[sdiy] FM spectrum question

[ニコラス・ケント]

On May 5, 2006, at 4:02 PM, jbv wrote:

> Thanks for the reply.
> But I already have some litterature about Chowning theory.
> The link you pointed at features all useful info to compute
> spectrum from carrier & modulator freqs, but what I'm
> looking for is the opposite : to determine C & M from
> spectrum content...
> I was hoping there was some math method (possibly
> implemented in some pice of software) for that purpose...
> Thanks anyway.
> JB
What struck me when you asked was what configuration of FM would the  
target model be? I think you need a defined set of variables before  
you can complete the formula.

Do you mean just one carrier and modulator? You would  only cover a  
relatively narrow range of waveshapes. Certainly that simple a  
configuration would yield no match most of the time and you'd have to  
devise criteria as to including a number of partial matches. (you  
know, how much deviation from an existing ratio and index worth  
considering)

  It goes without saying that if you have a multi carrier system and  
then configure it with no effective modulators then you have additive  
synthesis and there are commonly known transform functions available  
to you.

One thing I've heard rumor of. Yamaha was rumored to be developing  
some sort of an analysis and resynthesis software app that would  
create a patch for the F1SR based on a waveform or audio file. But  
the idea was dropped as were subsequent formant synthesizers. The  
F1SR has a superset of the DX configuration (and a huge number of  
parameters compared to the original DX7). It significantly adds  
several resonant digital formant filters.

nick kent