[FFmpeg-soc] [soc]: r2021 - in?eac3:?ac3dec.h?ac3dec_data.c?ac3dec_data.h eac3dec.c

[Justin Ruggles]

Bartlomiej Wolowiec wrote:
> 2008/3/26, Michael Niedermayer <michaelni at gmx.at>:
>>  I do not like generic code full of
>>  If(codecA)
>>  else
>>
>>  The sync function should return if it has a complete frame or if (maybe) not.
>>  That is
>>
>>  flags= FRAME_START | FRAME_END;
>>
>>
>>  You do this in stream_type but this name is totally wrong this is not a
>>  stream_type. It is not even constant for a stream. NEVER even think of using
>>  a name from the a52* spec, the people who wrote it are idiots.
>>  Its a flag indicating the start/end of a frame. And IMHO it should be
>>  returned more directly than as a context variable but thats rather minor
>>  the name is a major issue. It totally confused me until looked it up in the
>>  spec.
>>
>>  Code should be understandable without looking things up in the specs.
>>
>>  Also the "stream_type" has the wrong type (should be enum) and the types used
>>  in AAC-AC3 code must be generic types not *AC3 specific types in headers which
>>  arent #included.
> 
> Ok, what do you think about such a plan:

Also keep in mind (in case you are not already aware) that a normal AC3
frame can be followed by a dependent E-AC3 frame, usually containing the
7th and 8th channels of a 7.1 stream.  This is supposedly the only way
to carry a Dolby Digital 7.1 stream on Blu-ray.  I have not found any
samples though.

-Justin