Emax

--- In emax@yahoogroups.com, "Rob Keeble" <rob@e...> wrote:
> channels of digital audio. The chip performs three main jobs:
> 
> 1)   It replays the samples from memory and pitch shifts them as 
necessary
> using a fixed sample replay rate that adds and drops samples to 
achieve the
> transposition in pitch. It understands the sample rate of the 
sample.

Well, that would account for the distortion.  From what little I 
know, adding/dropping samples is a poor method of resampling.  With a 
single multiplier op, one can do at least linear interpolation where 
samples will sound smoother.  I'm not complaining, I love the pitch-
down/up artifacts on my SP.  I imagine emax owners love their 
artifacts too.


> The Akai S950 (variable rate) and Emax have very different frequency
> responses, with the Akai having 5dB boost from 3.5kHz to 21kHz, 
whilst the
> Emax drops 9dB at 18kHz. 

I didn't know that!  Why does akai do that?  Is their input filter 
this bad?  The Emax spec makes more sense to me.


> the Emax distortion is due to the compression/expansion of samples 
more than
> shifting, which also affects dynamic range. The E chip may also be 
over
> sampling, but the precise design is confidential.

Ok, so the resampling/pitch shifting is not the culprit!?

So that may be the difference between the SP and the Emax.  At zero 
pitch shift on the SP, the sample exhibits no distortion of the kind 
heard when pitch-shifting.  I would have imagined that the Emax was 
the same; only introducing artifacts when pitch-shifting.  If the 
compression algo does this much damage, I would guess that the Emax 
plays back with distortion even at the root key.  Is that true?

BTW Rob, thanks for the indepth responses.  These are super 
interesting!