My first post here, hello all!

As an exercise I'm making a patch that emulates an oldschool flipflop squarewave octaver like in an MXR bluebox or in some analog synths and organs.

I'm getting really close using {expr~ $v1 > 0] to convert a sine to a squarewave and with a short delay [expr~ $v1 > $v2] compares one sample to a delayed one to check if the value of the new sample is higher than the one before. This gives me a pulse every time the value went from 0 to 1. Those pulses go to an [rpole~ 1] as an integrator and [expr~ pow(-1, $v1) ] converts that ramp to 0 and 1 at half the original frequency.

This does work, but at higher frequencies I'm getting a 'ripple' on both rising and falling edges of about three or four alternating 0's and 1's that I can't seem to find the culprit of. Any ideas?

@rumblefuzz it would help if you could post the patch..

of course, make sure you're using a 1-sample delay. Maybe would be better to use [rzero~ 1] with [expr~ $v1 > 0] for this I think

I think you're going to get some kinds of aliasing problems with this just fyi.. you'd have to bandlimit the steps/account for sampling if you want it to sound analog..

Actually that first remark was the solution . The delay was set to 7 samples because at some point I was messing with that to get it to trigger right.

Now the only thing left is to get this so I can 'stack' these to transpose multiple octaves

Here is the patch if you want to have a look (hopefully this works):

divide down test.pd

I see you're clearing the [rpole~ 1] every 100 seconds using [metro].
Won't this cause a hiccup since it's running at control rate?

It does (sometimes)

That was still a temporary measure so I don't have to worry about that number getting too high. In the final implementation I plan to trigger the reset from a noise gate on the input, or, if it becomes a synth, from a note-off msg