> 
> > In particular,
> > codecs have their own implementations of endian conversion, bit reading,
> > windowing functions, huffman decoding, etc. and I'll change some of
> > these to use a common implementation in codeclib.
> 
> Isn't that as it is because we've imported codec code from other 
> projects? And don't we strive to do as little changes to that code as 
> possible to make upgrading easier?



No, its because no one has fixed that.  FWIW I don't think there are any codecs 
we're likely to upgrade.  We're pretty much the only people working on fixed 
point open source codecs. 

Mike

                                          

Reply via email to