intermediate-language.md

Types

The following are the types of the Sema intermediate representation

@lang

This is the top level node of the tree, and contains an array of branches

{ "@lang" : [branches]}

@spawn

Execute a branch of a tree

{ "@spawn":<branch>}

@num

{"@num":{value:val}}

@str

{"@string":val}

@setvar

Set a variable.

{"@setvar": {"@varname":<string>,"@varvalue":value}};

@getvar

Get a variable

{"@getvar":<string>}

@sigp

@sigp represents a signal processor or signal generation. It looks like this:

{"@sigp": {"@params":[params], "@func":<string>}}

It needs a function name, and an array of parameters. You can use any of the options below:

Audio inputs

adc

Get a signal from the system default input Parameters:

1. Gain

Audio outputs

dac

Output a signal to the system default audio interface

Parameters:

1. Signal
2. (optional) Channel number (starting from 0)

If parameter 2 is to given, then the signal is copied to all the outputs

Oscillators

saw

A saw oscillator Parameters:

1. Frequency (Hz)
2. Initial Phase (0 - 1)

sin

A sinewave oscillator Parameters:

1. Frequency (Hz)
2. Initial Phase (0 - 1)

tri

A triangle wave oscillator Parameters:

1. Frequency (Hz)
2. Initial Phase (0 - 1)

sqr

A square wave oscillator Parameters:

1. Frequency (Hz)
2. Initial Phase (0 - 1)

pha

A phasor, rising from 0 to 1 Parameters:

1. Frequency (Hz)
2. Initial Phase (0 - 1)

ph2

A phasor with configurable start and end levels Parameters:

1. Frequency (Hz)
2. Start level
3. End level
4. Initial Phase (0 - 1)

pul

A pulse oscillator with modulatable phase width Parameters:

1. Frequency (Hz)
2. Phase width (0-1)
3. Initial Phase (0 - 1)

imp

An impulse generator Parameters:

1. Frequency (Hz)
2. Initial Phase (0 - 1)

sawn

An band limited saw wave oscillator Parameters:

1. Frequency (Hz)
2. Initial Phase (0 - 1)

noiz

An white noise generator Parameters:

1. Amplitude

Sampling

sampler

Creates a sampler with a signal input, the sample plays when the input has a positive zero crossing

1. Input signal
2. Sample name

loop

Creates a sampler that plays in a continuous loop

1. Speed
2. Sample name

slice

Slice up a sample

1. Trigger, on which to set the sample position
2. The sample position to set when a trigger is received (0-1)
3. The sample name

sah

Sample and hold

1. Input signal
2. Hold time (ms)

Math Operations

gt

Outputs 1 if $A > B$, otherwise 0 Parameters:

1. A
2. B

lt

Outputs 1 if $A < B$, otherwise 0 Parameters:

1. A
2. B

mod

A modulo B Parameters:

1. A
2. B

add

$A + B$ Parameters:

1. A
2. B

sub

$A - B$ Parameters:

1. A
2. B

mul

$A * B$ Parameters:

1. A
2. B

div

$A / B$ Parameters:

1. A
2. B

pow

$A ^ B$ Parameters:

1. A
2. B

abs

The absolute value of A Parameters:

1. A

sum

Sums all parameters $\sum(x_1, x_2 ... x_n)$

prod

Product of all parameters $\prod(x_1, x_2 ... x_n)$

mix

Mean of all parameters $\frac{\sum(x_1, x_2 ... x_n)}{n}$

Modulation

env

ADSR envelope generator Parameters:

1. Trigger
2. Attack (ms)
3. Decay (ms)
4. Sustain (0-1)
5. Release (ms)

line

Triggered line generator Parameters:

1. Trigger
2. Time (ms) to rise from 0 to 1

Mapping

blin

Map input into a linear range, assuming bipolar source (between -1 and 1) Parameters:

1. Input signal
2. Lower bound of destination range
3. Upper bound of destination range

bexp

Map input into an exponential range, assuming bipolar source (between -1 and 1) Parameters:

1. Input signal
2. Lower bound of destination range
3. Upper bound of destination range

ulin

Map input into a linear range, assuming unipolar source (between 0 and 1) Parameters:

1. Input signal
2. Lower bound of destination range
3. Upper bound of destination range

uexp

Map input into an exponential range, assuming unipolar source (between 0 and 1) Parameters:

