From Gmarsh4719@aol.com Tue Nov 09 13:32:48 2004
Subject:Re: Why did Hammond get the virtual B3 right?


Dean,

Just to clarify (an already mish-moshy situation) one would not
sample the 9 drawbars... they do not in and of themselves make any
sound. One samples all 96 tone wheels. These 96 wheels (sample loops)
are sufficient to create every harmonic of every key of the manuals
and pedal tones.(Subject to foldback, of course) The drawbars and the
9key contact system (and the matching transformer, if you will)
combines all this spectral goodness into the composite sound that
each key makes. Therefore, a tonewheel sample set would consist of 96
loops. One of the contentions in play here is A> is it a single AC
cycle of the sine-like waveform? (which would be somewhat sterile) or
B> a longish loop which would include enough AC cycles to encompass a
full revolution or more of the wheel with respect to the pickup head?
This would allow the "wow and flutter" , crosstalk, motor hum, etc to
be a part of the waveform, and if well sampled and correctly trimmed
and looped, would seem to be preferrable to the single cycle loop.

In modelling, one can say (in software) OK, what we have is a sine
wave with the following nonlinearities in it, with the following
modulations on it (the aforementioned wow and flutter, hum,
crosstalk, etc.) and you have your raw single tonewheel sound. This
must still be combined in the same way. From here on out, everything
else on both types of algorythms is really modeling (amp model,
churus, perc, distortion, leslie, etc.) Not sampling.

Greg

>
> Dean
>
> [Non-text portions of this message have been removed]