[sdiy] saw vs ramp, audible?

[Mattias Rickardsson]

On Mon, 16 Dec 2024 at 07:49, Donald Tillman <don at till.com> wrote:

> On Dec 9, 2024, at 4:03 AM, Mattias Rickardsson <mr at analogue.org> wrote:
>
> Anyway, triangle/square/saw do have different overtone phase
> relationships, and I agree it's very important. Both when constructing new
> waveforms and when mixing oscillators together. The overtones cancel or
> reinforce in different ways, some of them useful and some not helpful at
> all.
>
> Are there any synths using modified waveforms with alternative overtone
> phase setups, in order to match each other (or avoid matching each other)
> for different results?
>
>
> I'm guessing I'm the only one who builds "phase correct" oscillators.
>

Not really, I've dabbled with it too lately. (But shhh, don't tell anyone!)


> I use the term "phase correct" for an audio oscillator where all the
> waveforms have their fundamental components in phase, and each waveform has
> as many of its harmonics in phase as possible.
>

But don't you sometimes run into problems with sawtooths blowing up into
infinite amplitudes if you move around the overtone phases?


> And I've found significant musical benefits mixing waveforms with this
> approach.  But that's me.
>

I can imagine there's lots of good applications for phase-correct
waveforms, thanks to the fact that they add up purely constructively
regardless of the waveform. When mixed and detuned they also "phase"
towards each other with a cancelling/reinforcing behaviour similar to
classic sawtooth phasing.

I'm interested though in what it would give to NOT have them only
phase-correct, but rather have the overtone series continuously
phase-adjustable. Both mixing and distortion would give interesting results
not available with static waveforms even when mixed together. Feels like
there are new useful synthesis methods hiding somewhere there...


> But no matter... It's also the case that the iconic ARP 2600 VCO is
> probably the least phase correct, as all four of its waveforms have
> fundamentals in different phases.  (!!!)
>

:-O


> Phase-correct waveforms are difficult for a sawtooth-core oscillator as
> the triangle and sine are derived from the sawtooth at a 90 degree phase
> difference; the triangle peak (90 degrees) is aligned with the sawtooth
> retrace (0 or 180 degrees).
>

Also the triangle has its overtones in other phases than the square for
instance, so there's no way of phase-aligning all the traditional waveforms
saw/sqr/tri/sin and get all their overtone phases to match. But
the triangle becomes a nice almost-half-circular waveform if you align them!

/mr
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20241219/0615374b/attachment.htm>