[vorbis] I hate myself for asking this, but...

Gian-Carlo Pascutto gcp at sjeng.org
Sun Sep 9 06:30:38 PDT 2001

On Sun, 9 Sep 2001, elfin wrote:

> > MPC is another matter. For what it does (VBR MP2-alike at 170kbps)
> > it is extremely good and without a doubt one of the, if not the
> > highest quality encoder available today.
> Yes I know it is, but it's not so free.
> Where is the secret of it's quality?

It built on something that was already good (MP2) and it
was focussed to a single specific problem. It also didn't
have to care about backwards compatibility.

It's main problem is that it's pretty much reached the end
of its possibilities. The author said it would be very hard
to improve quality even more.

> > I would hope that Vorbis will offer MPC-like quality at 160kbps
> > in the future, but I don't think it'll happen anytime soon.
> Why? Where is the problem?

If you try to focus on sounding good from 32kbps to 350kbps,
you are not going to progress as fast as if you focussed on
160kbps only.

One of the main problems Vorbis has right now (when trying
to reach MPC-like quality) is preecho. Vorbis will probably
not be as good as MPC in that regard untill after 1.0 comes

> > > Are there gonna be any "advanced user options" added to Vorbis?
> > Monty has said he would add something like this, although no
> > details are known.
> I'd suggest some options like in LAME
> -lowpass, -b (minimum bitrate)

Monty has said he would allow the users to pick the lowpass,
and that bitrate constraining would be added, so i think this
is already accounted for.

> or maybe some like MPC has
> -tmn, -nmt, -minSMR (different approach, you don't choose bitrate, but how
> much masking or noise is allowed)
> (You know, what I mean - something for us freaks to play with)
> :-)

You can do this already by selecting a different bitrate (internally
Vorbis doesn't use bitrate to determine what to encode, but similar
metrics to what you describe, and the bitrate actually toggles between
different presets), although it would be nicer to be  able to specify
this more finely of course. I guess that will be added too in the future.


--- >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.

More information about the Vorbis mailing list