[Ffmpeg-devel] Re: [PATCH] non conformance to 3GP format (3GP TS 26.244 v 6.5.0 2006/06) 2nd try

[Baptiste Coudurier]

Hi

Michael Niedermayer wrote:
>> [...]
>>
>> Michael, do you think we should keep using AVI fourcc ? The more I read
>> specs and the more I feel we should not do that and strictly mux
>> according to specs.
> 
> hmm, where do we use avi fourcc in mov/mp4 except if theres no mov fourcc
> maybe?

Yes when there is no mov fourcc. Personnaly I would prevent using them
except manually when they use vtag.

>> [...]
>> -    }
>> +    /* FIXME not sure, ISO 14496-12 specifies either 1 or 2, no samples found where channels > 2 */
>> +    if (track->mode == MODE_MOV || (track->mode == MODE_MP4 && track->enc->channels < 2))
>> +        put_be16(pb, track->enc->channels);
>> +    else
>> +        put_be16(pb, 2); /* Reserved */
> 
> does the spec just not specify how to store streams with more then 2 channels
> or does it explicitly say that >2 channels audio should have 2 in the channels
> field?


Quoting ISO 14496-12:
"template unsigned int(16) channelcount = 2;"

"ChannelCount is either 1 (mono) or 2 (stereo)"

Would mean 2 is default value.

Like I said I'm not sure. Im looking for ISO 14496-3 AMD 3 which
specifies mp3on4, mp3 5:1. Maybe they override specs for mp4, maybe not.

All mp3on4 5:1, vorbis 5:1, aac 5:1 samples I could find have 2 for
channelcount, while mov with 5:1 from apple have 6 as channelcount.

> [...]
> 
> ive no objections to this but IMO this should be split in 2 commits
> 1. adding the  if (track->mode == MODE_MOV) { ... } else { ... }
> 2. reindenting the code within the new if()
> it would be much more readable on cvslog that way
> 
> [...]

Sure. I'll do that.

-- 
Baptiste COUDURIER                              GnuPG Key Id: 0x5C1ABAAA
SMARTJOG S.A.                                    http://www.smartjog.com
Key fingerprint                 8D77134D20CC9220201FC5DB0AC9325C5C1ABAAA
Phone: +33 1 49966312