[vorbis] Bzip2 & Ogg Vorbis

[Ed Sweetman]

Shawn Riley wrote:
> John wrote
> 
>>>wrote a rough, hackerware tool to recalculate the optimal codes on
>>>an 
>>>existing file, and I think they got results similar to yours
>>>(though maybe 
>>>bzip2 does a bit better). 
>>
> HJ wrote-
> 
>>Segher Bossenkool (sp?) did, but Segher's been kinda quiet recently.
>>Search the archives for 'rehuff', or thereabouts.  Work has
>>presumably stopped or gone back underground for further tweakage
>>amidst other Dark Arts.  :D
> 
> 
> 
> Last I remember, he said it broke the decoders because of a bug in libvorbis or libogg. (?!) Something to do with seeking. So you end up with the same problem as when you use bzip2, & if you've got that working already.
> Something to look out for with non-seekable formats- I remember using RKA format to compress PCM audio on Windows. The decoder didn't allow seeking. And at one stage, the back half of the file was unusable, probably because it had a few bytes corrupted in the middle. It wouldn't seek past it, & It refused to read through it. So if bzip2 & rehuff have the same issue, I wouldn't use either for streaming.
> 
> - Shawn

<p>if you're streaming bzip2 compressed data, wouldn't that mean that each 
packet has to be compressed on it's own as it is sent? And if that's how 
it works, wouldn't that decrease the effectiveness of using bzip2 
because the compression is dependent on the other data in the file, if 
you use smaller samples it's less likely you'll find repetitive parts of 
data.

If that's not how it works then ignore this.

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.