2013-06-01 17:54:31 +00:00
|
|
|
/*
|
|
|
|
* This file is part of mpv.
|
|
|
|
*
|
|
|
|
* mpv is free software; you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
* (at your option) any later version.
|
|
|
|
*
|
|
|
|
* mpv 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 General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License along
|
|
|
|
* with mpv. If not, see <http://www.gnu.org/licenses/>.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <stdlib.h>
|
|
|
|
#include <assert.h>
|
|
|
|
|
|
|
|
#include <libavcodec/avcodec.h>
|
|
|
|
#include <libavutil/intreadwrite.h>
|
|
|
|
#include <libavutil/common.h>
|
|
|
|
|
2013-06-02 17:38:57 +00:00
|
|
|
#include "config.h"
|
|
|
|
|
2013-06-01 17:54:31 +00:00
|
|
|
#include "talloc.h"
|
2013-12-17 01:39:45 +00:00
|
|
|
#include "common/msg.h"
|
|
|
|
#include "common/av_common.h"
|
2014-08-29 10:09:04 +00:00
|
|
|
#include "misc/bstr.h"
|
2013-06-01 17:54:31 +00:00
|
|
|
#include "sd.h"
|
|
|
|
|
2014-03-16 09:04:46 +00:00
|
|
|
#define HAVE_AV_WEBVTT (LIBAVCODEC_VERSION_MICRO >= 100)
|
2013-08-24 18:06:45 +00:00
|
|
|
|
2013-06-01 17:54:31 +00:00
|
|
|
struct sd_lavc_priv {
|
|
|
|
AVCodecContext *avctx;
|
|
|
|
};
|
|
|
|
|
2013-08-24 13:17:37 +00:00
|
|
|
static const char *get_lavc_format(const char *format)
|
|
|
|
{
|
|
|
|
// For the hack involving parse_webvtt().
|
|
|
|
if (format && strcmp(format, "webvtt-webm") == 0)
|
|
|
|
format = "webvtt";
|
|
|
|
return format;
|
|
|
|
}
|
|
|
|
|
2013-06-01 17:54:31 +00:00
|
|
|
static bool supports_format(const char *format)
|
|
|
|
{
|
2013-08-24 13:17:37 +00:00
|
|
|
format = get_lavc_format(format);
|
2013-06-01 17:54:31 +00:00
|
|
|
enum AVCodecID cid = mp_codec_to_av_codec_id(format);
|
2015-09-01 21:48:14 +00:00
|
|
|
AVCodec *codec = avcodec_find_decoder(cid);
|
2013-06-01 17:54:31 +00:00
|
|
|
const AVCodecDescriptor *desc = avcodec_descriptor_get(cid);
|
2015-09-01 21:48:14 +00:00
|
|
|
return codec && desc && desc->type == AVMEDIA_TYPE_SUBTITLE;
|
2013-06-01 17:54:31 +00:00
|
|
|
}
|
|
|
|
|
sd_lavc_conv: strip style header
Normally, libavcodec subtitle converters will output a style header like
this as part of the extradata:
Style: Default,Arial,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,1,1,0,2,10,10,10,0,0
We don't want that, so use some bruteforce to get rid of them.
2013-06-02 18:34:20 +00:00
|
|
|
// Disable style definitions generated by the libavcodec converter.
|
|
|
|
// We always want the user defined style instead.
