Controls Js

Controls JS Library

Controls JS is a JavaScript library that makes it easier to code controller mappings for Mixxx. It lets you focus more on your mapping and less on the details of how Mixxx works. It is centered around JavaScript objects called Controls that represent a physical component on a controller, such as a button, knob, encoder, or fader. Control provides generic functions that can be made to work for most use cases just by changing some attributes of the Control, without having to write many or any custom functions. The library also provides more specialized Controls for common use cases. Controls can be organized into ControlContainer objects, making it easy to iterate over them and change their behavior to switch between different modes.

To use the library, in the <scriptfiles> element at the top of your mapping's XML file, load the Lodash library and the Controls library:

<file functionprefix="" filename="lodash.mixxx.js"/>
<file functionprefix="" filename="midi-controls-0.0.js"/>

Controls JS uses a few functions from Lodash, which is why they both need to be loaded. Importing the midi-controls-0.0.js file makes the library accessible by an object called control (lower case).

This documentation is a work in progress. Not every capability of the library is fully documented yet. Refer to the source code of the library if in doubt.

Control

A Control represents a physical component on a controller, such as a button, knob, encoder, or fader. It encapsulates all the information needed to receive MIDI input from that component and send MIDI signals out to the controller to activate its LED(s).

Controls should generally be properties of a #ControlContainer object. Most Controls should be properties of a custom #Deck object, which is a derivative of ControlContainer.

Setup

The input function of each Control needs to be mapped to the incoming MIDI signals in the XML file. For example:

<control>
    <group>[Channel1]</group>
    <!-- MyController.leftDeck would be an instance of a custom Deck. -->
    <key>MyController.leftDeck.quantizeButton.input</key>
    <status>0x90</status>
    <midino>0x01</midino>
    <options>
        <script-binding/>
    </options>
</control>

In the future Mixxx will be able to register MIDI inputs from JavaScript, so that will not be necessary. The output does not need to be mapped in XML. It is handled by the library in JavaScript.

Create Controls by calling the constructor with JavaScript's "new" keyword. The Control constructor takes a single argument. This is an options object containing properties that get merged with the Control when it is created, making it easy to customize the functionality of the Control. Most Controls need at least these properties defined:

• midi (array with 2 numbers): the first two MIDI bytes that the controller sends/receives when the physical component changes state
• group (string): the group that both the inCo and outCo manipulate, for example '[Channel1]' for deck 1
• inCo (string): the Mixxx ControlObject that this JavaScript Control manipulates when it receives a MIDI input signal
• outCo (string): when this Mixxx ControlObject changes value, the output function will be called

For example:

var quantizeButton = new controls.Button({
    midi: [0x91, 0x01],
    group: '[Channel1]'
    inCo: 'quantize',
    outCo: 'quantize',
});

Methods

The following methods (in JavaScript, methods are just object properties that happen to be functions) must be defined for every Control, but in most cases the defaults will work so you do not need to define them yourself:

• input: the function that receives MIDI input
• output: the function that gets called when outCo changes value. Typically this sends MIDI output to the controller to change the state of an LED, but it can do anything.
• connect: register output as the callback function that gets executed when the value of group, outCo changes. Implement a custom function if you need to connect callbacks for multiple Mixxx ControlObjects in one Control. Refer to the source code of SamplerButton.prototype.connect for an example.

The following methods are called by the default Control input and output methods, as well as the default input functions of #Button, #Pot, and #RingEncoder. If you do not need to implement complex custom behavior, overwrite these instead of the default input and output methods:

• inValueScale: takes the third byte of the incoming MIDI signal as its first argument and returns the value to set group, inCo to
• outValueScale: takes the value of group, outCo as its first argument and returns the third byte of the outgoing MIDI signal

Each Control also has these methods that you probably should not overwrite:

