Hi!
Sven Over wrote:
> So, the point is, by now the indexer works like this: it searches the first N
> megabytes of the input file for video and audio streams. Already this is some
> kind of limitation, as you might have a big TS file, in which new audio
> streams start in the middle, for example an additional language stream for a
> movie. These streams are completely ignored by DVBCUT. So a new indexer
> should start at the beginning of the file and create a catalogue of *all*
> usable presentation units. That is, of all pictures in the video stream and
> all audio units. Unlike the current indexer, which only indexes the video
> stream.
There's even more to consider. Video format changes, for example. Or
subtitles. And while we're at it, the indexer should also calculate the
average and maximum bitrates of each (packaged) elementary stream. That
way it would be easy to tell if you can create a conforming DVD or not.
> DVBCUT now just seeks the input file to a place 2 seconds (not sure about the
> value) before the START marker, and reads all streams and decides at which
> exact place it begins to copy the audio packets. In the future, it should
> know from the index, what streams are available and which is the right
> position to start copying the streams.
Maybe we need several indexes:
- video frame directory
- video GOP directory
- audio frame directory for each stream
- subtitle directory
- stream directory
> While indexing, it should correct for inconsistencies like timestamp
> discontinuities (the ARD problem). When exporting videos, all timestamp
> values for the output file should be taken from the index then, no longer
> from the original input file. These values of course come from the input file
> in the first place, of course, but might have been corrected by the indexer.
> These correction can be done best, when continously reading the input file
> from the beginning to the end.
Not so easy. How can you tell the difference between a missing
presentation unit (transport error) and a broken timestamp in a
contiguous stream?
> The second big space for improvements is the muxer. The best solution would
> be
> to get really familiar with the sources of mplex and adopt them.
I don't think that's a good idea. Not only because we already had so
many hassles with third-party software (ffmpeg).
--
Michael "Tired" Riepe <[EMAIL PROTECTED]>
X-Tired: Each morning I get up I die a little
-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
DVBCUT-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dvbcut-user
