[Flac-dev] Blocking and compression.

[Josh Coalson]

--- Wayde Milas <wmilas at rarcoa.com> wrote:
> On Tue, 2004-01-20 at 23:44, Josh Coalson wrote:
> > Miroslav did some experiments with searching for optimum blocksize.
> > from what I remember it made at best a couple percent difference.
> > there was a thread about it here a while back.
> > 
> Did his changes make it in? I can think of a coupla ways to approach
> this and I'd like to hear about what he tried.

no, I think he just tried a brute force search of all blocksizes
to find the lower bound.  you may be able to find the thread by
searching archives, though the sourceforge list search is not
that good.

> > you can get very close ot arithmetic coding with range coding,
> > which is much faster and believed to be patent free.  monkey's
> > audio uses range coding as part of its entropy coder.  I did
> > some experiments and couldn't get much of a compression
> > improvement, but the area is far from exhausted.
> > 
> 
> I was not aware that arithmetic coding was patented anywhere. The
> texts
> I have on the subject don't mention a patent, although that doesn't
> surprise me. I had considered range coding, but arithmetic is
> superior
> in every was except for speed. Having a nice reference encoder that
> you
> could then go back to and compare ordered rice or range to I thought
> would be ideal.

I think it's patented by IBM.  range coding gets very close
to arithmetic and is much faster.  maybe range coding was done
earlier and later turned out to be a variation of arith coding,
with renormalization is done in larger chunks.

Josh


__________________________________
Do you Yahoo!?
Yahoo! Hotjobs: Enter the "Signing Bonus" Sweepstakes
http://hotjobs.sweepstakes.yahoo.com/signingbonus