Spreading the FFT transform/process CPU load


I am doing a dozen chain.pv_collect from a single chain. Obviously that peaks at the transform point (when we get new fft data) and in MaxMSP there is a neat mention where we can offset an fft process by blocks so we have a less peaky and more steady CPU load.

(sorry to defile sc with max code here, but check the start-offset argument here: pfft~ Reference - Max 8 Documentation)

I’ve tried to find a way in SC to do that, but the FFT doesn’t seem to allow that.

Is there any good practice around this type of load-management?

Just schedule your Synths at different offsets?

I don’t understand how you access the trigger element of an PV Ugen…but if you could delay that, you would be golden.


thanks both. @Spacechild1 that would be good but I want the same FFT input - I could delay the inputs I reckon…

@Sam_Pluta the processes are triggered by the FFT chain, when not -1, as you know, but I’ve tried to delay that and I can’t neither as trig, as kr or as ar…