• disconnect: disconnect the output function from being called when group, outCo changes
• trigger: manually call output with the same arguments as if group, outCo had changed
• send: send a 3 byte (short) MIDI message out to the controller. The first two bytes of the MIDI message are specified by the midi property. The third MIDI byte is provided as the first argument to the send function.
• inGetParameter: returns the value of group, inCo normalized to a 0-1 scale
• inSetParameter: sets the value of group, inCo to the function's first argument, normalized to a 0-1 scale
• inGetValue: returns the value of group, inCo
• inSetValue: sets the value of group, inCo to the function's first argument
• inToggle: sets group, inCo to its inverse (0 if it is >0; 1 if it is 0)
• outGetParameter: returns the value of group, outCo normalized to a 0-1 scale
• outSetParameter: sets the value of group, outCo to the function's first argument, normalized to a 0-1 scale
• outGetValue: returns the value of group, outCo
• outSetValue: sets the value of group, outCo to the function's first argument
• outToggle: sets group, outCo to its inverse (0 if it is >0; 1 if it is 0)

Shift layers

Controls can be used to manage alternate behaviors in different conditions. The most common use case for this is for shift buttons. For that case, assign functions to the shift and unshift properties that manipulate the Control appropriately. If you ever need to check whether a Control is in a shifted state, set its boolean isShifted property in your shift/unshift functions (in most cases this is not necessary). In some cases, using the shift/unshift functions to change the Control's inCo, outCo, or group properties will be sufficient. Refer to the source code for #HotcueButton for an example. In more complex cases, overwriting the Control's input and output functions inside your shift/unshift functions may be required. Refer to #SamplerButton and #EffectUnit for examples. To avoid redundancy (like typing the name of the inCo both as the inCo property and in the unshift function), the Control constructor will automatically call the unshift function if it exists. The shift and unshift functions of #ControlContainer will call the corresponding function of all the Controls within it that have that function defined and will recursively decend into ControlContainers that are properties of the parent ControlContainer.

Optional properties

The following properties can be specified in the options object passed to the Control constructor to customize the Control's initialization. Changing their value after creating the Control does not do anything.

• outConnect (boolean, default true): whether to call connect in the constructor (assuming group and outCo were specified in the options object)
• outTrigger (boolean, default true): whether to call trigger in the constructor (assuming group and outCo were specified in the options object)

Some controllers send and receive two sets of MIDI signals for most physical components, one for when the shift button is pressed and one for when the shift button is not pressed. To avoid defining two Controls for every physical component of your controller, set the following options as appropriate:

• sendShifted (boolean, default false): whether to send a second, shifted MIDI message for every call to send
• shiftChannel (boolean, default false): whether the shifted MIDI message changes the MIDI channel (second nybble of the first byte of the MIDI signal)
• shiftControl (boolean, default false): whether the shifted MIDI message changes the MIDI control number (second byte) of the MIDI signal
• shiftOffset (number, default 0): how much to shift the MIDI channel or control number by

To avoid having to define those properties for every Control, you can change the properties of controls.Control.prototype in your controller's init function. For example:

controls.Control.prototype.shiftOffset = 3;
controls.Control.prototype.shiftChannel = true;
controls.Control.prototype.sendShifted = true;

Syntactic sugar

Controls JS provides more convenient shortcuts for common situations. If inCo and outCo are the same, you can specify 'co' in the options object for the constructor to set both inCo and outCo. For example:

var quantizeButton = new controls.Button({
    midi: [0x91, 0x01],
    group: '[Channel1]'
    co: 'quantize'
});

Setting the co property after calling the constructor will not automatically set inCo and outCo; you would need to do that manually if necessary.

Also, if a Control only needs its midi property specified for its constructor, this can be provided simply as an array without wrapping it in an object. For example:

var playButton = new controls.PlayButton([0x90 + channel, 0x0A]);

instead of

var playButton = new PlayButton({
    midi: [0x90 + channel, 0x0A]
});

To avoid typing out the group for the constructor of each Control, Controls that share a group can be part of a ControlContainer and the ControlContainer's #reconnectControls method can assign the group to all of them. Refer to the #Deck ControlContainer documentation for an example.

Button

A Button is a Control derivative for buttons/pads.

