[vorbis] Ogg Vorbis files can be compressed ?!?

[Corey Miller]

Hmmm... would it be all that difficult to implement some sort of system 
to find redundancy of blocks.  I know it seems like only 0.001% of all 
vorbis files can be compressed due to this.  But that's with lossless 
compression methods.  I'm sure many files have blocks that are audibly 
similar enough so you could use the same data for all of them.  It 
could be improved even further if the threshold of similarity to be 
considered redundant could be controlled.  Or if it also checked prior 
to compression, because maybe a piece of uncompressed audio is similar 
enough to another that when they are compressed to a certain level they 
can use the same block of audio.  If anything like this were 
implemented, it should obviously be optional.  Of course just another 
option in way a vorbis file can be compressed just adds more to the 
header.  I don't know how much.  I would probably look through the 
source and try to implement this right now if I were sure it was a good 
idea, but I have no idea, lol.

On Monday, January 20, 2003, at 03:01 PM, Peter Harris wrote:

> Nescafe wrote:
>> My final questions is :
>> What is to blame for this sample ?
>> Redundancy, or maybe window size, or what ever ?
>
> I'd guess redundancy.
>
> And, yes, MP3 and Vorbis do use different block sizes. Maybe that 
> accounts for MP3 not being recompressable with .zip (pure dumb luck in 
> where the blocks line up with your simialer sample)?
>
>> Shouldn't Vorbis take this redundancy into account ?
>
> No. That would prevent (or at least horribly complicate) both seeking 
> and streaming. Besides, it doesn't help on 99+% of the samples out 
> there.

--- >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.