The Logic Off Topic list

Yeah, but analog-modeled synths aren't samples. A basic square wave or 
saw wave is created mathematically (that means with numbers to start 
with), then run through some sort of d/a converter to create sound. It 
never started out as a sample. 

I have samples of analog synths on my Kurzweil K2vx - prophets, 
memorymoog, etc. And they don't sound anywhere near as alive as my Nova 
with its digitally recreated analog.]]]



>>> As Logic baby implied above, I'd say that the biggest difference between
>>> analogue and analogue modelling in digital synths is largely due to the 
'non
>>> linearities' that creep into the system in the various stages of a real
>>> analogue
>>> synth.
>
>
>No, the biggest difference is that digital is a sampled-time system, which
>means that it's subject to time aliasing (usually just called aliasing).
>It's truly a challenge to  make something sound close to it's analog
>counterpart while sampling at just above the hearing threshold (<24KHz
>bandwidth in a 48KHz system).  I was just reading an AES paper on how a
>decent peak limiter would need a 6MHz sample rate to keep aliasing noise to
>inaudible levels....  these days I guess that's a practicality, what with
>2GHz Pentiums....