Super Filter = No filter

[Bissell, Harry]

Yeah, it's called additive synthesis. To finish the job, the phase of each
harmonic (or non-harmonic) needs to be controlled also. I think the
Synclavier and Synergy either used this approach, or it was an option. The
Yamaha DX series were capable of a limited form of this, one "operator" had
all wave generators in parallel. The down side is how would you control it,
and how hard is it to program sounds.   :-) Harry

> -----Original Message-----
> From:	Gene Zumchak [SMTP:zumchak at cerg.com]
> Sent:	Wednesday, March 03, 1999 12:13 PM
> To:	DIY
> Subject:	Super Filter = No filter
> 
> List,
> 
>  A recent thread about waveform synthesis reminded me of an  idea I had
> about 25 years ago when I was into synths. I wanted to build a single
> voice channel (ala Moog).  Since I had visions of commercial
> possibilities, I needed to make it without a (patented) Moog VCLPF.  My
> background is digital, and the things I did for Moog (like the 960
> sequencer) were digital.  I was into microcomputers from the beginning.
> (I owned a KIM-1 (6502) serial #000005, which at $250, cost less than an
> 8080 chip set at the time.)
> 
>  My goal was to make the perfect filter.  Imagine a waveform, like a
> sawtooth, with a particular spectral recipe, and a filter bank in which
> each filter affects only one harmonic.  Each filter would have its own
> ADSR generator to control its harmonic.  If the waveform is the sum of
> the fundamental and some number of coherently generated harmonics, then
> a VCA/ASDR  for each harmonic, prior to the summing will give the
> identical effect as a filter bank with individual time control of each
> harmonic.  That is, the ideal LP filter would be no filter at all, but a
> bank of coherent harmonic generators each with its own VCA and envelope
> generator.
> 
>  To make the coherent fundamental/harmonic generator, I planned to use
> EPROMs tied to DACs for table lookup of sines.  Muxing data  out of a
> single EPROM modulo eight, I could supply waveforms for the first eight
> harmonics (enough?).
> 
>  Needless to say I never built it.  Today it would be a lot easier and
> cheaper.  In fact, I might be tempted to calculate everything in the
> micro (a 16 MHz 68HC12) and generate the already enveloped, filtered
> waveform directly.  Has anyone ever attempted this approach to a single
> voice channel?
> 
> Gene Zumchak
> 
>