Teletype workflow, basics, and questions



Thanks for sharing, it looks fun :slight_smile: I’ll take a closer look at the code


Recently got my TT so still grappling with the ultra basic stuff - I’m trying to construct a backbeat with Mutable Plaits, switching between the Kick and Snare models with CV and trig from TT.

Can someone explain why I’m getting a result where sometimes the CV switch doesn’t take place about 10-15% of the time? The TR output sends signal as expected.

Can’t see the fault in my code, but maybe there’s something I’ve missed. I was playing about with probabilities before this, but I still get the same result after turning off and on. Also tried both triggers externally and with the keyboard, with no change.

Here’s my code:

EVERY 2: CV 1 VV 0


And if you add a small delay between the Cv change and the pulse ?

DEL 10: TR.P 1 ?

Edit: And is VV50 enough to jump to the next model ?


Ahhh… DEL 20 fixes it! Thanks!


Depending on the module you’re sending the cv / trigger this is often a good solution. If I’m correct it’s more often for digital modules handling both pitch and envelope generation.


5 posts were merged into an existing topic: Teletype 3.0


Hi all, so just finished the TT studies, it is quite a big lot to have to register in the head, but definitely enjoying the process. There is something that came up twice that I have no idea what it really means, if someone could please shed some light as to what it means “Smoothing”? When you avg out some numbers, its referred to as smoothing? Lost.


the analogy is really with graphs and/or filtering. The idea is this: when you’re generating random numbers, you might see big jumps in the value of the number you get, and for some applications, this big jump is undesirable. But, if you average the last two values you got, now suddenly your big jump is much smaller. If you were to graph this, the graph of the averaged values would appear to have more gentle changes than the raw values.

It turns out that if you want to implement a low-pass filter digitally, this same idea ends up being what you do: the filtering “smoothes out” the high frequency content (aka the spikiness or big jumps), giving you something which changes more slowly and gently.


I think what’s missing to make sense is over time. You’re not just averaging a batch of numbers, but averaging them over time. For example, if you sample the param knob every 100ms, you could by turning it quickly get big differences between numbers. And that mean big/harsh jumps that could be jarring musically.

By averaging the old value with the new value you are effectively reducing the gap.

For example, the param knob is at 0. You turn it to 100.
If it takes the value every 100ms and average previous and current value you will get:
0ms: 0
100ms: (0 + 100) / 2 -> 50
200ms: (50 + 100) / 2 -> 75
300ms: (75 + 100) / 2 -> 87
400ms: (87 + 100) / 2 -> 93

and so on.


Thanks @alanza and @electret for the replies and generous explanations!

I think I understand now!


Happy to have made the first patch with just type today! Really enjoying the TT!


Thinking about how cool the 16n Faderbank would be with Teletype, I realized I already have a sort of 6-slider faderbank in my rack… Stages.

From there I threw together a quick scene to poll and memorize a hands-on sequencer, with some scaling to make voltage ranges a little friendlier:

Patch sequencer output to TT cv input
Patch TR.1 output to sequencer step input
Patch TR.2 output to sequencer reset, if available (Stages doesn’t have it)

Patch trigger input 3 to clock, or use internal metronome.
Patch CV.3 and TR.3 outputs to whatever you’re controlling.

Set X to 0-3 to select the pattern
Set Y in the init script (or manually) to the sequencer’s length.
Run script 1 to perform polling

TR.TIME 1 10
X 0
Y 6

PN.L X 0
TR.P 2
DEL.R Y 20: $ 2

TR.P 1

TR.P 3

I originally wrote a continuous polling version, but it was silly because I could just use Stages directly. (And if I had a 16n, I would almost certainly read the values in real time instead of messing with storing them in a pattern. Unless that’s slow.)

This has some trouble trying to poll in a more automated way while playing back (say, every time PN.I X is 0). I believe if you use an explicit counter instead of PUSH that will fix it.


