I've been trying to get something very simple happening: using groove~ in time stretch mode, I want to create a loop that alternates forward and backward play. I.e., it starts at point x, runs forward until point y, then reverses and runs backwards to point x, runs forward again etc. It would seem to be a very straightforward thing to do, and I've tried multiple solutions. Have a look at one of the obvious ones attached. Using a phasor~, I generate a ramp and then turn its output into alternating 1s and -1s. With all the appropriate settings activated, this should produce a sample-accurate 'turn around', and so we ought to hear exactly the same audio looped forwards and backwards: that's not what happens - take a listen and look at the output of minmax~. The loop points keep changing; my impression is that they change according to a regular pattern.

I'm aware of the old thread about making a forward/backward loop with groove~, and I've also solved this problem using gen~ to analyse the sync output itself and determine the 'turn around' points that way, but I'm just curious why this phasor~ solution doesn't work properly. Is there some kind of inconsistent relationship between the content of the buffer and the sync output (sample interpolation related, perhaps)?

abs~

change~

comment

loadmess

flonum

spike~

edge~

message

number~

minmax~

buffer~

ezdac~

phasor~

sig~

groove~

I'm aware of the old thread about making a forward/backward loop with groove~, and I've also solved this problem using gen~ to analyse the sync output itself and determine the 'turn around' points that way, but I'm just curious why this phasor~ solution doesn't work properly.  Is there some kind of inconsistent relationship between the content of the buffer and the sync output (sample interpolation related, perhaps)?

Hi all,
I've been trying to get something very simple happening: using groove~ in time stretch mode, I want to create a loop that alternates

odd-groove~-behaviour

Odd groove~ behaviour