[vorbis] Is this just anti-Ogg FUD?

[gtgbr at gmx.net]

Jem wrote:
> Ogg has to this day, fundemental problems with low frequency encoding.
> Sometimes refered to as "pre-echo aliasing" ... it is something that is a
> problem for music (primarly electronic which is why I don't use it) with
> low end bass and subsonics. If you have the ability to test this on a
> proper sound system, you will hear it dead out. Most people don't realise
> that you need a certain amount of distance from the generation source to
> your ear for the bass wave to resonate correctly. This is why sometimes
> when you are at a rave and you could be 10 feet from the bass bins and
> there is very little bass, but you step back 20 feet and its knocking you
> on your ass. This same principle holds with headphones. You are not
> typically going to get the "subsonics" out of it listening to the files on
> headphones. It may sound slightly out of tune, or out of phase. Good head
> phone definitely help, but even high end sony's like the MDR-7506's and the
> MDJ-700s are only just able to produce the subsonics effectively to hear
> it.

First of all, I have no clue what this guy is talking about. "Pre-echo
aliasing", whee ... examples please. Because lateron he rants about
subsonics getting lost, which, are not audible/easily masked by any
other sounds anyways (and this has nothing to do with artifacts). A
lossy/perceptual encoder can safely discard anything below 20Hz, so
what's the problem? It should do that by design - throw away useless
information.

By the way, by doing so, Vorbis (and any other reasonable lossy encoder)
does bad producers a favor. Filtering out subsonic rumblings that do
nothing but randomly shift the DC offset of a wave (please excuse my
probably wrong choice of words) and thus distort the whole thing
*should* be filtered out by the producer in the first place.

If he talks about audible bass that actually makes sense in music, I
really can't relate - this sounds like some subjective bias against
Vorbis for one reason or the other. Blind tests with hidden references
would be useful, and if he can prove that he hears some obvious and
important bug at a reasonably high bitrate, his feedback could be
valuable.

> [Ogg is too little, too late.] There are audio and video encoding systems
> coming out in the next year that will futher push the envelope without
> comprimise and many of those are based completely on wavelets which is
> really where Ogg should have gone in the first place ...

Afair, wavelets in audio compression are 99% buzzword and 1%
improvement. They ARE useful, but the result of previous discussions
were that they're simply not worth the effort. Vorbis I is already
revolutionary enough - it's good enough to compete with other Codecs for
quite a long time (until we get Vorbis II and major buttkicking, even
with wavelets :) ).

<p>Moritz
--- >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.