|
|
|
|
static void disable_styles(bstr header)
|
|
|
|
{
|
|
|
|
while (header.len) {
|
|
|
|
int n = bstr_find(header, bstr0("\nStyle: "));
|
|
|
|
if (n < 0)
|
|
|
|
break;
|
|
|
|
header.start[n + 1] = '#'; // turn into a comment
|
|
|
|
header = bstr_cut(header, 2);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-06-01 17:54:31 +00:00
|
|
|
static int init(struct sd *sd)
|
|
|
|
{
|
|
|
|
struct sd_lavc_priv *priv = talloc_zero(NULL, struct sd_lavc_priv);
|
|
|
|
AVCodecContext *avctx = NULL;
|
2013-08-24 13:17:37 +00:00
|
|
|
const char *fmt = get_lavc_format(sd->codec);
|
|
|
|
AVCodec *codec = avcodec_find_decoder(mp_codec_to_av_codec_id(fmt));
|
2013-06-01 17:54:31 +00:00
|
|
|
if (!codec)
|
|
|
|
goto error;
|
|
|
|
avctx = avcodec_alloc_context3(codec);
|
|
|
|
if (!avctx)
|
|
|
|
goto error;
|
|
|
|
avctx->extradata_size = sd->extradata_len;
|
|
|
|
avctx->extradata = sd->extradata;
|
|
|
|
if (avcodec_open2(avctx, codec, NULL) < 0)
|
|
|
|
goto error;
|
|
|
|
// Documented as "set by libavcodec", but there is no other way
|
|
|
|
avctx->time_base = (AVRational) {1, 1000};
|
|
|
|
priv->avctx = avctx;
|
|
|
|
sd->priv = priv;
|
2013-06-23 20:09:04 +00:00
|
|
|
sd->output_codec = "ssa";
|
2013-06-01 17:54:31 +00:00
|
|
|
sd->output_extradata = avctx->subtitle_header;
|
|
|
|
sd->output_extradata_len = avctx->subtitle_header_size;
|
sd_lavc_conv: strip style header
Normally, libavcodec subtitle converters will output a style header like
this as part of the extradata:
Style: Default,Arial,16,&Hffffff,&Hffffff,&H0,&H0,0,0,0,1,1,0,2,10,10,10,0,0
We don't want that, so use some bruteforce to get rid of them.
2013-06-02 18:34:20 +00:00
|
|
|
if (sd->output_extradata) {
|
|
|
|
sd->output_extradata = talloc_memdup(sd, sd->output_extradata,
|
|
|
|
sd->output_extradata_len);
|
|
|
|
disable_styles((bstr){sd->output_extradata, sd->output_extradata_len});
|
|
|
|
}
|
2013-06-01 17:54:31 +00:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
error:
|
2013-12-21 18:06:37 +00:00
|
|
|
MP_FATAL(sd, "Could not open libavcodec subtitle converter\n");
|
2013-06-01 17:54:31 +00:00
|
|
|
av_free(avctx);
|
|
|
|
talloc_free(priv);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
2013-08-24 18:06:45 +00:00
|
|
|
#if HAVE_AV_WEBVTT
|
|
|
|
|
2013-08-24 13:17:37 +00:00
|
|
|
// FFmpeg WebVTT packets are pre-parsed in some way. The FFmpeg Matroska
|
|
|
|
// demuxer does this on its own. In order to free our demuxer_mkv.c from
|
|
|
|
// codec-specific crud, we do this here.
|
|
|
|
// Copied from libavformat/matroskadec.c (FFmpeg 818ebe9 / 2013-08-19)
|
|
|
|
// License: LGPL v2.1 or later
|
|
|
|
// Author header: The FFmpeg Project
|
|
|
|
// Modified in some ways.
|
|
|
|
static int parse_webvtt(AVPacket *in, AVPacket *pkt)
|
|
|
|
{
|
|
|
|
uint8_t *id, *settings, *text, *buf;
|
|
|
|
int id_len, settings_len, text_len;
|
|
|
|
uint8_t *p, *q;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
uint8_t *data = in->data;
|
|
|
|
int data_len = in->size;
|
|
|
|
|
|
|
|
if (data_len <= 0)
|
|
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
|
|
|
|
p = data;
|
|
|
|
q = data + data_len;
|
|
|
|
|
|
|
|
id = p;
|
|
|
|
id_len = -1;
|
|
|
|
while (p < q) {
|
|
|
|
if (*p == '\r' || *p == '\n') {
|
|
|
|
id_len = p - id;
|
|
|
|
if (*p == '\r')
|
|
|
|
p++;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
p++;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (p >= q || *p != '\n')
|
|
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
p++;
|
|
|
|
|
|
|
|
settings = p;
|
|
|
|
settings_len = -1;
|
|
|
|
while (p < q) {
|
|
|
|
if (*p == '\r' || *p == '\n') {
|
|
|
|
settings_len = p - settings;
|
|
|
|
if (*p == '\r')
