[vorbis] Bit Rate Peeling Quality

[Graham Mitchell]

> I'm using oggenc on the assumption that it's state of the art (am
> I wrong?).

You are not.  There is no better encoder in terms of quality than oggenc.  
However, vorbis is still a (relatively) young codec, and there's no saying 
that future encoders won't offer even higher quality for a given bitrate.

> So following on from your answer (and also that given by Daniel):  Is
> there a quality higher than 2 with which c.ogg could be encoded to yeild
> it as good as d.wav, or is it just that quality doesn't (currently)
> scale in this way?

There's no magic number you can peel to that's guaranteed to match the 
equivalent pure-encoded quality.  Quality numbers are merely a reference 
point.

>> Anyway, it's assumed that eventually the (not-yet-written) code from the
>> [MAGIC] phase would be incorporated into the stock encoder.
>
> And, presumably, the into peeler for use when processing older files.

Maybe not.  The peeler probably won't, in general, be able to tell if you're 
feeding it a "peeling-optimized" ogg or not.  Plus, the peeling from, say, q6 
to q2 is intended to be much faster than decoding the q6 file to wav and then 
re-encoding to q2, and probably provide better quality to boot.  Adding extra 
"lossless re-encoding" steps to the peeler slows it down.

Of course, since tools like OggDrop exist, there's certainly a good chance 
that *someone* will make an all-in-one "losslessly re-encode, then peel" 
tool, but I doubt if it will be xiph.org.

> On the input side I have... [sources] which will never again be available.

(*cough*)

> So it seems to me that my best path is to encode now at q6, then in
> the future reencode available sources at q2 while peeling my q6s from
> no longer available sources at q2 (or maybe q3 if this would compare
> better with native q2 encoding).
>
> Sound reasonable?

Yes, it does.  I think it's possible that once real peelers appear, they may 
be able to peel, say a q6 file down to q2 and have the quality be 
indistinguishable from a "pure" q2 file.  But, then again, they may not.  
There may be a slightly lower quality relative to the "pure" q2 file.  I 
think at this point, no one really knows for sure.

Obviously everybody believes the quality of peeled files will compare well 
with natively-encoded files, or they wouldn't even mention peeling as a 
feature.

And finally, remember that quality values aren't integers.  You could 
conceivably peel a q6 to a q2.3 if you wanted.  At least, I think so.  Not 
sure if you'll be able to peel to non-integer qualities, but maybe so.

Quality 2.3 only takes up 5% more space (on my single test file) than quality 
2, and improved the bitrate from 82.6 kbps to 86.9.


-- 
Graham Mitchell - computer science teacher, Leander High School
"The music business is a cruel and shallow money trench, a long plastic
hallway, where thieves and pimps run free and good men die like dogs. There's
also a negative side."
	-- Hunter S. Thompson

--- >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.