Link search Menu Expand Document

How to talk COMPUTER

We’ve made it to part three without really explaining how commands work– I appreciate your blind trust. But now it is time to talk about syntax. This sounds possibly boring but it’s the key to understanding the great potential of the system, allowing us to make interesting and sophisticated interactions.

Before we begin, load up a new blank scene. We’ve described doing this in past studies (hint: hit ESC, navigate with ] until you find a blank scene, hit ENTER).

Right to Left

In many ways Teletype is just a fancy calculator. Syntax follows prefix notation (aka Polish notation)– the operator is to the left of the operands. Get into LIVE mode and try this:

ADD 2 8

You’ll see the result printed: 10. Here the operator is ADD, the operands (aka arguments) being 2 and 8. ADD wants two arguments (values). And why did it print? Teletype is designed to display the returned value, if one is returned.

ADD returns a value. Let’s use that fact to make a longer command:

ADD 2 ADD 4 8

What? It looks weird but it’ll feel better soon. Read right to left– find the rightmost operator and give it arguments. Pulled apart, first we have

ADD 4 8

which equals 12. Let’s mentally substitute that value into the remaining command:

ADD 2 (12)

which makes 14. Which should’ve been displayed as the result. Simple enough? Yeah! But things can get confusing:

ADD ADD 1 2 3

Yes, this works. Think of it like this:

ADD (ADD 1 2) 3

The rightmost operator uses the two closest arguments. Even this is legit:

ADD MUL 2 ADD 7 8 9

This decodes to ADD (MUL 2 (ADD 7 8)) 9 which is 39. Perhaps not the friendliest command at first glance, but not a monster after a closer look.

Break it down, save it for later

You’d possibly rather make sounds than deciphering strange looking strings of text. We can make commands easier to read by using variables.

X ADD 1 2

Instead of printing the value 3, X is now assigned 3. Type X alone on the command line and it’ll print its value.

KNOW THIS! → A variable (such as X) or parameter (such as CV 1) gets set if it is the leftmost WORD of a command. Otherwise the value is read and returned.

Consider that complicated command we had earlier: ADD ADD 1 2 3

Here’s a more readable way:

X ADD 1 2

X is set to the result of ADD 1 2. Then X gets read and returns 3 in the second line. The first line is effectively substituted into the second, breaking apart the command into multiple lines. Sometimes this is helpful– but you’ll also quickly run out of lines in a script. So there’s a balance to be found.

Real talk

  • Operators accept arguments.
  • Operators typically return values.
  • An argument is a value, so you can feed the returned value of an operator into another operator.
  • Variables can be assigned values, and read as arguments.
  • Parameters (CV, etc) can be read and used as arguments.

The takeaway: numbers are interchangeable. Once the flow of numbers makes sense, you’ll be able to put commands together in different ways to achieve a wide range of musical goals.


Basic arithmetic operators are ADD, SUB, MUL, DIV, and MOD. These all take two arguments.

Addition and multiplication are commutative– the order of the arguments don’t matter: ADD 1 3 and ADD 3 1 are the same. But this is not the case for the others:

  • SUB 4 2 reads 4 - 2
  • DIV 8 4 reads 8 / 4
  • MOD 7 2 reads 7 % 2

Reversing the arguments will mess with your calculation. And what the heck is mod? It’ll give you the remainder after a division. So MOD 7 2 is 1.

Short term memory

Variables are good for much more than simplifying commands. They store values which can then be manipulated in various ways.

Available variables: X, Y, Z, and T. T is typically used for time-based operations but can be used freely.

Variables can be both set and read in the same command. Consider this:

X 10

First X gets set to 10. In the second line the ADD reads X and adds 1, returning 11. So X gets set to 11. Push the up arrow to re-execute that last line. X is counting up by 1 each time.

Variables are global– they keep their value across scripts. For example X can be changed by script 1 and then read and further manipulated by script 2.

    X 0

    X ADD X 1

    Y N MUL X 2
    CV 1 Y

Here script 1 resets X to 0. Script 2 increments X by 1. Script 3 assigns CV output 1 to the note X multiplied by 2, using Y as a temp variable to break up the command (this could be easily combined, of course).

With various staggered triggering of these three scripts, you will likely find music!

More is more

You may feel that four variables just isn’t enough– c’mon, a four-note melody? (Enough for me most days.) Fortunately there is a whole different system for saving tons of numbers: patterns. We’ll cover this in a very-near-future part of this series.