For example:

var quantize = new controls.Button({
    midi: [0x91, 0x01],
    group: '[Channel1]',
    co: 'quantize',
});

By default, the inCo is toggled only when the button is pressed. For buttons that activate an inCo only while they are held down, set the onlyOnPress property to false. For example:

var tempSlow = new controls.Button({
    midi: [0x91, 0x44],
    inCo: 'rate_temp_down',
    onlyOnPress: false,
});

The button's LED is sent the value of the on property when outCo > 0 and off when outCo <= 0. By default, on is 127 (0x7F) and off is 0. For buttons/pads with multicolor LEDs, you can change the color of the LED by defining the on and off properties to be the MIDI value to send for that state. For example, if the LED turns red when sent a MIDI value of 127 and blue when sent a value of 126:

MyController.padColors = {
    red: 127,
    blue: 126
};
MyController.quantize = new controls.Button({
    midi: [0x91, 0x01],
    group: '[Channel1]',
    co: 'quantize',
    on: MyController.padColors.red,
    off: MyController.padColors.blue,
});

Derivative Buttons are provided for many common use cases, documented in the subsections below. These make it easy to map those kinds of buttons without having to worry about particularities of Mixxx's ControlObjects that can make mapping them not so straightforward. The PlayButton, SyncButton, HotcueButton, and SamplerButton objects also provide alternate functionality for when a shift button is pressed.

By default, this works for controllers that send MIDI messages with a different 3rd byte of the MIDI message (value) to indicate the button being pressed/released, with the first two bytes (status and control) remaining the same for both press and release. If your controller sends separate MIDI note on/off messages with on indicated by the first nybble (hexadecimal digit) of the first (status) byte being 9 and note off with the first nybble being 8, in your script's init function, set Button.prototype.separateNoteOnOff to true and map both the note on and off messages in XML to the Button object's input property.

PlayButton

Default behavior: play/pause
Shift behavior: go to start of track and stop

LED behavior depends on cue mode selected by the user in the preferences Refer to the manual for details.

CueButton

Behavior depends on cue mode configured by the user in the preferences
Refer to the manual for details.

SyncButton

Default behavior: momentary sync without toggling sync lock
Shift behavior: toggle sync lock (master sync)

LoopToggleButton

Toggle a loop on/off

HotcueButton

Default behavior: set hotcue if it is not set. If it is set, jump to it.
Shift behavior: delete hotcue

The LED indicates whether the hotcue is set.

Pass the number of the hotcue as the number property of the options argument for the constructor. For example:

var hotcues = [];
for (var i = 1; i <= 8; i++) {
    hotcues[i] = new controls.HotcueButton({
        number: i,
        group: '[Channel1]',
        midi: [0x91, 0x26 + i],
    });
}

SamplerButton

Default behavior: Press the button to load the track selected in the library into an empty sampler. Press a loaded sampler to play it from its cue point. Press again while playing to jump back to the cue point.
Shift behavior: If the sampler is playing, stop it. If the sampler is stopped, eject it.

Specify the sampler number as the number property of the object passed to the constructor. There is no need to manually specify the group. For example:

var samplerButtons = [];
for (var n = 1; n <= 8; n++) {
    samplerButtons[n] = new controls.SamplerButton({
        number: n,
        midi: [0x91, 0x02],
    });
)};

When the sampler is loaded, the LED will be set to the value of the "on" property. When the sampler is empty, the LED will be set to the value of the "off" property. These are inherited from Button.prototype if they are not manually specified. If your controller's pads have multicolor LEDs, specify the value to send for a different LED color with the playing property to set the LED to a different color while the sampler is playing. For example:

MyController.padColors = {
// These values are just examples, consult the MIDI documentation from your controller's
manufacturer to find the values for your controller. If that information is not available,
guess and check to find the values.
    red: 125,
    blue: 126,
    purple: 127,
    off: 0
};
var samplerButton = [];
var samplerButton[1] = new controls.SamplerButton(
    midi: [0x91, 0x02],
    number: 1,
    on: MyController.padColors.blue,
    playing: MyController.padColors.red,
    // off is inherited from Button.prototype
)};

Pot

A Pot is a Control for potentiometers (faders and knobs) with finite ranges, although it can be adapted for infintely turning encoders. Using a Pot Control is helpful because Pot.connect() and Pot.disconnect() take care of soft takeover when switching layers with ControlContainer.reconnectControls() and ControlContainer.applyLayer(). Soft takeover is not activated until the first input is received so it does not interfere with setting initial values for controllers that can report that information.

