From c_schonberger@yahoo.com Tue Dec 11 14:59:12 2012
Subject:Re: Korg CX-3 analog vs digital?

Well I think your idea is great, but there are so many more parameters to be considered. The most logical thing is addressing these in order of relevance. One of the first things after the correct gear ratios (or call it Hammond-tempered scale) is the correct waveform. It is supposed to be a sine wave, but it isn't. The specific signal path and electronics involved cause some form of distortion and of

course there is tonewheel leakage, which is not random, it follows a pattern due to proximity of certain tonewheels. Now about the key- and release click: is needs to be authentic and at least a little random (somethings some clones don't do and it sounds bad, only kind of made acceptable by a leslie or sim). Also: the pitch istn's stable, so I would add a smooth random LFO to modulate the pitch of each tonewheel.
 
It all comes down to: the only technology to faithfully emulate a vintage Hammond would be physical modeling with high resolution, extremely low latency and with all relevant parameters represented as faithfully as possible.
 
Sampling tonewheels is a good idea, but the sound already goes through electronics. So adding tonewheels (in a drawbar setting and/or playing chords) in a digital environment doesn't sound like adding the same tonewheels through a Hammond circuit. The question is: how relevant is it to the overall sound?
 
I guess, since we all know the details and parameters of a Hammond sound, the only question remains: which ones are the most relevant and how do I represent them all convincingly with different technology?
 
Christian

________________________________
From: donstavely
To: CloneWheel@yahoogroups.com
Sent: Tuesday, December 11, 2012 10:22 PM
Subject: Re: [CWSG] Korg CX-3 analog vs digital?

 
Thanks for the replies.

I agree that tone generation is key to a faithful Hammond sound. The idea of "harmonic beating" needs some explanation. The octave relationships of 16', 8', 4', 2', and 1' and C to C to C are perfect factors of two, since they are generated with tonewheels with 2, 4, 8, 16, 32, ... teeth.

The hard part is the other drawbars and the chromatic notes of the scale. They are generated with gear ratios that approximate a well-tempered scale. A well-tempered tuning generates unique "beating" for intervals other than octaves, since the intervals aren't perfect simple ratios of frequencies (as in just or pure intonation). The Hammond ratios are close enough that notes sound in tune, but intervals will "beat" in a slightly different way. I think this is a unique contribution to the Hammond sound.

Top Octave Generator chips used in early synths and clones also approximate a well-tempered scale by using divide-by-N circuits. The TOG ratios between chromatic notes are also close, but different, so intervals will "beat" in a way that is different than a real Hammond. Modern clone processors are fast enough to get the frequencies closer to Hammond's.

Believing that this distinction is important, my approach is to exactly replicate the gear ratios that Hammond used. I start with a master oscillator, which equates to the main drive motor shaft. I then multiply and divide this signal by the exact gear ratios that Hammond used in his driving- and driven-gear sets. So I get tone frequencies that replicate his exactly.

Sorry to be long-winded, but I wanted to explain this one reason why I think a modern analog tonewheel organ can do better than the first clonewheels, and maybe compete against modern digital clonewheels.

Don

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