So, I've finally got a buddy who wants to do some live sampling, on top of pre-recorded loops. We'll be dubbing with outboard analog gear, so basically I need to make a multitrack looper. My question is: say I've got a 1-second sample, looping via tabread4~, and a 1 sec delwrite~/delread~ combo. I play back the sample via a phasor, and use the delay with feedback to create a layered loop. Is it possible for the two to "drift" apart?

my guess is "no", as the delread~ SHOULD be triggered faithfully after exactly 1 second. And likewise a [phasor~ 1] SHOULD be exactly 1hz.

to test it out though, maybe just do a 10ms delay, and a 100hz phasor and let that run for a while to see how you go?

here's a little test patch. i couldn't notice any drift when the phasor rate and delay read are the same.

however, if you DO change the rates, you get cool flanging effects, which could maybe have some use?

http://www.pdpatchrepo.info/hurleur/feedback-loops.pd

ahh, excuse all those posts, as i hadn't properly read your question. you wanna do the tabread4~ and delread~'s separately yeah, in parallel.

i just tried that, and there does seem to be some unwanted flanging. it's only pretty minor though, so it might take hours before it'd ever give you trouble.

anyway, attaching my test patch for that too.

http://www.pdpatchrepo.info/hurleur/feedback-loops-parallel.pd

If you're using [tabread4~] instead of [tabread~], then I would suspect that you might be playing back at a rate different from the original sample. If that's the case, you're probably safer using [vd~], since [delread~] doesn't use interpolation.

i thought that at first too, but it doesn't matter. it's the frequency of the phasor~ driving the tabread4~ that is important.

i simplified the problem, and made a diagnostic patch. by sending a noise pulse into a delayline, and viewing it in an array, you can clearly see how to delayline drifts out of sync with the phasor~

so, unless i have messed up something, i think this proves my initial assumption wrong.
a phasor~ with 5ms wavelength does not seem to time perfectly compared to a delread~ after 5ms.

http://www.pdpatchrepo.info/hurleur/check-sync.pd

geez, i'm making a right royal mess of this thread,

but i just tried some more stuff, and it seems that phasor and tabread DO sync up properly at other values than the ones i tested.

Cool, I'll check out those patches. I made one for myself, but short of empirical knowledge, I think it is safe to say: close enough for Jazz. I'm never going to be sampling the same thing to a delwrite~ and a table, so it really isn't an issue. I was more concerned with rhythmic accuracy over a few minutes.