[sdiy] Modeling circuits

[Thomas Strathmann]

On 10/22/10 21:02 , Barry Klein wrote:
> I am wondering if they go and do spectrum analysis/ FFT response
> of say a Fender Twin and then only use their DSP platform to do
> a frequency subtractive overlay on the incoming signal - OR -
> could they also do an analysis of the incoming signal and then
> know better what to do to get it to be a better match to the original
> amplifier output response (that would be actually increasing frequency eq
> gain).
> I know right now it seems pretty lame to spend so much to get a VCO to
> match a Moog or whatever, but these costs are coming down and if you
> go this far you can then go much further obviously.  The control voltage
> indicates to the DSP the frequency, then analyze the input waveform and
> modify it as desired.  Such modification could be scripted too. So rather
> than just a wavetable thing this would be that and more.

I don't think I understand exactly what you mean. The first case you 
describe seems to be something like a convolution with the frequency 
response of a device. Then you generalize this to multiple frequency 
responses which are chosen according to certain characteristics of the 
input signal? The link I included in my previous mail has some papers 
about modelling non-linear circuits directly, i.e. building equations 
that can be implemented in a DSP.

	Thomas