It really depends on the size of the buffer~, with a larger buffer you limit the artifacts introduced by the interpolation for the "missing samples".

Thanks for the reply Emmanuel. I get what you're saying. I don't get why the "missing samples" don't effect cycle~ in the same way they effect wave~ though.

Also, if you set the the frequency to 30 so you can see the waveforms clearly on the scopes, and slowly bring the feedback in, by the time you get to .0045 you can see that there is a kink in the waveform at the bottom that isn't there in the top waveform, which starts making the two sound very different. Any idea what's causing that?