|
|
|
|
p++;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
p++;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (p >= q || *p != '\n')
|
|
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
p++;
|
|
|
|
|
|
|
|
text = p;
|
|
|
|
text_len = q - p;
|
|
|
|
while (text_len > 0) {
|
|
|
|
const int len = text_len - 1;
|
|
|
|
const uint8_t c = p[len];
|
|
|
|
if (c != '\r' && c != '\n')
|
|
|
|
break;
|
|
|
|
text_len = len;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (text_len <= 0)
|
|
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
|
|
|
|
err = av_new_packet(pkt, text_len);
|
|
|
|
if (err < 0)
|
|
|
|
return AVERROR(err);
|
|
|
|
|
|
|
|
memcpy(pkt->data, text, text_len);
|
|
|
|
|
|
|
|
if (id_len > 0) {
|
|
|
|
buf = av_packet_new_side_data(pkt,
|
|
|
|
AV_PKT_DATA_WEBVTT_IDENTIFIER,
|
|
|
|
id_len);
|
|
|
|
if (buf == NULL) {
|
2015-10-28 22:48:56 +00:00
|
|
|
av_packet_unref(pkt);
|
2013-08-24 13:17:37 +00:00
|
|
|
return AVERROR(ENOMEM);
|
|
|
|
}
|
|
|
|
memcpy(buf, id, id_len);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (settings_len > 0) {
|
|
|
|
buf = av_packet_new_side_data(pkt,
|
|
|
|
AV_PKT_DATA_WEBVTT_SETTINGS,
|
|
|
|
settings_len);
|
|
|
|
if (buf == NULL) {
|
2015-10-28 22:48:56 +00:00
|
|
|
av_packet_unref(pkt);
|
2013-08-24 13:17:37 +00:00
|
|
|
return AVERROR(ENOMEM);
|
|
|
|
}
|
|
|
|
memcpy(buf, settings, settings_len);
|
|
|
|
}
|
|
|
|
|
|
|
|
pkt->pts = in->pts;
|
|
|
|
pkt->duration = in->duration;
|
2015-09-29 16:43:28 +00:00
|
|
|
#if !HAVE_AV_AVPACKET_INT64_DURATION
|
2013-08-24 13:17:37 +00:00
|
|
|
pkt->convergence_duration = in->convergence_duration;
|
2015-09-29 16:43:28 +00:00
|
|
|
#endif
|
2013-08-24 13:17:37 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2013-08-24 18:06:45 +00:00
|
|
|
#else
|
|
|
|
|
|
|
|
static int parse_webvtt(AVPacket *in, AVPacket *pkt)
|
|
|
|
{
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif
|
|
|
|
|
2013-06-01 17:54:31 +00:00
|
|
|
static void decode(struct sd *sd, struct demux_packet *packet)
|
|
|
|
{
|
|
|
|
struct sd_lavc_priv *priv = sd->priv;
|
|
|
|
AVCodecContext *avctx = priv->avctx;
|
|
|
|
AVSubtitle sub = {0};
|
|
|
|
AVPacket pkt;
|
2013-08-24 13:17:37 +00:00
|
|
|
AVPacket parsed_pkt = {0};
|
2013-06-01 17:54:31 +00:00
|
|
|
int ret, got_sub;
|
|
|
|
|
av_common: add timebase parameter to mp_set_av_packet()
If the timebase is set, it's used for converting the packet timestamps.
Otherwise, the previous method of reinterpret-casting the mpv style
double timestamps to libavcodec style int64_t timestamps is used.
Also replace the kind of awkward mp_get_av_frame_pkt_ts() function by
mp_pts_from_av(), which simply converts timestamps in a way the old
function did. (Plus it takes a timebase parameter, similar to the
addition to mp_set_av_packet().)
Note that this should not change anything yet. The code in ad_lavc.c and
vd_lavc.c passes NULL for the timebase parameters. We could set
AVCodecContext.pkt_timebase and use that if we want to give libavcodec
"proper" timestamps.
This could be important for ad_lavc.c: some codecs (opus, probably mp3
and aac too) have weird requirements about doing decoding preroll on the
container level, and thus require adjusting the audio start timestamps
in some cases. libavcodec doesn't tell us how much was skipped, so we
either get shifted timestamps (by the length of the skipped data), or we
give it proper timestamps. (Note: libavcodec interprets or changes
timestamps only if pkt_timebase is set, which by default it is not.)
This would require selecting a timebase though, so I feel uncomfortable
with the idea. At least this change paves the way, and will allow some
testing.
