DVD Seamless playing trouble.

[Dmitry Vergheles]

Hi DVD authoring folks!
The problem description:

There is test DVD image (34MB):
ftp://ftp.elecard.net.ru/incoming/SolveigMM/test.iso

It consists of one VTS:
VIDEO_TS.IFO, BUP
VTS_01_0.IFO, BUP
VTS_01_1.VOB

Cell length = PG= Part of title (chapter) = about 10 secs.

DVD Player (MPC and WMP) plays only first chapter, and then repeats a recent GOP but don't start to play a next chapter.

Seamless info in VTS_01_0.IFO seems to be ok (below are screenshots of IFoEditor).

It is known that DSI parts of Nav Packets contain VOBU_SRI fields unfilled (zeroized).

Could anyone suggest what is wrong with this image concerning seamless playback? Is VOBU_SRI used during playback?

I'll be appreciative of any suggestions!
Thanks in advance

ftp://ftp.elecard.net.ru/incoming/SolveigMM/vts_m.JPG

ftp://ftp.elecard.net.ru/incoming/SolveigMM/vts_pgc.JPG

[mpucoder]

I don't know what was used to produce that DVD, but it is one really bad multiplex. I'm surprised it plays at all.

I was able to demux and feed it to MuxMan, however I suspect this is a DVB file as MuxMan reported a maximum of 54 fields in one GOP, and the limit is 30 for PAL.

[Dmitry Vergheles]

I don't know what was used to produce that DVD, but it is one really bad multiplex. I'm surprised it plays at all.

I was able to demux and feed it to MuxMan, however I suspect this is a DVB file as MuxMan reported a maximum of 54 fields in one GOP, and the limit is 30 for PAL.

Do you think this is the problem for seamless playback?

[mpucoder]

Do I think the GOP spec violation is causing the problem? No, it is the horrible multiplex. This is a problem for any kind of playback. While I won't go into great detail, the relationship of SCR and pts values is all wrong leading to buffer overflow. Software players can handle this better as they usaually do not use SCR to control buffering, but hardware players do. After the first chapter the incoming data is nine seconds ahead of its presentation time. Buffer size aside, one second is the absolute limit to preloading any buffer. And I don't understand why all the padding - that is not how DVDs control bitrate. The DVD is also full of illegal values.

I will be moving this thread once I determine whether it is development, authoring, or digital capture. If this is software you are developing, as I suspect, you need to study the specs and other programs' output more. And since multiplexing is somewhat of a black art and highly competitive, I don't think the other authors of multiplexers, or myself, are going to troubleshoot this for you.

So, what did produce this?

[Dmitry Vergheles]

Do I think the GOP spec violation is causing the problem? No, it is the horrible multiplex. This is a problem for any kind of playback. While I won't go into great detail, the relationship of SCR and pts values is all wrong leading to buffer overflow. Software players can handle this better as they usaually do not use SCR to control buffering, but hardware players do. After the first chapter the incoming data is nine seconds ahead of its presentation time. Buffer size aside, one second is the absolute limit to preloading any buffer. And I don't understand why all the padding - that is not how DVDs control bitrate. The DVD is also full of illegal values.

I will be moving this thread once I determine whether it is development, authoring, or digital capture. If this is software you are developing, as I suspect, you need to study the specs and other programs' output more. And since multiplexing is somewhat of a black art and highly competitive, I don't think the other authors of multiplexers, or myself, are going to troubleshoot this for you.

So, what did produce this?

Hi mpucoder,
Thanks for the answer.
Sure you are right the test dvd image is an output of a software we are developing. Really we use Mainconcept-Elecard MPEG-2 Muxer (DVD compliant feature) but Navigation Packets add by ourselves. I know about SCR values not to be noticed by software playback. And therefeore it is strange that a next chapter is not started.

After the first chapter the incoming data is nine seconds ahead of its presentation time. Buffer size aside, one second is the absolute limit to preloading any buffer.

is it real trouble?

I have been thinking the reason is some DVD fields (PSI,DSI or IFOs ones) concerned with seamless playback are invalid.

As for specification, we have it but can't find an answer...
Could you please suggest any programs that can help us by their outputs?

[Dmitry Vergheles]

thanks to all, the problem seems to be solved.
After setting Next и Previous Video Vobu in VOBU_SRI field, chapters start to be played seamless.

[Afrinux]

thanks to all, the problem seems to be solved.
After setting Next и Previous Video Vobu in VOBU_SRI field, chapters start to be played seamless.
Did this dvd_video pass the DVD-Video specifications verifier?
Sometimes, standalone players can play two cell seamlessly, even if there is still some buffer management errors.
Afrinux