In the meantime, if you’re really desperate for more variables– A, B, C, and D can be overwritten. By default these are initialized to 1-4 on startup.

As it happens, TR A it is exactly the same as TR 1. So if you overwrite A, be sure to use TR 1 instead.

Save save save

Be warned that variables are not stored with scenes. You can load a new scene and the variables will remain the same. If you want to have a scene recall with specific variable values you’ll need to use the INIT script. For example:

    X 8
    Y 22

Now we have some defined values for X and Y when the scene is loaded.

Lastly, variables aren’t saved when powering down. On power-up memory is cleared, but the INIT script is called on powerup which allows you to define startup values.

Here at the post office

A long time ago at the post office in Castaic, CA: the postmaster asked the lady ahead of us “How fast do you want this to ship?” There was a lot of confusion and shrugging and finally “I don’t care.” To which the postmaster responded resoundingly:

“Here at the post office we only deal in absolutes.”

Teletype is not the post office. (OK, weird transition, sorry.)

    CV 1 N RAND 12

We’ve arrived at the moment you’ve been waiting for: random semitones streaming out your modular synth.

RAND 12 will return a random number between 0 and 12.

Operators can have different numbers of arguments. The arithmetic operators so far have had two. RAND takes just one. This is important to remember when analyzing (and building) commands. Of course, the command sheet will help!

Let’s make this immediately more musical by creating a whole tone scale:

CV 1 N MUL 2 RAND 12

Now we’re creating a random number between 0 and 12, and multiply it by two. Recall that N does a note lookup for sending to CV outputs.

And what about this:


Yeah? Now we can manipulate the range (Y) and interval (X) from some other script!

Somewhere in between


A random value from 4 to 8 (inclusive) will be returned. RRAND (range random) takes two arguments.

Recall that a TR index can be set with 1-4 rather than A-D? So we can turn on TR A 1 with the identical command TR 1 1. But how about this:

TR RRAND 1 4 1
TR RRAND 1 4 0

The first line turns ON a random trigger output. The second line turns OFF a random trigger output. We can do the same with CV.

Infinite coins


This operator has no arguments! It returns a 0 or 1, randomly.


This command sets a random TR output to a random state, on or off.


Here we’re multiplying 5 by either zero or one, which gives us either zero or 5.

The musical qualities of stumbling


Appropriately, DRUNK isn’t quite normal. It’s a variable, but it changes by 1, 0, or -1 each time you read it. But you can also reset its position:


However, next time you read the value:


You may get 4, 5, or 6.

DRUNK does not have boundaries, so you may need to constrain it within a range to keep it useful:


Here we’re creating single-volt steps between 0 and 4. You might get something like this:

0 → 1 → 0 → 0 → 4 → 3 → 3 → 4 → 3 → 2 → 1 → 1 → 2

The MOD operator wraps the edges.


This scene is featured in the banner video above.

Like the previous study, we’re using two oscillators with frequencies controlled by CV outputs 1 and 2. Remember to tune them to the same note prior to plugging in CV.

  • Input 1 will randomly select a note for the first oscillator. TR out A will be pulsed.
  • Input 2 will choose between two notes for the second oscillator. I used this as a low root note.
  • Input 3 advances a drunk walk of single volts between 0 and 2.
  • Metro running at 1 second interval, randomly slewing CV output 4 between random voltages 0 to 5.

Values are initialized on startup.

Printable Blank Scene Template

Suggested explorations

Get into LIVE mode and try changing some variables:

X 2
Y 12
Z 12

X changes the note spread and Y the range for voice 1. Z the interval of the low tone for voice 2.



ADD x y         add x + y
SUB x y         subtract x - y
MUL x y         multiply x * y
DIV x y         divide x / y
MOD x y         modulus x % y (return remainder)

RAND x          return random value from 0 and x
RRAND x y       return random value from x to y
TOSS            return a random value 0 or 1

X,Y,Z,T         variables
A,B,C,D         variables, initialized to 1-4 on startup
DRUNK           variable that changes by -1, 0, or 1 when read
DRUNK x         set DRUNK value to x

Full Command Chart

Teletype Key Reference

You can also browse help within Teletype by using ALT-H to toggle help mode.

Teletype Studies Continued

Part 4: Collect and transform →

Part 3: Playing with numbers

Part 2: Curves and repetition ←

Part 1: Navigating and making edges ←