Casio CZ/ VZ/ FZ - Pro Series

>For your interest, Poly-800 uses KORG MSM5232 for DCO, and KORG NJM2069
for VCF and VCA.
>(See: http://www.asahi-net.or.jp/~wz4k-tnk/semi/poly800.html
>It is a great site but written in Japanese.)

right.  I think that page is where I initially got the info.  A search for 
MSM5232 will give you a lot of arcade game pages...

And not that it has anything to do with anything but the 2069 filter chip 
has a beautiful sound and was also used in the DW-8000 and DSS-1.

>I opened my HT-3000 to compare the chips, and found what I believe to
>be the main circuit board. Besides the CPU PD78C10 and a memory chip,
>there was mysterious chips called MSM6294-03 and MSM6294-04, manifactured
>by OKI semiconductors. The name MSM is used by OKI, so KORG's chips may
>also be made by OKI. But I found no information on the net about these
chips.

Interesting that you mention the PD78C10, because my VZ-10M uses a PD78C14, 
both NEC processors in the same family...but I have no idea what the MSM 
chips would be... Let's see...to get really speculative, the custom chips in 
the VZ-10M were made by Hitachi, and I think I've noticed some other brands 
in other casios, but never a "casio" chip.  This would indicate to me that 
Casio doesn't (or at least didn't at the time) manufacture their own 
chips...thus the OKI chips are likely custom creations, which unfortunately 
means that it's the end of the road as far as digging deep into the 
synthesis methods of the HTs goes (unless you want to try 
reverse-engineering the whole thing...and I'm sure you have something better 
to do there...)

>I haven't thought of the possibility of creating triangle wave out of
>square waves. I don't know much about phase issues.
>As mp3s on my site shows, HT's DCO can't generate such a soft-sounding
>wave like triangle. They produce so many variations of pulse and square
>waves.

Well..a 4-bit triangle wave would hardly sound soft.  It would have the 
basic triangle tone but with a lot of high ringy harmonics.  The Nintendo (I 
guess that would be a Famicom to you?) triangle wave was 4-bit and had a 
very interesting quality to it..

>'Additive square wave' synths may not be able to control phases and pulse
>width. I think the pulse widths are fixed to 50%. In fact, Poly-800 has
no Pulse width parameters.
>Poly-800's DCO and Roland's sub-oscillators generate square waves by
>'dividing' the master clock. So, the phases of each harmonics are always
>synced, I think.

That's what I was getting at.  you can't get truly complex waves when 
they're all phase-synced and fixed pulsewidths.  So if you were really 
determined, you could get a bunch of graph paper or something, look at a 
good variety of the waveforms and try to recreate them by using octaved, 
phase synchronized square waves of different amplitudes.  This would is 
probably a lot more effort than it's worth.

 > Seeing as how it has 16 steps, the saw wave looks more like it comes
out of
 > a 4-bit DAC.  To me this would seem pretty likely.  It still doesn't
explain
 > how the waveforms are made, and particularly why they're affected by
the
 > amplitude envelope, but it makes some sense...

>When I first saw the waveform displayed, I thought my HT might be a 4-
>bit PD synth. But after a while, I came to think that PD needs at least
>8-bit DAC or so. So I'm confused again as to its synthesis engine...
>I'm still rather for my 'additive square' hypothesis, but I'm not so
>certain about whether my HT has a ring modulator...

I don't know...In theory phase distortion could be 4-bit but it would have 
too much quantization noise to be useless to all except people like me :) I 
think one thing you need to do is see for sure what the waveform is doing 
when it decays...It could be getting distorted slightly by the filter at 
higher amplitudes but it doesn't look like it when I see the saw wave.  
Maybe you could cut a bunch of short samples from a long decay, normalize 
them all and see if they're any different.  If it's actually undergoing any 
sort of meaningful change then it effectively rules out PCM samples (unless 
they used some sort of wavetable thing like a PPG, in which case it would be 
quite overcomplicated as well as very poorly implimented...)

You might also want to check out the noise source too and try to figure out 
how that works.

_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com