return cnt + 1;
}
static void dump_code(AVFilterContext *ctx, ScanItem *scan, int len, int item)
{
av_log(ctx, AV_LOG_DEBUG, "%d:", item);
for (int i = 0; i < len; i++) {
av_log(ctx, AV_LOG_DEBUG, " %03d", scan->code[i].size);
}
av_log(ctx, AV_LOG_DEBUG, "\n");
}
#define READ_LINE(type, name) \
static void read_##name(AVFrame *in, int nb_line, LineItem *line, int lp, int w) \
{ \
const type *src = (const type *)(&in->data[0][nb_line * in->linesize[0]]);\
\
if (lp) { \
for (int i = 0; i < w; i++) { \
int a = FFMAX(i - 3, 0); \
int b = FFMAX(i - 2, 0); \
int c = FFMAX(i - 1, 0); \
int d = FFMIN(i + 3, w-1); \
int e = FFMIN(i + 2, w-1); \
int f = FFMIN(i + 1, w-1); \
\
line[LAG + i].input = (src[a] + src[b] + src[c] + src[i] + \
src[d] + src[e] + src[f] + 6) / 7; \
} \
} else { \
for (int i = 0; i < w; i++) { \
line[LAG + i].input = src[i]; \
} \
} \
}
READ_LINE(uint8_t, byte)
READ_LINE(uint16_t, word)
static int config_input(AVFilterLink *inlink)
{
AVFilterContext *ctx = inlink->dst;
ReadEIA608Context *s = ctx->priv;
const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(inlink->format);
if (!desc)
return AVERROR_BUG;
s->depth = desc->comp[0].depth;
s->max = (1 << desc->comp[0].depth) - 1;
s->read_line[0] = read_byte;
s->read_line[1] = read_word;
return config_filter(ctx, s->start, s->end);
}
static void extract_line(AVFilterContext *ctx, AVFrame *in, ScanItem *scan, int w, int nb_line)
{
ReadEIA608Context *s = ctx->priv;
LineItem *line = scan->line;
int i, j, ch, len;
uint8_t codes[19] = { 0 };
float bit_size = 0.f;
int parity;
memset(line, 0, (w + LAG) * sizeof(*line));
scan->byte[0] = scan->byte[1] = 0;
scan->found = 0;
s->read_line[s->depth > 8](in, nb_line, line, s->lp, w);
build_histogram(s, scan, line, w);
find_black_and_white(s, scan);
if (scan->white - scan->black < 5)
return;
thresholding(s, scan, line, LAG, 1, 0, w);
len = periods(line, scan->code, w);
dump_code(ctx, scan, len, nb_line);
if (len < 15 ||
scan->code[14].bit != 0 ||
w / (float)scan->code[14].size < SYNC_BITSIZE_MIN ||
w / (float)scan->code[14].size > SYNC_BITSIZE_MAX) {
return;
}
for (i = 14; i < len; i++) {
bit_size += scan->code[i].size;
}
bit_size /= 19.f;
for (i = 1; i < 14; i++) {
if (scan->code[i].size / bit_size > CLOCK_BITSIZE_MAX ||
scan->code[i].size / bit_size < CLOCK_BITSIZE_MIN) {
return;
}
}
if (scan->code[15].size / bit_size < 0.45f) {
return;
}
for (j = 0, i = 14; i < len; i++) {
int run, bit;
run = lrintf(scan->code[i].size / bit_size);
bit = scan->code[i].bit;
for (int k = 0; j < 19 && k < run; k++) {
codes[j++] = bit;
}
if (j >= 19)
break;
}
for (ch = 0; ch < 2; ch++) {
for (parity = 0, i = 0; i < 8; i++) {
int b = codes[3 + ch * 8 + i];
if (b == 255) {
parity++;
b = 1;
} else {
b = 0;
}
scan->byte[ch] |= b << i;
}
if (s->chp) {
if (!(parity & 1)) {
scan->byte[ch] = 0x7F;
}
}
}
scan->nb_line = nb_line;
scan->found = 1;
}
static int extract_lines(AVFilterContext *ctx, void *arg,
int job, int nb_jobs)
{
ReadEIA608Context *s = ctx->priv;
AVFilterLink *inlink = ctx->inputs[0];
const int h = s->end - s->start + 1;
const int start = (h * job) / nb_jobs;
const int end = (h * (job+1)) / nb_jobs;
AVFrame *in = arg;
for (int i = start; i < end; i++) {
ScanItem *scan = &s->scan[i];
extract_line(ctx, in, scan, inlink->w, i);
}
return 0;
}
static int filter_frame(AVFilterLink *inlink, AVFrame *in)
{
AVFilterContext *ctx = inlink->dst;
AVFilterLink *outlink = ctx->outputs[0];
ReadEIA608Context *s = ctx->priv;
int nb_found;
ff_filter_execute(ctx, extract_lines, in, NULL,
FFMIN(FFMAX(s->end - s->start + 1, 1), ff_filter_get_nb_threads(ctx)));
nb_found = 0;
for (int i = 0; i < s->end - s->start + 1; i++) {
ScanItem *scan = &s->scan[i];
uint8_t key[128], value[128];
if (!scan->found)
continue;
//snprintf(key, sizeof(key), "lavfi.readeia608.%d.bits", nb_found);
//snprintf(value, sizeof(value), "0b%d%d%d%d%d%d%d%d 0b%d%d%d%d%d%d%d%d", codes[3]==255,codes[4]==255,codes[5]==255,codes[6]==255,codes[7]==255,codes[8]==255,codes[9]==255,codes[10]==255,codes[11]==255,codes[12]==255,codes[13]==255,codes[14]==255,codes[15]==255,codes[16]==255,codes[17]==255,codes[18]==255);
//av_dict_set(&in->metadata, key, value, 0);
