video: add insane hack to work around FFmpeg/Libav insanity

So, FFmpeg/Libav requires us to figure out video timestamps ourselves
(see last 10 commits or so), but the methods it provides for this aren't
even sufficient. In particular, everything that uses AVI-style DTS (avi,
vfw-muxed mkv, possibly mpeg4-in-ogm) with a codec that has an internal
frame delay is broken. In this case, libavcodec will shift the packet-
to-image correspondence by the codec delay, meaning that with a delay=1,
the first AVFrame.pkt_dts is not 0, but that of the second packet. All
timestamps will appear shifted. The start time (e.g. the time displayed
when doing "mpv file.avi --pause") will not be exactly 0.

(According to Libav developers, this is how it's supposed to work; just
that the first DTS values are normally negative with formats that use
DTS "properly". Who cares if it doesn't work at all with very common
video formats? There's no indication that they'll fix this soon,
either. An elegant workaround is missing too.)

Add a hack to re-enable the old PTS code for AVI and vfw-muxed MKV.
Since these timestamps are not reorderd, we wouldn't need to sort them,
but it's less code this way (and possibly more robust, should a demuxer
unexpectedly output PTS).

The original intention of all the timestamp changes recently was
actually to get rid of demuxer-specific hacks and the old timestamp
sorting code, but it looks like this didn't work out. Yet another case
where trying to replace native MPlayer functionality with FFmpeg/Libav
led to disadvantages and bugs. (Note that the old PTS sorting code
doesn't and can't handle frame dropping correctly, though.)

Bug reports:

 https://trac.ffmpeg.org/ticket/3178

 https://bugzilla.libav.org/show_bug.cgi?id=600
This commit is contained in:
wm4 2013-11-28 15:10:45 +01:00
parent d9b5dedfe9
commit dc0b2046cd
4 changed files with 13 additions and 3 deletions

View File

@ -441,6 +441,9 @@ static void handle_stream(demuxer_t *demuxer, int i)
/ (float)(codec->height * codec->sample_aspect_ratio.den);
sh_video->i_bps = codec->bit_rate / 8;
// This also applies to vfw-muxed mkv, but we can't detect these easily.
sh_video->avi_dts = matches_avinputformat_name(priv, "avi");
mp_msg(MSGT_DEMUX, MSGL_DBG2, "aspect= %d*%d/(%d*%d)\n",
codec->width, codec->sample_aspect_ratio.num,
codec->height, codec->sample_aspect_ratio.den);

View File

@ -1338,6 +1338,7 @@ static int demux_mkv_open_video(demuxer_t *demuxer, mkv_track_t *track)
sh_v->disp_h = track->v_dheight;
}
mp_msg(MSGT_DEMUX, MSGL_V, "[mkv] Aspect: %f\n", sh_v->aspect);
sh_v->avi_dts = track->ms_compat;
return 0;
}

View File

@ -76,6 +76,7 @@ typedef struct sh_audio {
} sh_audio_t;
typedef struct sh_video {
bool avi_dts; // use DTS timing; first frame and DTS is 0
float fps; // frames per second (set only if constant fps)
float aspect; // aspect ratio stored in the file (for prescaling)
int i_bps; // == bitrate (compressed bytes/sec)

View File

@ -262,9 +262,12 @@ static double retrieve_sorted_pts(struct dec_video *d_video, double codec_pts)
d_video->num_unsorted_pts_problems++;
d_video->unsorted_pts = codec_pts;
if (opts->user_pts_assoc_mode)
if (d_video->header->video->avi_dts) {
// Actually, they don't need to be sorted, we just reuse the buffering.
d_video->pts_assoc_mode = 2;
} else if (opts->user_pts_assoc_mode) {
d_video->pts_assoc_mode = opts->user_pts_assoc_mode;
else if (d_video->pts_assoc_mode == 0) {
} else if (d_video->pts_assoc_mode == 0) {
if (codec_pts != MP_NOPTS_VALUE)
d_video->pts_assoc_mode = 1;
else
@ -292,7 +295,9 @@ struct mp_image *video_decode(struct dec_video *d_video,
int drop_frame)
{
struct MPOpts *opts = d_video->opts;
bool sort_pts = opts->user_pts_assoc_mode != 1 && opts->correct_pts;
bool sort_pts =
(opts->user_pts_assoc_mode != 1 || d_video->header->video->avi_dts)
&& opts->correct_pts;
double pkt_pts = packet ? packet->pts : MP_NOPTS_VALUE;
double pkt_dts = packet ? packet->dts : MP_NOPTS_VALUE;