[Flac-dev] Decoding Type=Independent

[Josh Coalson]

--- MVallevand <mvallevand at gmail.com> wrote:
> Hello, I am new to the list, to FLAC and to FLAC development but I am
> having a problem in trying to make FLAC work with MediaMVP based the
> mvpmc project (ref.  http://www.wvpmc.org) that hopefully someone can
> help me with.  This device is possibly minimum spec. for a FLAC
> decoder.
> 
> In any case, I have had some success cross-compiling libFLAC 
> (without
> the ogg or metadata modules).for the embedded environment upon which
> mvpmc is based. In fact according to top the CPU utilization for FLAC
> decoding and playing is only 10%.  I was very please and was going to
> release a beta release to our developers when I noticed that the code
> I use (based on the test_libFLAC encoders.c program and flac123.c)
> would not play "type =INDEPENDENT" files.
> 
> The way the MediaMVP device works is there is an onboard audio
> processor which can decode PCM (in either OggVorbis or MS WAV
> formats)
> , various MPEG, and AC3.  Basically for the LPC type I just write()
> the date to the hardware file descriptor and it seems to play fine.
> When the INDEPENDENT data is written this way, I can just make out
> the song through the noise and static and the speed is reduced.
> 
> Right now I've spent a lot of time trying to playing with blocksize,
> samples, and channels but I can't seem to get these files to work.
> Can anyone point me to sample code (no floating point on this device)
> which would explain how to transform the INDEPENDENT sample frames 
> to LPC.  Obviously preserving lossless is important for this logic
too.
> 
> Thanks in advance.

I'm not sure I understand... I think the INDEPENDENT you are referring
to means the interchannel decorrelation method.  this stage is before
and independent of LPC analysis.

for stereo data, the encoder chooses whether to encoder the L R
channels as is, or to compute mid (M) and side (S) channels and
send M/S L/S or R/S.

after that, each of the 2 channels can be encoded with LPC or
other methods.  e.g. 1st channel LPC, 2nd channel verbatim

I'm not sure why "independent" frames would not decode correctly
though.  more info on exactly how it is not working would help.

Josh


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com