H.264 and PCM in .mp4 wrapper - ever done?

[benwaggoner]

Hello all,

Continuing to work away at the H.264 Mezzanine question, I was wondering if anyone has ever tried using H.264 and PCM audio in a .mp4 wrapper together?

I have tools that can do this in a transport stream. And there certainly is an ISO mapping for uncompressed audio. But I don't seem to have any good way to make .mp4 like this myself. Anyone ever done this, or have any suggestions for how to do so?

[JEEB]

And there certainly is an ISO mapping for uncompressed audio.
Waitwaitwait, can you specify which spec actually specifies that? Because as far as I've understood, neither part 12 or 14 support raw PCM audio. Naturally MOV, the thing that both of those are based on, does support raw PCM audio though.

If it really is specified, I think nothing really stops people from implementing it in either libavformat or L-SMASH.

Edit:
OK, so the amendment draft "ISO/IEC 14496-1:2010/PDAM 2 (Support for raw audio-visual data)" seems to have it, but it's still in draft... It's not implemented in L-SMASH yet as the decoder specific info structure might still change. Same for libavformat most probably.

[VFR maniac]

http://www.itscj.ipsj.or.jp/sc29/open/29view/29n12734.zip
At my glance, the decoder specific info structure seems somewhat awkward.
e.g. about raw video, 'fps' field doesn't consider NTSC timebase.
e.g. about raw audio, there is no indication of the number of audio samples in the last audio frame.

So, I expect the definition will change later.

[poisondeathray]

Hello,

any updates on this topic?

ffmbc can do pcm audio in mp4

[JEEB]

I will have to note here that the specification is still not ratified (http://www.iso.org/iso/home/store/catalogue_tc/catalogue_detail.htm?csnumber=62114) (stage 40.20 as per the ISO site).

edit: And since the only thing we have available is that one draft, the problems VFR maniac mentioned are still valid.

[JEEB]

Thanks for your replies.
What do they mean by "raw audio-visual data"?

What kind of raw visual data would be specified in that specification?

PCM = "raw audio" is somewhat self-explanatory. But what about "raw visual"?
Well, the draft specification (here (http://www.itscj.ipsj.or.jp/sc29/open/29view/29n13095c.htm), scroll down and click "Text" for the .doc) handles both raw audio and video streams as far as I can see.