Home | ER-101 | ER-102 | ER-301 | Wiki | Contact

Custom Unit Build: 4 Poly Round Robin Sample Player



I still don’t really get this unit, although I sure appreciate it! (Kind of like the Room of Requirement haha).

I just set up the polysampler using joes video as a step by step guide, and I noticed that for the bump scanner to work for my sample the width had to be at least 133 or so. I don’t know why, but it works and I’m happy!


Hmm, the specific sample you’re using shouldn’t have any bearing on what the width parameter of the bump scanner needs to be set to. In this case since we’re doing a “hard” window with very controlled input values (from the count custom unit), the width should be able to be even smaller than what I used in the video.

I would go watch the output of the Count custom unit and send it triggers. Does it appear to be stopping at exactly 0.0, 0.25, 0.5, 0.75? If not, try going into the Count’s reset trigger and sending it an initial reset from the on-screen fire button just to get things lined up the first time.


Thanks, Joe. When I set it up, the count custom unit was going up in four increments and resetting, and it sounded as it should. I’m finally getting a chance to look at it again today. But now, for some reason, it’s not working at all. So I’ll start again from scratch.


While trying to understand the bump scanner I was looking in all the wrong directions because of the suggestion of an object with an X and a Y axis, so yes changing the height parameters name would be good. Maybe it could just be ‘value’ or ‘output’, at the end of the chain?


I think I posted this before but just a reminder. The above picture shows how the Sample and Bump Scanners work. The Sample Scanner (blue line) is the same as the Bump Scanner (black line) but with the trapezoid replaced by whatever sample that you assign (in the picture, a single cycle of a sine wave was assigned).


Of course the first part where I go through setting up the count and do the calculations was obsoleted 3 days after I posted the video. :slight_smile: (Not that I’m complaining. It’s awesome!) You can now set up the subchain on the count custom control in a much more direct (and probably CPU efficient) way using the Counter factory unit introduced in firmware 0.3.17.

In my opinion, it’s still worth understanding Brian’s original count custom unit though. It’s a great example of how you can take a unit (really the concept applies to external modules as well though, too) and coerce it to do something that is useful, even though that’s not really what it was designed for by thinking outside the box.

Yep, looking at this diagram, the control nomenclature makes perfect sense. I have somehow built myself a slightly different cognitive map of how the bump scanner works. I think the models are functionally equivalent. It was good to review this after using the bump scanner some more. I’d stick with it!

What do you think about having 2 fade controls, one for each side? I think this might be useful for controlling crossfade durations sending in a simple ramp as the input. Does that make it too confusing?


I was wondering @Joe, did you end up posting this as a custom unit somewhere? Or is the way it is set up here obsolete due to software updates?


I didn’t post it but I can. The way it’s done in the video still works but I think building it with the Counter factory unit would be better - more straightforward and I would guess a bit CPU lighter but haven’t tested that yet.


Oh yes, we have the factory count unit now. I didn’t have time yet to play with all the new features! If it’s not too much trouble to post that custom unit, I might take a shot at rebuilding it with the new Count unit. There’s still so much uses for this concept.


I added the custom unit preset (which uses the original count custom unit rather than the new Counter built in unit) to the first post.


Hi Joe, Love the work you are doing on here and this round robin sampler is what ive been looking for. I loaded up your lua preset but once ive assigned a sample im not getting it to play back. if i go deep and bypass the bump scanner im getting triggers.
any idea whats going on?


I would try going into the count paramter’s sub chain and give the count custom unit a reset (just fire a single one from the UI). The bump scanners are looking for very specific output values from this chain. If it were to not start at zero that could screw things up.

Also just to confirm you are running the 48k firmware. The count unit math would need to be tweaked for the 96k firmware. It won’t work out of the box on the 96k version. I’ll try to post an updated unit that uses the new Counter built in unit instead of the saw oscillator. That should make it firmware independent.


Resetting the count unit did it!
Thanks so much, this custom unit is amazing :raised_hands::raised_hands:


Finally got around to trying this and its so useful! Great CU!


First, I love this unit, but I have a bit of a stupid question: What is the limiting feature or parameter that prevents expansion beyond 4 voice pseudo-polyphony? What was mentioned in the video describing this unit is that going beyond 4 voices would clip at the output, but is that a limitation of the hardware or something that that could be addressed in firmware updates? Another stupid question is whether or not the Count, V/Oct, and Trig could be adapted and made into Global parameters that could be expanded to, say, create a 2x4 pseudo-polyphonous sample player over two outputs?


It’s been a while since I’ve made this… but I don’t think there is any current limitation that would prevent you from going beyond 4 voice pseudo-polyphony, or moving some of the pieces to global chains to create a 2x4. I think I did talk about clipping but the solution was just turning the mixers down a bit to prevent that. You might have to turn them down a bit more if you go over 4 voices, but no big deal.

So, I’d say go for it. I think you can make either work!


when i use this 4 poly CU with an external sequencer (bsp), the guitar sample i use sounds perfectly in tune, but when i sequence it with varigate 8+, it never gives the correct pitch and sounds out of tune.
Anyone has a idea about it ?


Do you see the same behavior when using a simple Sine Osc unit? Are you always feeding the 1V/oct signal to the calibrated inputs: ABCDx?


no problem with the built-in sources. Yes i used the calibrated inputs for the pitch and took the gate out from V8+ to G1. The problem seems to occure only using samples in 4 poly.


Have you got slew on the v8? Does it work ok if you put a delay on the trigger from the v8? so the cv deffo arrives before the gate?