On Tue, 8 Dec 2015, Moritz Barsnick wrote:

On Mon, Dec 07, 2015 at 08:17:06 -0800, Roger Allen wrote:
I included a pointer at the end of the email.  I regret that was not more
obvious to you.  Please see
https://gist.github.com/rogerallen/426203d3a580898bd864

Apparently, the concat demuxer assigns DTS and PTS values to each
frame. For some peculiar reason, these alternate between 0 and
AV_NOPTS_VALUE:

(grep extract of loglevel debug - "ffprobe -show_frames" gives the same
info essentially:)

[concat @ 0xac22900] file:140 stream:0 pts:0 pts_time:0 dts:0 dts_time:0 -> 
pts:0 pts_time:0 dts:0 dts_time:0
[concat @ 0xac22900] file:141 stream:0 pts:0 pts_time:0 dts:0 dts_time:0 -> 
pts:-9223372036854 pts_time:-3.68935e+11 dts:-9223372036854 dts_time:-3.68935e+11
[concat @ 0xac22900] file:142 stream:0 pts:0 pts_time:0 dts:0 dts_time:0 -> 
pts:0 pts_time:0 dts:0 dts_time:0
[concat @ 0xac22900] file:143 stream:0 pts:0 pts_time:0 dts:0 dts_time:0 -> 
pts:-9223372036854 pts_time:-3.68935e+11 dts:-9223372036854 dts_time:-3.68935e+11
[concat @ 0xac22900] file:144 stream:0 pts:0 pts_time:0 dts:0 dts_time:0 -> 
pts:0 pts_time:0 dts:0 dts_time:0
[concat @ 0xac22900] file:145 stream:0 pts:0 pts_time:0 dts:0 dts_time:0 -> 
pts:-9223372036854 pts_time:-3.68935e+11 dts:-9223372036854 dts_time:-3.68935e+11
[concat @ 0xac22900] file:146 stream:0 pts:0 pts_time:0 dts:0 dts_time:0 -> 
pts:0 pts_time:0 dts:0 dts_time:0
[concat @ 0xac22900] file:147 stream:0 pts:0 pts_time:0 dts:0 dts_time:0 -> 
pts:-9223372036854 pts_time:-3.68935e+11 dts:-9223372036854 dts_time:-3.68935e+11
[concat @ 0xac22900] file:148 stream:0 pts:0 pts_time:0 dts:0 dts_time:0 -> 
pts:0 pts_time:0 dts:0 dts_time:0

(I had made a 150 image concat input file instead of one with only 15
images.)

Is this a regression?

Does this work OK if duration is specified in the concat file for each file?

I have a feeling that the concat demuxer never worked well with static images and unknown duration.

Regards,
Marton
_______________________________________________
ffmpeg-user mailing list
[email protected]
http://ffmpeg.org/mailman/listinfo/ffmpeg-user

Reply via email to