snprintf(key, sizeof(key), "lavfi.readeia608.%d.cc", nb_found);
snprintf(value, sizeof(value), "0x%02X%02X", scan->byte[0], scan->byte[1]);
av_dict_set(&in->metadata, key, value, 0);
snprintf(key, sizeof(key), "lavfi.readeia608.%d.line", nb_found);
av_dict_set_int(&in->metadata, key, scan->nb_line, 0);
nb_found++;
}
return ff_filter_frame(outlink, in);
}
static av_cold void uninit(AVFilterContext *ctx)
{
ReadEIA608Context *s = ctx->priv;
for (int i = 0; i < s->nb_allocated; i++) {
ScanItem *scan = &s->scan[i];
av_freep(&scan->histogram);
av_freep(&scan->code);
av_freep(&scan->line);
}
s->nb_allocated = 0;
av_freep(&s->scan);
}
static int process_command(AVFilterContext *ctx, const char *cmd, const char *args,
char *res, int res_len, int flags)
{
ReadEIA608Context *s = ctx->priv;
int ret, start = s->start, end = s->end;
ret = ff_filter_process_command(ctx, cmd, args, res, res_len, flags);
if (ret < 0)
return ret;
ret = config_filter(ctx, s->start, s->end);
if (ret < 0) {
s->start = start;
s->end = end;
}
return 0;
}
static const AVFilterPad readeia608_inputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_VIDEO,
.filter_frame = filter_frame,
.config_props = config_input,
},
};
static const AVFilterPad readeia608_outputs[] = {
{
.name = "default",
.type = AVMEDIA_TYPE_VIDEO,
},
};
const AVFilter ff_vf_readeia608 = {
.name = "readeia608",
.description = NULL_IF_CONFIG_SMALL("Read EIA-608 Closed Caption codes from input video and write them to frame metadata."),
.priv_size = sizeof(ReadEIA608Context),
.priv_class = &readeia608_class,
2021-08-12 13:05:31 +02:00
FILTER_INPUTS(readeia608_inputs),
FILTER_OUTPUTS(readeia608_outputs),
avfilter: Replace query_formats callback with union of list and callback If one looks at the many query_formats callbacks in existence, one will immediately recognize that there is one type of default callback for video and a slightly different default callback for audio: It is "return ff_set_common_formats_from_list(ctx, pix_fmts);" for video with a filter-specific pix_fmts list. For audio, it is the same with a filter-specific sample_fmts list together with ff_set_common_all_samplerates() and ff_set_common_all_channel_counts(). This commit allows to remove the boilerplate query_formats callbacks by replacing said callback with a union consisting the old callback and pointers for pixel and sample format arrays. For the not uncommon case in which these lists only contain a single entry (besides the sentinel) enum AVPixelFormat and enum AVSampleFormat fields are also added to the union to store them directly in the AVFilter, thereby avoiding a relocation. The state of said union will be contained in a new, dedicated AVFilter field (the nb_inputs and nb_outputs fields have been shrunk to uint8_t in order to create a hole for this new field; this is no problem, as the maximum of all the nb_inputs is four; for nb_outputs it is only two). The state's default value coincides with the earlier default of query_formats being unset, namely that the filter accepts all formats (and also sample rates and channel counts/layouts for audio) provided that these properties agree coincide for all inputs and outputs. By using different union members for audio and video filters the type-unsafety of using the same functions for audio and video lists will furthermore be more confined to formats.c than before. When the new fields are used, they will also avoid allocations: Currently something nearly equivalent to ff_default_query_formats() is called after every successful call to a query_formats callback; yet in the common case that the newly allocated AVFilterFormats are not used at all (namely if there are no free links) these newly allocated AVFilterFormats are freed again without ever being used. Filters no longer using the callback will not exhibit this any more. Reviewed-by: Paul B Mahol <onemda@gmail.com> Reviewed-by: Nicolas George <george@nsup.org> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2021-09-27 12:07:35 +02:00
FILTER_QUERY_FUNC(query_formats),
.uninit = uninit,
.flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC | AVFILTER_FLAG_SLICE_THREADS,
.process_command = process_command,
};