[M4IF Technotes] Is standard conform to put mp3 sound in aMPEG-4 stream ?

[DaveMackie]

In ISO/IEC 14496-1:2001, Section 8.6.7.2 DecoderSpecificInfo Semantics 
(pp 31-32), I see
"For values of DecoderConfigDescriptor.objectTypeIndication that refer 
to streams complying with ISO/IEC 13818-3 the decoder specific 
information is empty since all necessary data is in the bitstream 
itself. The access units in this case are the frame() bitstream element 
as is defined in ISO/IEC 11172-3."
There is no specific mention of the case of 11172-3, but it would seem 
to me bizare to apply different rules for 11172-3 and 13818-3.
Dave Mackie
Cisco Systems
AVARO Olivier FTRD/DIH/REN wrote:
>  
>
>
>     Well yes, it does mean that QT6 doesn't do layer 3 in MP4.  QT6
>     supports a few profiles in MP4, and we currently don't cover the
>     'old' formats (MPEG1 and MPEG2) in an MP4 context.
>
>     I'd be curious to know, however, if we have the same understanding
>     of the structures which would say "this is layer3 audio".  can you
>     briefly summarize your understanding of the profile, level,
>     streamtype, objecttype,  decoderspecificinfo, and access unit format? 
>      
>     Since MP3 is a spec. that MPEG *owns*, I think it will be a very
>     valuable contribution to clarify all this in the MPEG spec.
>      
>     I think streamtype and objecttype are very clear :
>     - streamtype shall be 0x05 (this is an AudiioStream)
>     - objecttype shall be 0x6B (this is Audio ISO/IEC 11172-3 MPEG-1
>     Audio)
>     (Maybe we could have more depending on P&L of ISO/IEC 11172-3).
>      
>     For decoderspecficinfo, although we could make something
>     elaborate, I guess we can also leave it empty and rely on the
>     *in-band* signaling of MP3.This needs further thought however
>     (such as to investigate if P&L for 11172-3  are properly signalled
>     in MP3 files).
>      
>     Concerning Access Unit definition, yes, this needs to be clarified
>     and findthe could trade-off between being consistent with the
>     MPEG-4 framework and minimize the transformation of MP3 files.
>      
>     As far as MPEG-4 Audio Profile and Level are concerned, I concur
>     with Rob, if there is a demand for MPEG-4 P&L supporting MP3, this
>     can be done. This could also be done in industry fora like ISMA
>     (ex : "we support this and this MPEG-4 P&L + MP3").
>      
>     Kind regards,
>      
>     Olivier
>