[sdiy] Analog division

[cheater cheater]

On Tue, Jul 20, 2010 at 14:53, Tim Daugard <daugard at cox.net> wrote:
>> a logarithm can only be evaluated at a positive number. If this module
>
> accepts negative numbers then it does not count their logarithm. It
> can be said to be a generalized logarithm function.
>
> It effectively takes the log of the absolute value of the difference between
> the ref and the input and then puts the sign back on. Or at least simulates
> this function. It functions as an individual module as a signal compressor
> with no time delay. It compresses both halves of the waveform symetrically.
>
> Tim
>

Hi Tim,
Yeah - that's what I'd do, if I knew how to design electronics ;-)

I'm just making this distinction because you have to remember that,
otherwise the math doesn't work out, it's a bit of a gotcha if you're
used to all signals being always able to attain positive and negative
values and then suddenly you are confronted with something that has to
stay positive always, at least during some stages. Then there's the
question of what the output is doing when the signal is completely
disconnected. I see the circuit you posted seems to output a 0V when
this happens, right? I'm a bit unclear :-) it does look fairly cool.
TBH I wonder what it sounds like on audio. Possibly similar to a
softer tanh distortion?

Cheers,
D.

On Tue, Jul 20, 2010 at 14:53, Tim Daugard <daugard at cox.net> wrote:
>> a logarithm can only be evaluated at a positive number. If this module
>
> accepts negative numbers then it does not count their logarithm. It
> can be said to be a generalized logarithm function.
>
> It effectively takes the log of the absolute value of the difference between
> the ref and the input and then puts the sign back on. Or at least simulates
> this function. It functions as an individual module as a signal compressor
> with no time delay. It compresses both halves of the waveform symetrically.
>
> Tim
>
>
>
>
>
>
> Cheers,
> D.
>
> On Mon, Jul 19, 2010 at 21:25, Tim Daugard <daugard at cox.net> wrote:
>>
>> From: "cheater cheater" <cheater00 at gmail.com>
>> you can only take logarithms of positive numbers, so it wouldn't work
>> without modifications. However you can rectify each signal and extract
>> the sign with a comparator, then multiply them with e.g. an AND gate,
>> and apply that to the output. Then you should be fine, I think.
>>
>> This module:
>>
>> http://members.cox.net/synthfred/h_alg121.htm
>>
>> provides a signal that is the log of the signal centered on the reference
>> voltage at the top of C7. If you look at the description, it takes the log
>> of both the positive and the negative going portion of the signal. The
>> waveforms from the theory section match the waveforms from the built
>> module.
>>
>> The math would be:
>>
>> ref+(log(sig-ref)) --> if the signal is greater than the ref
>>
>> OR
>>
>> ref-(log(ref-sig)) --> if the signal is less than the ref.
>>
>> I'm not positive I got the math right but, the figures are 100% verfified.
>>
>> Tim Daugard
>> AG4GZ 30.4078N 86.6227W Alt: 12 feet above MSL
>> http://members.cox.net/synthfred/h_toctop.htm (Fred's online edition)
>>
>>
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at dropmix.xs4all.nl
>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>>
>
>
>