To adapt a Pot for an infinitely rotating encoder, replace its inValueScale() function with a function that increments or decrements the parameter depending on the direction the encoder is turned. For example, if the encoder sends a MIDI value of 1 for a left turn and 127 for a right turn:

MyController.SomePot.inValueScale = function (value) {
    if (value === 1) {
        return this.getParameterIn() - .05;
    } else if (value === 127) {
        return this.getParameterIn() + .05;
    }
}

RingEncoder

RingEncoder is a Control for encoders with LED rings around them. These are different from Pots because they are sent MIDI messages to keep their LED rings in sync with the state of Mixxx and do not require soft takeover.

These encoders can often be pushed like a button. Usually, it is best to use a separate Button Control to handle the MIDI signals from pushing it.

ControlContainer

A ControlContainer is an object that contains Controls as properties, with methods to help iterate over those Controls. Documentation for each method is below.

forEachControl

Iterate over all Controls in this ControlContainer and perform an operation on them.

Function arguments:

1. operation, function that takes 1 argument: the function to call for each Control. Takes each Control as its first argument. "this" in the context of the function refers to the ControlContainer.
2. recursive, boolean, optional: whether to call forEachControl recursively for each ControlContainer within this ControlContainer. Defaults to true if ommitted.

reconnectControls

Disconnect and reconnect output callbacks for each Control. Optionally perform an operation on each Control between disconnecting and reconnecting the output callbacks. Arguments are the same as #forEachControl.

shift

Call each Control's shift() function if it exists. This iterates recursively on any Controls in ControlContainers that are properties of this, so there is no need to call shift() on each child ControlContainer. This function takes no arguments.

unshift

Call each Control's unshift() function if it exists. This iterates recursively on any Controls in ControlContainers that are properties of this, so there is no need to call unshift() on each child ControlContainer. This function takes no arguments.

applyLayer

Activate a new layer of functionality. Layers are merely objects with properties to overwrite the properties of the Controls within this ControlContainer. Layer objects are deeply merged. If a new layer does not define a property for a Control, the Control's old property will be retained.

In the most common case, for providing alternate functionality when a shift button is pressed, using applyLayer() is likely overcomplicated and may be slow. Use shift()/unshift() instead. ''applyLayer() ''may be useful for cycling through more than two alternate layers.

For example:

someControlContainer.applyLayer({
    someButton: { inCo: 'alternate inCo' },
    anotherButton: { outCo: 'alternate outCo' }
});

By default, the old layer's output callbacks are disconnected and the new layer's output callbacks are connected. To avoid this behavior, which would be desirable if you are not changing any output functionality, pass false as the second argument to applyLayer().

Deck

Deck is a ControlContainer with methods for conveniently changing the group attributes of contained Controls to switch the deck that a set of Controls is manipulating. The setCurrentDeck() method takes the new deck as a string and sets the Controls' group property appropriately, including for equalizer knobs and QuickEffect (filter) knobs.

The Deck constructor takes one argument, which is an array of deck numbers to cycle through with the toggle() method. Typically this will be [1, 3] or [2, 4].

To map your own controller, create a custom derivative of Deck and create instances of your custom Deck objects in your controller's init() function. Use a constructor function to create all the Controls you need for your particular controller and assign your custom derivative's prototype to Deck. For example:

MyController.init = function () {
    this.leftDeck = new MyController.Deck([1, 2]);
    this.rightDeck = new MyController.Deck([2, 4]);
};
MyController.Deck = function (deckNumbers, midiChannel) {
    // Call the Deck constructor to setup the currentDeck and deckNumbers properties.
    controls.Deck.call(this, deckNumbers);
    this.playButton = new PlayButton([0x90 + midiChannel, 0x01]);
    this.CueButton = new CueButton([0x90 + midiChannel, 0x02]);
    this.hotcueButtons = [];
    for (var i = 1; i <= 8; i++) {
        this.hotcueButtons[i] = new HotcueButton({
            midi: [0x90 + midiChannel, 0x10 + i],
            number: i
        });
    }
    // ... define as many other Controls as necessary ...

