MOTM

In a message dated 3/22/2000 10:50:32 AM, ken.tkacs@... writes:

>Really? I find that very strange. I've *always* approached synthesis with
>the harmonic spectrum in mind. The waveform is the *result* of the sound's
>harmonic content. 

Or is the harmonic spectrum the *result* of the waveform? HA!
Obviously, they are each a different perspective of the same phenomenon. But 
I agree entirely about an approach (from a musical point of view) to a 
particular sound, either add harmonic content, or subtract. My comments about 
waveforms were more (as I think Bradley saw) directed at the engineering 
point of view -- how do you build such a device?


>Also, additive synthesis and Fourier analysis predate digital by a pretty
>good margin. These were terms that we used to discuss in the old analog
>days
>long before the first digital synthesizers hit the market. 


Cologne school.


While I agree
>that additive synthesis is probably easier to implement in digital, I don't
>think it's at odds with analog at all. That's why we use more than one
>VCO!


I think a similar idea has been batted around here in the past, and I seem to 
recall that a statement about DSP being a cinch for this kind of thing -- 
VC-DSP, hmmmm.

>I agree that waveform manipulation is easier in analog, and I think that
>the
>discussions of waveform shapers, folders, and multipliers can be very
>fruitful. But as a generalization, they still only affect the amplitudes
>of
>the existing harmonics (I usually use the term "partials" because it was
>beaten into me that the Fundamental is not a harmonic, so when you want
>to
>refer to *all* of the Fourier components, including the fundamental, you
>need to use 'partial' rather than 'harmonic.

>Know what I mean though?  After the fundamental, the harmonics in simple
>waveforms are specified as period-times-1/2 (octave, or first harmonic),
>times-1/3 (octave-fifth, or second harmonic), times-1/4 (two octaves, or
>third harmonic, etc.). You can filter this, waveshape it, and so on but
>you
>still aren't changing these relationships unless you run the signal through
>a ring modulator. The fact that these relationships are 'set in stone'
>while
>so many other aspects of the sound are open to us for manipulation ...


>The reason this kinda thing keeps me up at night is because I'm interested
>in microtonal music. Our musical scale closely parallels the harmonic series
>of simple waveforms like we've been talking about (at least until you get
>up
>in the range where the compromises of equal temperament cause a divergence).
>When you look at the unusual scales of other cultures, like the music of
>the
>gamelan, they aren't just playing weird tunings with Western
>instruments---their instruments produce tones with harmonics that echo
>and
>fall in line with their scales, just as ours does. We have instruments
>that
>produce "simple waveforms" (please allow me that ridiculous statement)
>that
>work with our equal tempered scale, and they have instruments that have
>complex tones and scales to match them.


I would be quite interested to know more about this! Do you have any 
references I could look at in this regard? One thing that occurs to me at the 
moment is that many of the metalaphones present in gamalen music might be 
seen as having an FM or AM component to them. Have you looked in to mimicking 
this behavior via AM and FM?

>I find this whole thing very interesting. I'd love to play with those
>relationships. Sure, digital's always an option, but I want to do it with
>patchcords and knobs! Typing on a keyboard to specify harmonics is for
>the
>birds.


As you rightly point out, additive EM is an expensive proposition outside of 
digital. But, as I said above, I wonder if extremely precise use of FM (or 
AM) might yield some similar timbres to what you're after. In my previous 
experiments with one VCO modulating a second which then modulates the first, 
I've gotten many interesting timbres including lots of non-harmonic 
overtones, but I've found that I can only use these voices over a very 
limited range since every error is multiplied and becomes profound rather 
quickly. I've had a bit of luck using sync to reduce this error without 
totally eliminating the complexity of the timbre.

Good points Ken! Keep em coming!
JB