[vorbis] bitpeeler

[Chris]

How does turn on use of optimal huffman tress?
would be a nice option for those who are just ripping CDs.

Chris

Graham Mitchell wrote:

>>>Yes.  But libvorbis-1.0 encoded streams are not very nicely peelable --
>>>that's why I losslessly reencode them, first.
>>>      
>>>
>>Am I correct in guessing that you are referring to re-huff'ing them?
>>    
>>
>
>Both are lossless processes, but I don't think they are the same.  Rehuffing 
>refers to replacing the generic huffman trees used for a given file with an 
>optimal which is specific to that file.  This ends up reducing the file size 
>a few percent, but doesn't provide any hinting for peelers.
>
>The reason the optimal huffman trees aren't used by default is because doing 
>so requires the ability to scan through the entirety of the data stream 
>before starting encoding.  This would make it impossible to encode 
>on-the-fly, which is a major goal for vorbis.
>
>The reason hinting for peelers isn't done yet is because a) it's hard, b) up 
>until recently it hasn't been anywhere near the front of the priority queue, 
>and c) no one wants it bad enough to pay for development, so other things are 
>being done first.
>
>Optimal huffman codes will never be used by default, though there's been talk 
>of adding an option to oggenc to enable a pre-encoding pass over the data to 
>create them if you've got it all in advance.  Not sure if this is really 
>planned anytime soon, but it's been mentioned.
>
>Peeler hinting will eventually be done, but since it may be more 
>computationally-intensive than just "normal" encoding it may not be the 
>default.
>
>  
>

<p>--- >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.