[vorbis] Is this just anti-Ogg FUD?

gtgbr at gmx.net gtgbr at gmx.net
Fri Dec 27 01:13:36 PST 2002



Brandon High wrote:
> > First of all, I have no clue what this guy is talking about. "Pre-echo
> > aliasing", whee ... examples please. Because lateron he rants about
> If you have the correct speaker spacing, and are standing in the correct
> spot, the sound waves from multiple sources will combine and reinforce
> each other. This includes subsonics.

That's what interfering waves do, adding up or deleting each other on
different locations - depending on how out of phase they are. It's like
throwing two stones in water. This is most noticable if you wrongly
connect your speakers (one playing the audio shifted by 180°) - if you
play a test tone and move around in front of your speakers, you'll
notice different amplitudes depending on where you stand. Another
example would be an unlucky setup where the reflections of a wave
delete/interfere with the valid sound source.

But what does this have to do with something called "pre-echo aliasing"?
These problems don't even arise from lossy audio encoding in general.
Pre-Echo is what Vorbis makes sound a bit squishy on very low bitrates,
and to hear aliasing take a sinewave and pitch it up until ... uh, let's
say 200 kHz, and sample the whole thing with a sample rate of 44.1kHz.
Funny things will happen after the sinewave disappeared into inaudible
pitches for the first time.

Channel coupling may not completely fool around with phases of the
waves. If this was actually true, every song we encode with Vorbis would
have some randomly broken stereo picture, or worse, many drop-outs if
the stereo sample contains two equal channels (e.g. a mono recording
stored as a stereo .wav/.ogg). From what I hear, Vorbis does things very
correctly, and those claims are really just FUD. Unless ... someone
shows me "pre-echo aliasing" in the bass area. I'd be very interested.

<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.



More information about the Vorbis mailing list