linear vs exp FM

[Martin Czech]

> And then the last; all synth manufactureres have used exponential 
> keyboards in their synths except KORG wich has used the V/Hz norm. So 
> was KORG the only one with linear FM (MS20?)? I don't understand what 
> the difference between lin FM and exp FM is actually; the oscs in a 
> synth are controlled by a CV from the keyboard which increases it's 
> output by V /oct right? The two oscs in say a jp8 track each other 
> quite evenly so why has the jp8 still got exp FM? Does it sound very 
> much different than linear FM?
> 
All Theory a la Chowning and FM-Radio applies only for linear FM.
This includes that the phase pointer will run backwards for some time
if the frequency modulation is very strong.

If you use the exponential input for FM, your input signal will be
predistorted via the e-function. The basic sound is allmost the same,
however the effect is different spectra AND a shiftet frequency mean
value, ie. pitch. This is undesireable.

It is no problem to modulate the reference current of the e-converter,
this will give linear FM. In the case of a triangle osc you can also
reverse the osc direction if the linear input crosses zero, that is
real through zero modulation. This will enshure proper operation even
for large FM.

If the vco is linear right from the start things are easier, of course.
But in this case you may need additional cicuitry to allow for equal
distance vibratio (up = down in musical terms).

The latest discussion showed that linear keyboard control is nice
(due to less error) and this can be combined with expo lfo modulation
(etc.), ie. you use still the (infamous) expo converter, but you normaly use
only the linear input (ref. current) whereas the expo input is normally 0,
thus no tempco compensation is needed.

m.c.