[vorbis] OGG IS BEST, BUT UNUSABLE! (very important)]

[Michael Smith]

Monty wrote:
> 
> On Sun, Jan 28, 2001 at 04:23:48PM +0100, Aleksandar Dovnikovic wrote:
> > "Geoff Shang" <gshang at uq.net.au> wrote:
> >
> > > True.  And anyway, the VBR stuff in beta4 apparently does a better job at
> > > allocating bit rates than did beta3.
> >
> > It does on almost all test samples I have, except for the samples that
> > have lot of (potential) pre-echo events (like castanets.wav and a few
> > other that I have) - it seems that the VBR in pre-beta4 gives less
> > bits then beta3 for these events, so pre-echo is more noticeable in
> > beta4 (in comparison to beta3).
> 
> I'm checking out what you've found.  Beta 4 does limit bit allocation
> in short blocks more aggressively than beta 3 did, but the preecho
> triggering should be the same.
> 

>From a quick look at the code (diff between HEAD and the beta3 tag), it
looks like preecho triggering is changed in a fairly major way (i.e.
_ve_deltai() is more or less completely different, and this is the core
function for preecho trigger). Certainly it isn't the same.

As I recall, this is one of the bits you did take from Segher's big
patches. Perhaps the replacement code here doesn't work as well as the
older/slower code.

Michael

--- >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.