    // Set the group properties of the above Controls and connect their output callback functions
    // Without this, the group property for each Control would have to be specified to its
    // constructor.
    this.reconnectControls(function (c) {
        if (c.group === undefined) {
            // 'this' inside a function passed to reconnectControls refers to the ControlContainer.
            c.group = this.currentDeck;
        }
    });
};
MyController.Deck.prototype = new controls.Deck();

EffectUnit

This ControlContainer provides Controls designed to be mapped to the common arrangement of 4 knobs and 4 buttons for controlling effects. 3 knobs are used for controlling effect metaknobs or parameters, depending on whether the effects' parameters are shown. The other knob is used for the dry/wet knob of the whole chain or the superknob when shift is pressed. 3 buttons are used for enabling effects and the other button toggles the effect unit between hiding and showing effect parameters. The Controls provided are:

• dryWetKnob (#Pot)
• showParametersButton (#Button)
• enableButtons[1-3] (#ControlContainer of #Buttons)
• knobs[1-3] (#ControlContainer of #Pots)
• enableOnChannelButtons (#ControlContainer of #Buttons)

When the effect unit is showing the metaknobs of the effects but not each parameter, the knobs control the metaknobs. The enableButtons control whether each effect is enabled. Pressing an enableButton with shift switches to the next available effect.

When the effect unit is showing all the parameters, the knobs behave differently depending on whether an effect is focused. When there is no focused effect (the default state), the knobs control the effect metaknobs like they do when parameters are not showing. When an effect is focused, the knobs control the first 3 parameters of the focused effect. An effect can be focused by pressing shift + its enableButton or clicking the focus button on screen. Pressing shift + the enableButton for the focused effect again unfocuses the effect.

The enableOnChannelButtons allow assigning the effect unit to different channels and are named after the Mixxx channel they affect. Not all controllers have buttons to map these. The following Buttons are provided by default:

• Channel1
• Channel2
• Channel3
• Channel4
• Headphones
• Master
• Microphone
• Auxiliary1

You can easily add more, for example for additional microphones, auxiliary inputs, or samplers by calling enableOnChannelButtons.addButton('CHANNEL_NAME') (do not put brackets around the CHANNEL_NAME).

Setup

To map an EffectUnit for your controller, call the constructor with the unit number of the effect unit as the only argument. Then, set the midi attributes for the showParametersButton, enableButtons[1-3], and optionally enableOnChannelButtons. After the midi attributes are set up, call EffectUnit.init() to set up the output callbacks. For example:

MyController.effectUnit = new controls.EffectUnit(1);
MyController.effectUnit.enableButtons[1].midi = [0x90, 0x01];
MyController.effectUnit.enableButtons[2].midi = [0x90, 0x02];
MyController.effectUnit.enableButtons[3].midi = [0x90, 0x03];
MyController.effectUnit.knobs[1].midi = [0xB0, 0x01];
MyController.effectUnit.knobs[2].midi = [0xB0, 0x02];
MyController.effectUnit.knobs[3].midi = [0xB0, 0x03];
MyController.effectUnit.dryWetKnob.midi = [0xB0, 0x04];
MyController.effectUnit.showParametersButton.midi = [0x90, 0x04];
MyController.effectUnit.enableOnChannelButtons.Channel1 = [0x90, 0x05];
MyController.effectUnit.enableOnChannelButtons.Channel2 = [0x90, 0x06];
MyController.effectUnit.init();

Controllers designed for Serato and Rekordbox often have an encoder instead of a dry/wet knob (labeled "Beats" for Serato or "Release FX" for Rekordbox) and a button labeled "Tap". If the encoder sends a MIDI signal when pushed, it is recommended to map the encoder push to the EffectUnit's showParametersButton, otherwise map that to the "Tap" button. To use the dryWetKnob Pot with an encoder, replace its inValueScale() function with a function that can appropriately handle the signals sent by your controller. Refer to the #Pot documentation for an example.

For the shift functionality to work, the shift button of your controller must be mapped to a function that calls the shift()/unshift() functions of the EffectUnit on button press/release. If the EffectUnit is a property of another ControlContainer (for example a Deck), calling shift() and unshift() on the parent ControlContainer will recursively call it on the EffectUnit too (just like it will for any other ControlContainer).