ffmpeg - MPEG TS - wrong ES stream_type in PMT for AC3

[palettenheinze]

Hi folks,
I am using ffmpeg to encode and mux an MPEG - TS with h264/x264 video and AC3 2.0 audio. So a pretty straightforward thing, output shall be used for VoD on DVB cable platform.

Now I have two major problems:

- using the standard AC3 encoder (or ac3_fixed) I get a wrong stream_type for the audio in the PMT. PIDs, PES IDs, everything correct but in DVB it should put in the PMT as stream_type for AC3 audio a 6/0x6. But ffmpeg puts the value for ATSC, being the 129/0x81. Using the ffmpeg E-AC3 encoder it works correctly, putting "6" there but i am supposed NOT to use E-AC3..

- second thing: is there a way to get rid of the service description table stream in the TS? It is using the PID 0x11 and I am supposed to leave it out completely

I am using a self-compiled ffmpeg version on Ubuntu, source straight from ffmpeg.org, Nov., 18th.

Thanks for help!
PH

[palettenheinze]

-bump-

wrong place for this issue?

[El Heggunte]

-bump-

wrong place for this issue?

Probably

You MAY have better luck @ Doom9.org . Even better, why not go directly to one of the ffmpeg mailing lists
And I think it wouldn't hurt to try to get in touch directly with Herr Niedermayer or Reimar Döffinger themselves...

[Selur]

posting to the bug tracker 8https://trac.ffmpeg.org/9might be a good idea too

[poisondeathray]

Not sure if they will help, but you might have a look at dvblast or opencaster

http://www.videolan.org/projects/dvblast.html
http://www.avalpa.com/the-key-values/15-free-software/33-opencaster

There are a lot of compliance issues with ffmpeg ts muxes

[palettenheinze]

Hey guys,
yes I´ll use a different channel then. Interesting tools, too. Not sure yet if they serve my purpose.

From what I read in the lists, MPEG TS muxing with ffmpeg is indeed not in the phase of being standards compliant. Still, I wanted to try how far I´d get, the error I found being one seemingly easy to correct.

But of which I am unsure is which criteria should be used for checking compliance, "T-STD" being mentioned in the list. I researched a bit further and found the "ETSI TR 101-290, measurement guidelines for monitoring MPEG Transports Streams". Don´t know how these two exactly relate to each other as the TR 101-290 contains certain buffer tests.

I found the TR 101-290 tests in some Software: many of the issues are covered with the "StreamGuru" MPEG analyzer. For ex. it found a too large variation of PCR interval. Elecard Stream Analyzer would cover all the issues from TR 101-290. Do you think there still would the need for some really costly stuff from Tektronix, or similar!? In the ffmpeg list they would use this as a tool of reference for further development.

Cheers,
PH