Home Digital Drive

Upsamplers, DACs, jitter, shakes and analogue withdrawals, this is it.

You say 'potato', I say 'potahto'

Todd Krieger wrote:

> Extrapolation is predicting future values based on
> existing data. (Stock analysts use "extrapolation" to
> predict the market and various stocks.) It differs from
> interpolation which fills in values *between* existing
> data points.

My sense of the word "extrapolation" is a more general one
of "placing or going beyond what's directly known or
believed with confidence". In the time domain, going
beyond the present means projecting into the future, as
you point out. However, I don't think the sense of the
word is restricted to time. In the audio case under
discussion, we're going beyond the 16 bits/sample we
have direct knowledge of (from the CD) to 24 bits/sample
that we're projecting, guessing at, estimating, modelling,
what have you. I'd call it "extrapolation". Your
favorite lexicographer may disagree, of course, and I
don't have the _Oxford English Dictionary_ handy. You can
do the lexicographical/etymological research yourself.

Also, AFAIK, the Audio Alchemy/Perpetual "resolution
enhancement" algorithm is operating strictly in the
amplitude domain (by adding bits to existing samples),
not in the time domain (adding additional samples
between the existing ones). Though, in the case of the
Perpetual P-1A, the issue is confused by the fact that
an asynchronous SRC (the CS8420 chip) is adding additional
samples, too. Though the upsampling can be bypassed
by using the P-1A's I2S input (as I'm doing).

Jim F.



This post is made possible by the generous support of people like you and our sponsors:
  VH Audio  


Follow Ups Full Thread
Follow Ups


You can not post to an archived thread.