View Single Post
Old 18th March 2020, 05:20   #653  |  Link
markfilipak
Registered User
 
markfilipak's Avatar
 
Join Date: Jul 2016
Location: Mansfield, Ohio (formerly San Jose, California)
Posts: 280
Quote:
Originally Posted by Mosu View Post
I'm actually not quite sure which particular feature that person is talking about. Normally mkvmerge will re-use existing timestamps for the container a track is read from.

A couple of possibilities:
  1. The "--default-duration" option is the most likely candidate in my opinion. You specify how long each frame or field (depending on how the video was encoded: interlaced or progressive) should be displayed, and mkvmerge will calculate the timestamps from it. mkvmerge will also set the track's "default duration" header field accordingly. With this option mkvmerge does indeed "generate timestamps" from your specification.
  2. With the "--timestamps" option you can specify a file which contains the timestamps for all the frames of a track. mkvmerge will use those timestamps instead of the ones from the source container. There are several file formats for that particular "timestamps file". The option is meant to be used by other tools, not really by humans. With this option mkvmerge does not "generate timestamps" itself, it just uses timestamps someone or something else has generated for mkvmerge.
  3. Maybe the person wasn't talking about timestamps after all but about chapters? mkvmerge can generate chapters based on timestamps, e.g. one chapte every X seconds.
Thank you. Yes, I was familiar with all these options. Below is the mailing list thread for your information and perspective. As you see from my query (at the end of the thread), I thought that MKVToolNix might provide a clue regarding the 4400ms audio lead. The early-audio was present only for PGCs 9 & 10 and the only thing I can think of was that the PTS was wrong.
On 03/15/2020 09:41 AM, Mark Filipak wrote:
> On 03/15/2020 09:02 AM, Crazy Red Elephant via ffmpeg-user wrote:
>>> Are they actually disruptive, or could you just keep the original
>>> stream as is, knowing 1 out of 120 frames or something will be
>>> dropped when playing back?
>>
>> To me, yes. I know some other users also reported something about
>> playback issues numerous times but the stream provider doesn't seem to
>> care.
>>
>> I think I mentioned before that remuxing the video with MKVToolNix and
>> the "generate timestamps" option checked...
>
> Can you help me find the "generate timestamps" option? I can't find it.
> I'd like to see what it does with a VOB that has 8 PGCs that are okay,
> but the 9th & 10th PGCs have audio in advance of video by about 4400ms.
>
> Regards,
> Mark.

Last edited by markfilipak; 18th March 2020 at 05:45. Reason: Clarified; "PTS" was "PTC" oops
markfilipak is offline   Reply With Quote