2013-12-04 19:12:14 +00:00
|
|
|
mp_set_av_packet(&pkt, packet, &avctx->time_base);
|
2013-06-01 17:54:31 +00:00
|
|
|
|
2013-08-24 13:17:37 +00:00
|
|
|
if (sd->codec && strcmp(sd->codec, "webvtt-webm") == 0) {
|
|
|
|
if (parse_webvtt(&pkt, &parsed_pkt) < 0) {
|
2013-12-21 18:06:37 +00:00
|
|
|
MP_ERR(sd, "Error parsing subtitle\n");
|
2013-08-24 13:17:37 +00:00
|
|
|
goto done;
|
|
|
|
}
|
|
|
|
pkt = parsed_pkt;
|
|
|
|
}
|
|
|
|
|
2013-06-01 17:54:31 +00:00
|
|
|
ret = avcodec_decode_subtitle2(avctx, &sub, &got_sub, &pkt);
|
|
|
|
if (ret < 0) {
|
2013-12-21 18:06:37 +00:00
|
|
|
MP_ERR(sd, "Error decoding subtitle\n");
|
2013-06-01 17:54:31 +00:00
|
|
|
} else if (got_sub) {
|
|
|
|
for (int i = 0; i < sub.num_rects; i++) {
|
2015-09-01 21:48:14 +00:00
|
|
|
if (sub.rects[i]->w > 0 && sub.rects[i]->h > 0)
|
|
|
|
MP_WARN(sd, "Ignoring bitmap subtitle.\n");
|
2013-06-01 17:54:31 +00:00
|
|
|
char *ass_line = sub.rects[i]->ass;
|
|
|
|
if (!ass_line)
|
|
|
|
break;
|
|
|
|
// This might contain embedded timestamps, using the "old" ffmpeg
|
|
|
|
// ASS packet format, in which case pts/duration might be ignored
|
|
|
|
// at a later point.
|
|
|
|
sd_conv_add_packet(sd, ass_line, strlen(ass_line),
|
sub: do not clear subtitle list on seeking
This affects non-ASS text subtitles (those which go through libavcodec's
subtitle converter), which are muxed with video/audio. (Typically srt
subs in mkv.)
The problem is that seeking in the file can send a subtitle packet to
the decoder multiple times. These packets are interlaved with video,
and thus can't be all read when opening the file. Rather, subtitle
packets can essentially be randomly skipped or repeated (by seeking).
Until recently, this was solved by scanning the libass event list for
duplicates. Then our builtin srt-to-ass converter was removed, and
the problem was handled by fully clearing the subtitle list on each
seek.
This resulted in sub-seek not working properly for this type of file.
Since the subtitle list was cleared on seek, it was not possible to
do e.g. sub-seeks to subtitles before the current playback position.
Fix this by not clearing the list, and intead explicitly rejecting
duplicate packets. We use the packet file position was unique ID for
subtitles; this is confirmed working for most file formats (although
it is slightly risky - new demuxers may not necessarily set the file
position to something unique, or at all).
The list of seen packets is sorted, and the lookup uses binary search.
This is to avoid quadratic complexity when subtitles are added in
bulks, such as when opening a text subtitle file.
In some places, the code has to be adjusted to pass through the packet
file position correctly.
2015-12-17 23:51:57 +00:00
|
|
|
packet->pts, packet->duration, packet->pos + i);
|
2013-06-01 17:54:31 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-08-24 13:17:37 +00:00
|
|
|
done:
|
2013-06-01 17:54:31 +00:00
|
|
|
avsubtitle_free(&sub);
|
2015-10-28 22:48:56 +00:00
|
|
|
av_packet_unref(&parsed_pkt);
|
2013-06-01 17:54:31 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void reset(struct sd *sd)
|
|
|
|
{
|
|
|
|
struct sd_lavc_priv *priv = sd->priv;
|
|
|
|
|
|
|
|
avcodec_flush_buffers(priv->avctx);
|
|
|
|
sd_conv_def_reset(sd);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void uninit(struct sd *sd)
|
|
|
|
{
|
|
|
|
struct sd_lavc_priv *priv = sd->priv;
|
|
|
|
|
|
|
|
avcodec_close(priv->avctx);
|
|
|
|
av_free(priv->avctx);
|
|
|
|
talloc_free(priv);
|
|
|
|
}
|
|
|
|
|
|
|
|
const struct sd_functions sd_lavc_conv = {
|
2013-06-03 19:49:39 +00:00
|
|
|
.name = "lavc_conv",
|
2013-06-01 17:54:31 +00:00
|
|
|
.supports_format = supports_format,
|
|
|
|
.init = init,
|
|
|
|
.decode = decode,
|
|
|
|
.get_converted = sd_conv_def_get_converted,
|
|
|
|
.reset = reset,
|
|
|
|
.uninit = uninit,
|
|
|
|
};
|