[FFmpeg-devel] [PATCH 2/3] RSD demuxer

[James Almer]

On 18/04/13 6:35 AM, Paul B Mahol wrote:
> On 4/18/13, James Almer <jamrial at gmail.com> wrote:
>>
>> Signed-off-by: James Almer <jamrial at gmail.com>
>> ---
>>  Changelog                |   1 +
>>  doc/general.texi         |   1 +
>>  libavformat/Makefile     |   1 +
>>  libavformat/allformats.c |   1 +
>>  libavformat/rsd.c        | 195
>> +++++++++++++++++++++++++++++++++++++++++++++++
>>  libavformat/version.h    |   2 +-
>>  6 files changed, 200 insertions(+), 1 deletion(-)
>>  create mode 100644 libavformat/rsd.c
> 
> Ineresting I already write something similar in my repo.
> 
> I'm not sure about GADP, as its not mono only... and I'm
> not sure its same as ADPCM_THP.

It is, it just doesn't have the previous sample stored within each frame 
(Like thp) or in a table (Like brstm). Which is why i modified the decoder
to accept frames like that.

I used the samples from http://hcs64.com/files/in_cube_samples.zip to 
test both GADP and PCMB from rsd, and http://hcs64.com/files/jingle04.rsd
and http://hcs64.com/files/jingle05.rsd for RedSpark.
I couldn't find samples of WADP (Also adpcm_thp but can be stereo) or the
other codecs, so i decided to not write code i couldn't test.

That being said, i think we can safely rename adpcm_thp into something like
adpcm_dsp, since it's not just used by thp containers. But that's purely a
cosmetic thing.

Regards.