MOTM

> As you mix more of the output back to the
> input, the excursions from one sample to the next will be smaller.
> (There must be a mathematical formula for this.)

Yn = A*Yn-1 + B where B is the random sample

...otherwise known as an Infinite Impulse Response Digital Filter, also
referred to as a recursive transversal filter.  While the implementation is
different, the result is the same as simple lag applied to the input of the
S&H.  Both are filtering, one analog, one digital.  Both give the same
result, correlation of the samples to a more confined set of values.  As A
increases in the above equation, the S&H output will approach a steady state
output.  Taking the analog lag example, as more lag is applied to the sample
B, the S&H output will approachy a steady state output as well.

John Loffink
The Microtonal Synthesis Web Site
http://www.microtonal-synthesis.com
The Wavemakers Synthesizer Web Site
http://www.wavemakers-synth.com


> -----Original Message-----
> From: motm@yahoogroups.com [mailto:motm@yahoogroups.com] On Behalf Of
> Richard Brewster
> Sent: Sunday, January 01, 2006 9:15 PM
> To: jwbarlow@...
> Cc: motm@yahoogroups.com
> Subject: Re: [motm] Re: Another thought
> 
> Correlation is done by mixing some of the present output of the S&H back
> with the input.  You need a bit of lag between the output and the input
> to the mixer, so that the old output voltage holds for a moment as the
> sample occurs.  That is because the output will change at the instant of
> the sample and you need the old output for a few milliseconds for your
> mix.  You can patch this up with any S&H that has a lag following it.
> The MOTM-101 has a lag section.  To try this, patch your random signal
> (say the Pink Out of the 101) into a mixer, patch the mixer out to the
> S&H input, and patch the output of the S&H to another input on the
> mixer.  If you patch the S&H out to the FM input of a VCO it will be
> easy to monitor the effect.  As you mix more of the output back to the
> input, the excursions from one sample to the next will be smaller.
> (There must be a mathematical formula for this.)  The correlation
> feature is provided in the Oakley S&H, and that's what I am asking for
> in a new MOTM S&H.
> 
> You could add a DC-coupled VCA in the feedback loop.  But I don't think
> it would be terribly useful, due to the fact that the effect of the
> correlation is subtle.  The output is still random, but it has a
> different pattern.  I find just being able to dial it up with a knob is
> plenty.  Adding more VC inputs to this proposal would push it toward a
> 2U size, and most people seem to prefer 1U.  You can always patch a VC
> feedback loop.
> 
> -Richard Brewster
> 
> jwbarlow@... wrote:
> 
> > In a message dated 1/1/2006 4:50:04 P.M. Pacific Standard Time,
> > jloffink@... writes:
> >
> >     S&H Correlation is typically implemented as a lag circuit prior to
> the
> >     sample and hold.
> >
> > I haven't kept up with all the permutations of this module, but I'm a