VCO or VCF

Mon Aug 4 10:23:39 CEST 1997

3)
	id sma001777; Sun Aug  3 10:46:31 1997
Received: from horus.sara.nl (horus.sara.nl [192.16.188.40]) by ns1.sara.nl (8.8.
3/8.7.1) with ESMTP id LAA03689; Sun, 3 Aug 1997 11:45:16 +0200 (MET DST)
Received: by horus.sara.nl (8.8.3/4.04)
          id JAA99777; Sun, 3 Aug 1997 09:42:01 GMT
Received: from mail.enterprise.net by horus.sara.nl (8.8.3/4.04)
          id JAA45751; Sun, 3 Aug 1997 09:41:56 GMT
Received: from oakley ([194.72.196.221])
	by mail.enterprise.net (8.8.5/8.8.5) with ESMTP id KAA06073
	for <synth-diy at horus.sara.nl>; Sun, 3 Aug 1997 10:42:00 +0100 (BST)
Message-Id: <199708030942.KAA06073 at mail.enterprise.net>
From: "Tony Allgood" <oakley at enterprise.net>
To: <synth-diy at horus.sara.nl>
Subject: VCO or VCF
Date: Sat, 2 Aug 1997 23:04:42 +0100
X-Msmail-Priority: Normal
X-Priority: 3
X-Mailer: Microsoft Internet Mail 4.70.1161
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Sender: owner-synth-diy at horus.sara.nl
Precedence: bulk
-------------- next part --------------
     <turning lurk mode off...>

     Hi there list!  This is my first real posting, so here goes...

     In reply to Tony's posting:

     Your idea of building a Walsh function-based synth sounds great way!!  
     I built a 16 term Walsh synth as a demonstrator for my MSc research 
     project a couple of years ago.

     Virtually _all_ of the synth was discrete analogue/digital hybrid (the 
     only microprocessor was an 8051 to implement the MIDI interface and 
     control the digital master oscillator from a large pitch table).  The 
     rest of the beast - three cabinets in all - consisted of LOTS of knobs 
     for the term mixers, three envelope generators, a pseudo-sinewave LFO 
     and the multiplication matrix.

     The route I chose for the Walsh terms was to invert a control voltage 
     and then chop between the normal and inverted signals.  The chopping 
     was controlled from the digital Walsh function generator, feeding a 
     bank of 4053s.  Doing it this way meant I could then use simple mixers 
     to create the control voltages from a number of sources.  My synth had 
     six: two EGs, an LFO, Pitch Bend, After Touch and Mod.Wheel.

     BTW: combining several EGs is a great way of generating complex 
     multi-stage envelopes.  Using two simple Attack-Decay envelopes I 
     could create weird, bipolar four-stage envelopes just by summing the 
     two outputs.  Anyone else played with this idea?  Stick four AD EGs 
     together and you have an eight-stage EG!

     Results?  For a piece of demo kit that was lashed up on breadboard, 
     without any thoughts on layout, it was predictably noisy.  But so 
     what!  For bass sounds, it was, umm, frightening..!  Having all the 
     terms set up so that the shape changed from sine to square or 
     triangle, depending on how hard the key was hit (or maybe After 
     Touch?) was FUN.

     After all this excitement, I've calmed down a bit.  My back-burner 
     project now is a 128-term digital Walsh function synth with many more 
     modulation options - the thing with Walsh function synths is the 
     infinite number of ways of modulating the Walsh coefficients: some of 
     my ideas:

        * feedback from the output into the terms
        * external input, perhaps via a set of filters (like some evil 
     vocoder beasty...)
        * more input from the player (pedals, breath, poly after touch)
        * more complex envelopes
        * more LFO's
        * modulating LFO's with envelopes/other LFOs/After Touch/Mod Wheel, 
     etc
        * reverse-processing of audio signals, extract coefficient 
     envelopes, then swap them around and play it back - resynthesis????

     One thing I found with my Walsh synth was that I fell between the 
     all-digital and all-analogue camps.  On the one hand I got
        "Analogue? Isn't that noisy? Why not do it all on a DSP?",
     then I got 
        "Ugh! What's this microprocessor doing in here?  And your 
     oscillator is DIGITAL!!"

     Me, I just turned up the volume so I couldn't hear them anymore!

     Well, I hope Tony has as much fun designing, building, and playing his 
     synth as I did with mine.  If anybody else has built Walsh function 
     synths, I'm sure we'd all be interested in hearing from you.

     Its a great list guys.  I'm addicted!

     Neil

     --
     +---------------------------+-----------------------------+
      Neil Johnson               * All thoughts and opinions
      Cambridge Consultants Ltd  * expressed here are my own.
      email: nj at camcon.co.uk     * Heck, even _I_ don't believe
      http://www.camcon.co.uk    * what I say at times.....
     +---------------------------+-----------------------------+ 

______________________________ Reply Separator 
_________________________________
Subject: VCO or VCF
Author:  "Tony Allgood" <oakley at enterprise.net> at Internet-PO
Date:    02/08/97 23:04

Reading some of the notes in the archive, it seems to me that some people 
think that its the VCF that determines the overall sound of the instrument. 
Do the oscillators actually make a difference? One person said a squarewave 
is a squarewave whatever the VCO design. Surely, this is not true. Slew rate 
and HF performance of the semis will have an effect, won't they? And 
sawtooth outputs, well, some sawtooths can have terrible flyback times. For 
instance, my old Rogue (which I sold for 40 UKP about 5 years ago, what a 
plonker! ), had a sawtooth with a flat top, like it had been clipped at 
12volts.

And what about VCAs, all my modular stuff uses LM13600 or 3080s, does a 
AD633 really make a difference that can be heard? Or does it only make a 
difference when you try to modulate with AF rather than slow EG outputs?

I will agree that VCF types really make a difference, but for my next 
project, I'm going to be working on oscillator waveforms. I am going to use 
a 3340 running at a very high frequency driving a Walsh generator bank. I 
tried ages back but I tried to be clever and use a digital EG circuit I had 
designed for each of its 32 outputs. This time I'll use a bank of pots. 
Anyone out there tried using hardwired digital chips to make analogue 
waves? I will keep you posted of any positive results.

TTFN

Tony

PS anyone got the circuit schematics for the powertran vocoder or the FAT 
procoder?