[sdiy] Synthex Oscillator

[Ove Ridé]

On 8 July 2017 at 11:49,  <rsdio at audiobanshee.com> wrote:
> You seem to be thinking that a variable sample rate based on an analog clock has no aliasing to worry about, but that's totally false on all counts.
>
> The fact that the rate is variable does not remove aliasing. A variable sample rate simply has a variable Nyquist frequency, such that aliasing moves around as the rate varies.

There's a difference between aliasing and aliasing that you need to
worry about. The waveshaper's sampling frequency (and thus Nyquist
frequency) is a fixed multiple of the fundamental frequency it
outputs. This is the stepped response, when looking at in the time
domain. Given a perfectly jitter-free clock, this will only introduce
aliasing that's in harmonic bands to the fundamental, and the aliasing
will be consistent across pitches. Thus, it will appear musically as
the "character of the oscillator", rather than the "nasty" type of
aliasing. Or to put it plainly, the oscillator would not sound "weird"
if you were to sweep its pitch.

In this circuit, candidates for "bad" aliasing would instead be jitter
on the clock feeding the waveshaper, due to the pulse swallowing
mechanism, as well as the sync input, since the time from the sync
event, to the next phase incrementation even will vary between
different sync events.

-- 
/Ove

Blog: <http://blog.gg8.se/>

"Here is Evergreen City. Evergreen is the color of green forever."