[vorbis-dev] BCB build successful?

OmegaDan omegadan at thehumbleguys.com
Wed Jan 31 02:03:14 PST 2001



heh its ok, I didn't understand rehuff untill someone explained it to me
..

Huffman compression is essentially a substitution cypher where short bit
patterns are substituted for byte patterns ... to do the substitutions you
need a translation device called a "huffman tree" (a special case of a
binary tree) ... but to make the whole thing work you have to have seen
EVERY byte in the stream so you can count what patterns there are and how
many of them there are, and a simple algorithim can build an optimal tree
from that data.  

-- what threw me -- was -- during streaming you don't know what data is
coming next -- so you can't generate an optimal tree -- so they've come up
with "codebooks" which (I haven't looked at but) must contain guesses at
pattern frequencies to build a tree from ...

So when all was said and done, you weren't using an optimal tree but a
aproximation of one based on historical data.  Where rehuff comes in is it
decompresess the data, generates an optimal tree for it, and recompressess
it (lossless), and saves you a few extra percent space.

IMHO rehuff should be a standard part of the encoding process :)

So if the binaries you had have different codebooks, it would output
different files which both should reduce (rehuff) down to a 3rd identical
file ... though I suspect its different sources at this point?

OD

On Tue, 30 Jan 2001, David K. Gasaway wrote:
> Date: Tue, 30 Jan 2001 23:00:36 -0800
> To: vorbis-dev at xiph.org
> From: "David K. Gasaway" <dgasaway at bigvalley.net>
> Reply-To: vorbis-dev at xiph.org
> Sender: owner-vorbis-dev at xiph.org
> Subject: Re: [vorbis-dev] BCB build successful?
> 
> > At this point I have to ask -- are you absoultley sure these are the
> > same exact sources doing this ?  I wonder if they have different
> > codebooks? That could be tested with the rehuff util Segher
> > Boessenkool wrote -- theoretically when optimized they would be
> > identical if the codebooks were different?. 
> 
> No, I don't know if it's the same sources; that's the problem.  I 
> downloaded the other version from a link that someone posted in 
> the list.  I don't remember who or exactly when.  Although I've read 
> all the messages about rehuff, I haven't gathered what it does.  In 
> other words: I don't know what you're talking about :)
> 
> > OD
> 
> > F:\oggs>oggenc-bcb "125bpm 8 bar main line 1 FX.wav"
> >         Average bitrate: 121.0 kb/s
> > 
> > F:\oggs>oggenc-msvc "125bpm 8 bar main line 1 FX.wav"
> >         Average bitrate: 121.8 kb/s
> 
> Ah, same deal.  Thanks for your help.
> -----
> David K. Gasaway
> dgasaway at comports.com
> http://dkg.home.dhs.org
> -----
> Unless you're the lead dog, the view never changes.
> -----
> Sig by Kookie Jar 5.98b http://go.to/generalfrenetics/
> 30 January 2001
> 
> --- >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-dev-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.

--- >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-dev-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.



More information about the Vorbis-dev mailing list