Artefacts in add1 or my brain?

[marlon]

Hello,

I have an issue in add-1 I don’t quite understand: In a sound synthesis process using “add-1” with 50 partials at random frequencies between two limits (e.g. 100 and 1000, all remaining params set to defaults), there are pulses in the resulting sound at a constant time interval of 1 sec. I have attached some screen caps. I was presuming this to be due to phase interference, however, the time interval of the pulses seems to be steady at an interval of 1 second, regardless of the range of the frequencies. It doesn’t happen (or at least not to this extent) for synthesis with fewer partials. I’m curious to what might be causing this effect?

Best,
M

[marco]

Dear Marlon,

The result is normal. If om-random is given integer inputs, it will generate only integer numbers, whose frequency resolution is 1Hz. If you give a floating point input, the randomly generated frequencies will be floating point, and the “beatings” disappear.

Hope this helps you.

Good “add-1”-luck!

Marco

[marlon]

Haha - I knew it must have been something like this. A nice example indeed.

I did a small modification to understand a bit better though…
e.g. in the attached screen cap all delta-frequencies of 1 Hz are removed - yet the beatings persist since the effect is caused due to all frequencies being integer-multiples of 1Hz (since all frequencies are “quantized” to 1 Hz this is like producing partials of a phase-aligned harmonic spectrum with an f0 of 1Hz). Ahh, the wonderful world of phases and frequencies…

Thank you

[marco]

This is exactly the point. Since all the oscillators, with add-1, start with the same phase, even if you eliminate the frequencies that are 1Hz apart, since all the others have the same multiple distance, you get something like a phase-aligned signal (not necessarily harmonic). If you closely look at the “clicks”, you will see that they are just phase-aligned sinus tones, and not distortions or discontinuities in the waveform.
Good “phase”, then!

Marco