[vorbis-dev] DWIT

Segher Boessenkool segher at wanadoo.nl
Sun Jan 14 14:02:08 PST 2001



Philipp Kempgen wrote:
> 
> Steve underwood wrote:
> 
> > Segher Boessenkool wrote:
> >
> > > We can't hear phase differences above 2kHz.
> >
> > Actually I think we both forgot the biggie - "well".  We can't  detect
> > bass direction well.
> 
> So what about an option for filtering frequencies below 10 Hz?

This is generally a bad idea. Sometimes, removing very low frequencies
causes modulation effects (or removes them), so we would need to detect
if this would happen; not a trivial task. And, more importantly, removing
the very low freqs causes clipping at playback (the low freqs have large
amplitudes). The savings in bits is not that big, btw. (Ignoring phase
differences in >2kHz saves about 40% of the bits of a stereo signal,
on most audio streams).

> And in 'joined stereo' mode (yet to be done) we could simply leave out any information about frequencies below 40 Hz and above 16kHz or so on the second channel - thus making the decoder play back these very low resp. high freqs in 'mono'.

Phase difference on <40Hz is very detectable. Amplitude difference on
high frequencies is very hearable.

To really exploit stereo redundancy, we'll need to do a lot of
(literature) research. The normally-used methods are not "good
enough" and full of patent minefields as well.

Dagdag,

Segher

p.s. Anyone who knows about good stereo masking/unmasking literature:
please mail me pointers and/or pdf's :-)

--- >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-dev-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-dev mailing list