Use Matroska A_QUICKTIME instead of deprecated A_QUICKTIME/**** (A_QUICKTIME/QDMC, A_QUICKTIME/QDM2) Codec IDs.

[Vika Apelsinova]

FFMPEG should produce files with Matroska A_QUICKTIME instead of deprecated A_QUICKTIME/ (A_QUICKTIME/QDMC, A_QUICKTIME/QDM2) Codec IDs.

Please see #3256 for more info.

[Vika Apelsinova]

Please anybody can fix it? Looks like it is not too difficult, but I'm looking in the FFMPEG source code and don't know what should be changed in the sources to fix it and therefore can't make the patch. :-[

Thanks in advance!

[Carl Eugen Hoyos]

Which player / application fails for files created by current FFmpeg?

[Vika Apelsinova]

Unfortunately, I can't test QDMC. Looks like QDMC is not supported by any open source software. Is FFMPEG support it? Is there open source QDMC decoder at all?

Regarding QDM2, please look ​https://trac.ffmpeg.org/ticket/3256#comment:12. I just can't make the patch because don't know what should be changed in the source code.

Over 5 years (see ​https://trac.bunkus.org/ticket/354#comment:1) there was no player that I know that supports QDM2 in MKV until ​https://trac.videolan.org/vlc/ticket/11038 was recently closed.

All players now that support A_QUICKTIME, support deprecated A_QUICKTIME/QDM2 CodecID too (for legacy compatibility purposes), therefore I can't find player that fails for files created by current version of FFMPEG, but according to Matroska specs (​http://matroska.org/technical/specs/codecid/index.html) A_QUICKTIME/QDMC and A_QUICKTIME/QDM2 CodecIDs is deprecated in favor of A_QUICKTIME (without a trailing codec name) and FFMPEG should not produce files with deprecated CodecIDs and should produce files that correspond to current specifications.

[Carl Eugen Hoyos]

Why?

I mean: Why do you suggest that FFmpeg should create files that some players (applications) will not support while you apparently found out that all relevant players support files created by current FFmpeg?
Don't you agree that we should at least wait some time before we update the code to output incompatible files by default?

[Vika Apelsinova]

Replying to cehoyos:

Why?

Because it is deprecated.

Don't you agree that we should at least wait some time before we update the code to output incompatible files by default?

Please explain what you mean when you wrote about "output incompatible files by default"?

Is there players that support A_QUICKTIME/QDM2 and/or A_QUICKTIME/QDMC, but not support A_QUICKTIME (without a trailing codec name)?

BTW, MKVToolNix has always produced MKV files using A_QUICKTIME (without a trailing codec name) Codec ID according to specs.

[Carl Eugen Hoyos]

Replying to Vika Apelsinova:

Is there players that support A_QUICKTIME/QDM2 and/or A_QUICKTIME/QDMC, but not support A_QUICKTIME (without a trailing codec name)?

Not very old versions of MPlayer and FFmpeg come to mind...

[Carl Eugen Hoyos]

Replying to Vika Apelsinova:

Over 5 years (see ​https://trac.bunkus.org/ticket/354#comment:1) there was no player that I know that supports QDM2 in MKV until ​https://trac.videolan.org/vlc/ticket/11038 was recently closed.

I just tested a five year old version of ffplay and it works fine and I would be surprised if it didn't also work with MPlayer.

[Vika Apelsinova]

Replying to cehoyos:

Don't you agree that we should at least wait some time before we update the code to output incompatible files by default?

Looks like the latest versions of all popular open source players at this moment support A_QUICKTIME (without a trailing codec name) scheme (VLC, MPayer, FFMPEG, etc.). How much time is necessary to wait and for what? Is there hardware players that don't support decoding files with A_QUICKTIME (without a trailing codec name) Codec ID?

FFMPEG producing MKV files using A_QUICKTIME/QDM2, MKVToolNix producing MKV files using A_QUICKTIME (without a trailing codec name) according to specs. Don't you agree that we need standardization?

[Carl Eugen Hoyos]

I thought that there is a standard (that mentions two ways of muxing QDM2 in mkv), am I wrong?

Could you explain why this is so important for you? Maybe I miss something but breaking a use case for some (a few) users seems not a very good idea to me, especially since according to you, it would not fix anything for any user.