After having revisited the Just Type studies, Teletype has flown back onto my radar. With a burgeoning interest in coding (zero experience thus far) and the new edition just coming out, now seems like a great time to get involved.

Are there many people out there using Teletype as a small part of a larger system? To clarify, I have a 7U setup that I’m pretty happy with right now but Just Type seems to be along the lines of what I’ve been looking to get out of modular music for a while. JF is probably my favorite module and I’m really interested in the pair, particularly in terms of polyphony. I don’t know that I’d necessarily want it to be the foundation of my setup but it almost seems like a waste not to get one to pair with JF.

I guess to reduce my question down to its most basic form - is it crazy to consider getting TT only for JT stuff? I’m sure it would inevitably lead down a deeper Monome rabbit hole but that’s mostly all I’m interested in for now.


I started with TT which led me to the TXo’s and TXi’s along with Just Friends, Ansible and modding my ER-301 so everything could talk i2c. With so many voices talking in the background I stepped up for a Grid and an ARC. Add a W/ to top it all off, thinking about a second one. The rabbit hole is real, be aware.


I was intrigued by W/, and that led me to Teletype (as soon as I realized I could replace my two Euclidean modules with it and do a lot more besides). I didn’t keep W/, and don’t have a JF, but I do love TT.

I expect to go for a 16n Faderbank when the built ones are available. TXo and/or Crow are possibilities. Even an ER-301 someday, if I decide to replace a bunch of other stuff with it.


I typically use just type as a voice as part of a larger system—granted, where the other voices are controlled by cv/gate from teletype as well, tho that doesn’t necessarily need to be the case if you have other sequencers you are happy with. You could easily sync teletype as the leader or follower.

Note that if you happen to have a grid and a trilogy module or ansible, you can explore Just Type’s synth mode with the new polyearthsea firmware. You just need a powered i2c bus and a ribbon cable.


That’s good to hear. My modular stuff tends to be melodic & generative - I have Rene at the moment but I’m finding it to be a bit overkill for this case and end up using Marbles to generate melodies more often than not.

Can TT run multiple scripts simultaneously? E.g. could I use it to control JF internally and use the physical outs for something else?

Also thanks for the heads up about Polyearthsea, I didn’t realize there were other avenues to get into JF’s synth mode. I’ll have to look into that, Grid & Ansible could be a different option to explore for sure.


I had Teletype in a larger system a while back, but like it better in a small skiff and felt I was underutilizing it before in my 208 hp case. Other people have used it really effectively in larger set ups, in particular @laborcamp and @egon77 come to mind.

I’m much happier with my smaller set up with Teletype, driven by Walk or Ansible, and a set of Mannequins modules…feels like there is much to explore, especially with Just Type and W/ Type now.

Yes, definitely! Each of the 8 scripts can do whatever you want.


to answer your initial question - buying TT for JT isn’t crazy, but you’re almost definitely going to end up using it for lots of other stuff.

i’m in a similar boat, but reversed! 7u x 84hp, roughly 75% full. i was attracted to TT first because of how many different functions it could pack into my system (particularly sequencing, random, s&h, and JT once i got JF), but i was also a little wary of all the coding. i got JF roughly 6 months after TT and i’m glad i had time to learn the ropes before leveling up. the learning curve with TT is steep, but after you’ve done the studies once or twice and start coming up with ideas on your own, the functions start to stick in your mind. i definitely don’t fully understand a lot of stuff going on in the code exchange threads yet, but i know enough for TT to be a capable brain for my system.

it’s definitely less hands-on than either module, but for what it’s worth TT can definitely perform the functions of both marbles and rene.


Ah that does remind me, there was a learning hump for me personally, in which it felt more natural to use an “immediate” sequencer, and it took me putting other sequencers away for a bit to get comfortable and really start to gel with the teletype. In the sense of diving into it, it does require an investment to learn (at least it did for me personally)