[FFmpeg-devel] [PATCH] QCELP decoder

[Reynaldo H. Verdejo Pinochet]

Hello

Kenan Gillet wrote:
> 
> First, I want to say that I am willing to help :)

Great, thanks.

> My local subversion start at the inclusion of the SoC qcelp code at  
> r2373.
> Following this particular revision are only 2 minor revisions of  
> cosmetics
> and grammar.

Good, this two should be easy to merge.

> So we are more talking the SoC svn tree and its continuation in my  
> subversion.
> Maybe a less painful path might be to import my local commits into the
> SoC?

It could be but I'm afraid 300 commits are too
much for the real changes done, we need to find
a way to group them or just cherry pick your changes
and import them to the SoC repo in an ordered
fashion.

> Furthermore, my first intent was to implement I_F_Q and RATE_OCTAVE

RATE_OCTAVE was already implemented (mostly), although never
really tested so it might had some errors. Insufficient
frame quality handling in the other side was kind of a
left-for-latter compromise due to its two-way communication
needs. The way I saw it back then is that IFQ handling
only makes sense if you have a way to talk back to the
encoder. (like on a real phone converstion).

> let ken know your thoughts

I'm glad to have your help and if you're willing to work
with me on the SoC repo version I'm sure we can make it
happen in a short amount of time.

Bests and thanks again

--
Reynaldo