1. Input signal
2. Lower bound of destination range
3. Upper bound of destination range

linlin

Map input into a linear range, specifying the range of the source Parameters:

1. Input signal
2. Lower bound of source range
3. Upper bound of source range
4. Lower bound of destination range
5. Upper bound of destination range

linexp

Map input into an exponential range, specifying the range of the source Parameters:

1. Input signal
2. Lower bound of source range
3. Upper bound of source range
4. Lower bound of destination range
5. Upper bound of destination range

Effects

dist

Tanh distortion

1. Input signal
2. Distortion level (0 upwards)

asymclip

Asymmetric wave shaping

1. Input signal
2. The curve shape for values below zero (e.g. 2 = squared, 3 = cubed, 0.5 = square root)
3. The curve shape for values above zero

dl

Delay

1. Input signal
2. Delay time (in samples)
3. Feedback

flange

Flanger

1. Input signal
2. Delay (ms)
3. Feedback (0-1)
4. Speed (Hz)
5. Depth (0-1)

chorus

Flanger

1. Input signal
2. Delay (ms)
3. Feedback (0-1)
4. Speed (Hz)
5. Depth (0-1)

freeverb

Reverb Arguments:

1. Input signal
2. Room size (0-1)
3. Absorption (0-1)

Filters

lpf

One pole lowpass filter

1. Input signal
2. Filter amount (0-1)

hpf

One pole highpass filter

1. Input signal
2. Filter amount (0-1)

lpz

Resonant lowpass filter

1. Input signal
2. Filter frequency (Hz)
3. Resonance

hpz

Resonant lowpass filter

1. Input signal
2. Filter frequency (Hz)
3. Resonance

svf

State variable filter

1. Input signal
2. Cutoff frequency (Hz)
3. Resonance
4. Low pass filter amount (0-1)
5. Band pass filter amount (0-1)
6. High pass filter amount (0-1)
7. Notch filter amount (0-1)

Machine Learning

toJS

Creates a transducer for sending a signal to a javascript model

1. A trigger, on which to send data
2. The channel number
3. Data value (this can be a list or a number)
4. The block size of the data

fromJS

Creates a transducer for receiving a signal from a javascript model

1. Channel (a number)
2. A trigger, indicating when to read from the channel. A zero here indicates that the channel will be read whenever data is available.

These functions are paired with 'input' and 'output' in the machine learning window

Clock functions

quantise

Choose when to bring in new code changes

1. 1 = at the start of the next bar, 0 = immeadiately

clp

Phasor derived from constantly running clock phasor

1. Number of cycles per bar
2. Phase offset (0-1)

clt

Trigger derived from constantly running clock phasor

1. Number of triggers per bar
2. Phase offset (0-1)

clk

Set the clock

1. Beats per minute
2. The number of beats in a bar

Triggers

onzx

Create a trigger on a zero crossing

1. A signal

onchange

Create a trigger when a change occurs in the input signal

1. A signal
2. Tolerance (a trigger will be generated if the change is more than +/- this value)

count

Counts up when receiving a trigger

1. Input trigger
2. Reset trigger

idx

Index into a list

1. Trigger input - output a value when triggered
2. The index of the value to be output when a trigger is received (normalised to between 0 and 1)
3. A list of values

Mouse

mouseX

Mouse X coordinate in the sema client window, normalised to between 0 and 1

mouseY

Mouse Y coordinate in the sema client window, normalised to between 0 and 1

Helper functions

Some simple helper functions are available within the scope of the grammar scripts to aid in putting together trees in a more readable way:

// create the tree structure for a number
function num(val) {
	return { "@num": { value: val } };
}

// create the tree structure for a string - useful for naming samples
function str(val) {
	return { "@string": val };
}

// create the tree structure for a DSP function
function synth(functionName, params) {
	let branch = {
		"@sigp": { "@params": params, "@func": { value: functionName } }
	};
	return branch;
}

// create the tree structure for setting a variable
function setvar(name, value) {
	return { "@setvar": { "@varname": name, "@varvalue": value } };
}

// create the tree structure for reading a variable
function getvar(name) {
	return { "@getvar": name };
}

You can create your own helper functions in the first portion of the grammar between @{% and %}.