[Vorbis-dev] Ogg vorbis decoder - Independent implementation questions

[John Ripley]

Vinoth Kumar wrote:
> Hi All,
>  
>  Ours is an embedded software solutions company that focuses on 
> developing embedded software solutions specific to the DSP market. I'm 
> happy to inform the ogg vorbis developer community that today we have 
> release trial version of our Ogg vorbis decoder made available on PC, 
> TMS320C67x platform (More platforms to follow). More information about 
> the same can be fetched from
>  
>  www.vinjey.com/ogg.html <http://www.vinjey.com/ogg.html>
>  
>  Now there is couple of questions for which we are looking for the answers.
>  
> a. What are the set of conditions we need to oblige before using the 
> Xiphophorus and Ogg Vorbis logo on our website?
> 
> b. Ours is an independent implementation of the Ogg vorbis decoder. We 
> are interested in knowing about whether there are any other independent 
> implementations of ogg vorbis decoder available?

I've written one, although the only version I have up on a web page is 
rather old (Jan 2004!) and not the nice tidy version it's become since 
then (honest!) It's at http://www.pslam.demon.co.uk/specbis-0.05.tar.gz

I wrote it to test out the viability of strictly controlled memory 
management, and to test out some ideas for encode/decode more easily 
than the reference codebase could. It's also a bitrate peeler.

What method have you used to implement bitrate peeling? The method I 
used was to patch the encoder to force "low" and "high" quality residue 
into passes 0-2 and 3-7 respectively. This generates a file slightly 
larger than just encoding at "high" without peeling support. The peeler 
can then go through the bitstream and throw away passes 3-7. This 
results in a file roughly the same size as it would be when encoded at 
"low". It's extremely fast and works surprisingly well, but it's a very 
brutal and unpolished method and the sound quality is barely passable.

I take it there's no freely available source code?

- John Ripley.