(I may not be the best person to explain this, but I'll tell you what I've
been told, or at least, what I've been 'led to understand.')
Yes, the oscillators on the DX7, being programmable digital oscillators and
therefore quite "flexible" in their design, if you want to think of it that
way, function in a "through-zero" manner. Because in digital, it's just
"math." In analog, you actually have to design circuitry to do that kind of
thing.
Analog oscillators tend to work by charging and discharging a capacitor,
giving a sawtooth behavior; all other waveforms are derived from that
sawtooth with waveshaping circuitry. So they leverage the physical behavior
of a physical component. If the input voltage goes "negative," it doesn't
invert the sawtooth--the capacitor discharges the way that capacitors
discharge, you know? It doesn't "reverse its behavior in time."
Some--a ∗very∗ few--"through-zero" analog VCOs have been designed for this
purpose. I was sure that Mr. Haible had something about this on his web site
but I can't seem to find it today. But the fact that the MOTM VCOs aren't
through-zero is not a design deficiency at all; through-zero behavior is
very special purpose. The Moog, ARP, etc. VCOs were never through zero
either. That kind of FM is really a child of the digital age, because it's
simple (cough, choke) number juggling. But of course, it CAN be done with
analog electronics; it's just that it's a unique design problem. I don't
believe that any commercial analog VCO manufacturer ever produced a TZ-VCO.
At least, I've never heard about it. (Unless Big D has leveraged JH's
designs or something...)
There was some discussion about this on the list two years ago or so, and it
seemed that there was some mild interest in a TZ-VCO. I don't know if it
will ever really make it on to Paul's agenda though. The only thing that the
TZ behavior is good for is this flavor of FM, and since the VCO would
certainly be more expensive, it becomes a real niche thing. Personally, I'd
love to play with one and see what it's all about.
As a side note, while the DX7 used only sine oscillators, some later
implementations such as the Yamaha TX-81Z had eight (I think) waveforms.
Sine plus seven others. They tended to still be "rounded" waveforms, I have
to believe selected after a lot of experimentation, but to look at them,
they were in great part sine waves that were rectified in various ways.
Depending on how they were used, they did lend coloration to the sound. I
think Yamaha was trying to get more harmonic content with less oscillators
(they actually called them "operators" of course), since the DX7 had 6 per
voice and the TX-81Z only 4.
An important aspect of FM, too, is the algorithms used. Don't underestimate
the importance of feedback--an oscillator modulating itself.
Well, that's my two cents. Like I said at the beginning, this is how it was
explained to me; if someone jumps on the list and says, "no, no, no you've
got it all WRONG!" then listen to them. <g>
As a post-script, I have not used a through-zero VCO myself, so I don't know
if the effects are radical or subtle compared to a 'normal' VCO. But I do
remember Wendy Carlos groaning, after the hoopla surrounding John Chowning's
algorithms and the release of the DX synthesizers, "What's the big deal?
Rachel and I were doing that stuff fifteen years ago." So reading into that,
and knowing that she was using Moog modules, it makes you wonder if it's
really worth losing sleep over the TZ aspect of the VCOs. I just don't know.
-----Original Message-----
From: Scott Evans, Gen Mgr [mailto:
esresource@...]
Sent: Wednesday, 04 December, 2002 1:09 AM
To: Tkacs, Ken; MOTM
Subject: Re: FW: [motm] FM Synthesis w/ MOTM
Ken and group,
I am curious if someone would elaborate a bit on this subject. Is it the
case that the DX7 oscillators are "through zero"? Am I correct that this
comment says that the MOTM VCO will not go through zero? I cannot see
much relevance in the phase of the VCO output signal being important
sonically, unless I am really missing something here. FM generates
sidebands that extend equal sides of the carrier so what difference does
the phase make? Is the sonic effect that is desired the actual passing
through zero or is it some other issue?
Thanks in advance for a discussion.
Scott
----------------------------------------
"Tkacs, Ken" wrote:
>
> If you're going for the whole DX7 kind of thing, remember that these VCOs
> are not the "through zero" type (extremely few analog VCOs are), meaning
> that as the modulating signal goes negative, the output phase of the
second
> VCO will not "reverse."
>