[vorbis] Low bitrate encoding

Joerg Bergmann bergmann at rcs1.urz.tu-dresden.de
Tue Jan 9 07:41:49 PST 2001



On Tue, 9 Jan 2001 10:19:54 -0500, Gregory Maxwell wrote:

>> To get good quality downsampling (or A/Ding) you must filter first. This is
>> non-trivial especially when the rates in noninteger.
>
>To get a mental handle on this: Imagine sampling a 15KHz signal at 8KHz.
>Would you get all zeros?
>
Low-pass cannot meen: _No_ frequencies above a limit. At my feeling,
_exactly_ such a low pass would violate the signal theorem mentioned.
You can only decrease their amplitudes by a very low (but non-zero)
value, with the exception of some singular (resonance) points, in ideal.
Half the sampling rate is such a resonance point. In principle, the
discussion about a fourier-transformed weighting function (which is a serie
for an already sampled signal) in time holds.
The problem: For non-integer sampling down ratios, you need a multiple
set of weighting series. For rational ratios, you need as many series
as the denominator of the ratio. In principle, for irrational sampling
rates, each sampling serie is different from each other.

J"org Bergmann

email at jbergmann.de

--- >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