[vorbis] Spectrum duplication (2)

Sune Foldager cryo at diku.dk
Mon Jun 16 17:55:13 PDT 2003



Headless wrote:

>> Speaking of which, would there be anything to stop Vorbis, maybe not 
>> v1
>> but in the future, to use similar techniques as the ones emplyoed by
>> mp3pro and the similar thing for aac? I am refering to the spectrum
>> duplication..
>
> I'm no expert, but imo the technique used by MP3Pro is a kludge to
> maintain backward compatibility with regular MP3.

Yes, the implementation is done to remain backward compatible. This, 
however, is hardly related to the technique itself, which is 
duplicating part of the spectrum using some auxilary data.

> Ogg doesn't need that sort of nonsense.

The backward compatibility stuff, no. But a technique that can maybe 
improve percieved quality? Would make sense to consider it at least. 
IMHO.

>> even though it sounds a bit weird to me, that it can work
>> well in all circumstances; but mp3pro does sound significantly better
>> than regular mp3.
>
> Only at low bitrates, MP3 isn't capable of competing with for example
> WMA at low bitrates. MP3Pro was divised to address that, but because
> they wanted it to be backward compatible with MP3 they had to
> implement it in that kludge type way.
>
> There is nothing to "fix" for MP3Pro at higher bitrates, which is why
> proper MP3Pro encoders only allow encoding up to a certain bitrate.

Right.. but none the less, they made AACpro or whatever it's called, 
which apparently employes the same "technology", viz. the spectrum 
duplication stuff, to improve on AAC's quality in low bitrates. Since 
aac is pretty comparable to vorbis (ogg is still the container format 
right? ;-) ), it makes sense to ask if a similar technique could be 
applied to the latter and thus be able to put more bits into a smaler 
part of the spectrum, relying on the duplication.. weird as it sounds.


-- 
Sune.

--- >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