Branches feedback

Hey people,

I know a few of you here have bought some of the branches modules I sold 2 weeks ago. I haven’t received any single piece of feedback, so I assume it’s all good to launch production, but if you have any comment, speak now or forever hold your peace.

The only comment I have is that it’s perfect, no issues whatsoever.

All in all it’s getting lots of use, and I’m always thinking of new ways of using it.
Another MI winner.


missed the fb post on that batch and am still kicking myself over it! should a unit become available, i will beat it as mercilessly as i’m hitting peaks :wink:

Presumably Branches is digital? In which case, I wonder if, as well as the Bernoulli gate and Markov chain modes, it couldn’t also just act as a clock divider? In other words, instead of a being a biased stochastic gate, it could act deterministically in a third mode and just count pulses (“one pulse to the left output, three to the right output, one to the left, three to the right and so on”. It could potentially offer very high levels of clock division, perhaps with mode three being integer clock division, mode four being multiples of 2 division, mode five being multiples of 3 division and so on (to accommodate many time signatures)? You only need clock division modes for multiples of the the first seven primes to cover all possible time signatures up to 19 (and many beyond that), I think. There’s a button to push for each half of Branches to set modes now, isn’t there? That should be enough of a user interface e.g. to set clock division mode, hold down button for 5 seconds and then tap button the required number of times up to 19 to set the lowest divisor.

Yes, it’s possible to implement a clock division feature in Branches - actually any gate/trigger processing function with two outputs controlled by a CV. It could be an euclidean sequencer too, a clock shuffler, a clock delay…


  • The CPU is the tiniest ATmega - 6kb or flash and clocked at 8 MHz.
  • There are already many clock dividers on the market, and their user interface is designed for this function.
  • I think I have done enough “crammed with barely usable secret functions” modules for now.

The module will be open-source, can be reprogrammed with an AVR programmer and has a tiny audio bootloader, so I’ll leave that as an exercise to the reader.

Yup, that seems like a very sound plan. There are very distinct limits to how many features can be crammed into 6kb of flash, and how many secret handshakes into user’s brains.

since branches is made for logic inputs only i’d love to have a module that has the same abilities but for audio ins and outs.

The particularity of Branches is that the trigger (the signal that triggers the random decision) is the same as the input signal to be dispatched to the two outputs.

If you want to design an “audio” version of Branches, you have two options:

  • Use a separate input to trigger the random decision logic. There’s actually an existing module from another brand doing exactly that - like Branches but with a switch IC. Anyone here remembers its name? Warning: this creates super clicky results on audio when the signal is abruptly cut and routed to the other output!
  • Keep the same layout as Branches, and use signal processing techniques to detect sharp transients on the input. Take advantage of the DSP power to gently fade out the signal before it is being cut -> EXPENSIVE. To me this is the kind of utility task that does not deserve the burden of an extra AD/DA conversion stage.

But what about patching this thing yourself using a VC-switch like the good old A-150 connected to an output of Branches? Or, if you want to avoid clicks, use a VC panner (or dual VCA) and slew-limit the output of Branches?

I found the panner yesterday during my research and i figured this being close to what i am searching for. Since i’m obviously a newb to modular synths i couldnt figure out to eg just patch branches->a134. Thanks again for your quick response (feel stupid once again :D).

Just installed Branches and I’m having trouble with Latch mode. It doesn’t seem to respond to every trigger, regardless of the pot’ setting. I’ve tried both triggers and gates from different sources (+5V) that work on other Branches modes, but not Latch. Is there anything special about Latch mode that I’m missing? I could try it in a different case with a different PSU later today, but I’m pulling 2.2A off a 2.5A supply so it should be OK. Any ideas on things to test? Thanks!

Solved. My bad/laziness for not spending more time with it before I posted. Fully CW it latches on every trigger on alternate outs. Every trigger/gate I can throw at it. Backing off from fully CW introduces randomness.

That is exactly the behavior I expected and wanted. Welcome to my world Branches!

The latch function can be enabled/disabled independently of the toggle function.