Home HelloWorld! Doc API Examples GitHub

Sequencer API

This describes the API of Sequencer.


The core idea is that programmers express temporal validity of objects by associating them with Intervals. For instance, a subtitle may be associated with [12.1, 17.44] indicating when the object should be active, in reference to some media timeline. Furthermore, in order to decouple Sequencers from the data model, Intervals are not directly associated with data objects, but indirectly through some unique identifier. For instance, programmers may use identifiers from an application specific data model as keys into a Sequencer.

A Sequencer manages a collection of (key,Interval) associations, also called SequencerCue objects. An Interval is simply the mathematical concept [12.75, 13.1>. Singular points [3.02] are considered a special case of Interval where length is 0. An Interval defines when a key is active or inactive, according to a time source.

Sequencers in the timingsrc library uses TimingObject as timing source. The main function of a Sequencer is to emit change and remove events at the correct time, as cue intervals dynamically transition between active and inactive. Change means that the cue has become active, or that it remained active while the cue data changed in some way. Remove means the cue became inactive, or was removed from the Sequencer all together. Sequencers also maintain a list of active cues, always consistent with the history of event callbacks and the state of the timing object. The Sequencer API is similar to the TrackElement API.

A Sequencer is data agnostic and may therefore be used by any application-specific data model, provided only that application data can be associated with unique keys, and that temporal aspects can be expressed in terms of intervals or singular points.


This documentation includes the following sections:


An Interval is expressed by two floating point values low, high, where low <= high. -Infinity or Infinity may be used to create un-bounded Intervals, e.g. [low, Infinity) or (-Infinity, high]. If low === high the Interval is said to represent a singular point [low].

Intervals may or may not include its endpoints; [a,b], [a,b>, <b,a], <a,b>. This is defined by optional boolean flags lowInclude and highInclude. If lowInclude and highInclude are omitted, [a,b> is the default setting. When multiple Intervals have the same endpoint, these endpoint flags influence event ordering. The Sequencer implementation also depends on this feature internally for correctness.

Interval objects are immutable.

Interval: Constructor

var i = new TIMINGSRC.Interval(low, high, lowInclude, highInclude);
  • param: {float} [low] value of lower endpoint of interval
  • param: {float} [high] value of higher endpoint of interval
  • param: optional {boolean} [lowInclude] lower endpoint included in interval : default true
  • param: optional {boolean} [highInclude] higher endpoint included in interval : default false
  • returns : {Interval} Interval object

Interval: Properties

var low = i.low,
    high = i.high,
    lowInclude = i.lowInclude,
    highInclude = i.highInclude,
    length = i.length;

Interval: Methods


  • returns: {string} string representation of the interval


  • returns: {boolean} true if (low === high)
    if (i.isSingular()) {}


  • param: {float} [point]
  • returns: {boolean} true if point is inside interval


  • param: {Interval} [otherInterval] another Interval
  • returns: {boolean} true if interval covers all points covered by other interval


  • param: {Interval} [otherInterval] another Interval
  • returns: {boolean} true if interval covers at least one point also covered by other interval


SequencerError is thrown by the Sequencer as a response to illegal input data. The SequencerError is defined as follows.

	var SequencerError = function (message) {
		this.name = "SequencerError";
		this.message = (message || "");
	SequencerError.prototype = Error.prototype;

Sequencer Cue

SequencerCue is a simple datatype used by Sequencer for query responses (and in some cases as parameter to event callback parameters). A SequencerCue is essentially an association between a key (string), an Interval, and a data object.

var cue = {
    key : "string",                  // unique string key
    interval : new Interval(12,13),  // interval object
    data : {}                        // javascript object

Sequencer EArg

SequencerEArg is a simple datatype used by the Sequencer as argument for event callbacks. SequencerEArg is a SequencerCue but includes additional properties relevant for specific event types.

var eArg = {
	// SequencerCue properties
    key : "string", 				 // unique key
    interval : new Interval(12,13),  // interval 
    data : {}, 						 // application object
    // additional properties
    src: object, 					 // reference to emitter of event, i.e. the Sequencer object
    point : 12.0, 					 // position of timing object when event was (should have been) triggered
    pointType : "low",				 // how point relates to the interval {"low"|"high"|"inside"|"outside"|"singular"}
    dueTs : 1441266518486, 			 // timestamp when the event should ideally be emitted - from performance.now()
    delay : 0.9,					 // lateness in milliseconds relative to dueTs
    directionType : "forwards",		 // direction of timingobject at point {"backwards"|"forwards"|"nodirection"}
    type : "change"				     // cue event type {"change","remove"}
    cause : "add"                    // operation type {"init", "add", "update", "remove", "motion"}
    enter: true                      // enter flag {true, false}
    exit: false                      // exit flag {true, false}


Sequencer: Constructor

Returns a Sequencer object. There is no need to start the Sequencer. Execution is driven by the given timing object, and the Sequencer is operational when the constructed finalizes.

var s = new TIMINGSRC.Sequencer(timingObject);
  • param: {object} [timingObject] The TimingObject that drives the execution of the Sequencer.

The Sequencer additionally supports window sequencing mode. To do window sequencing with the Sequencer, simply specify two timing objects in the constructor. Note that in window sequencing mode the Sequencer provides SequencerCues with events instead of EArg.

var s = new TIMINGSRC.Sequencer(timingObjectA, timingObjectB);
  • param: {object} [timingObjectA] Timing object A represents one endpoint of the active interval of the IntervaSequencer.
  • param: {object} [timingObjectB] Timing object B represents the other endpoint of the active interval of the IntervaSequencer.


The sequencer becomes ready when its timing object is ready and necessary initialization is done.

  • Event handlers may be registered before the sequencer is ready.
  • Cues may also be added before the sequencer is ready

The sequencer also defines a ready promise, although this is typically not important as handlers and cues may be added without regard for readiness.


  • returns : {boolean}
    // now the sequencer is ready


.addCue(key, interval, data)

  • param: {string} [key] unique key identifying an Interval.
  • param: {Interval} [interval] defines when the associated key is active.
  • param: {Object} [data] application object associated with key and interval
  • returns : {undefined}

Associate a unique key with an Interval and a data object. addCue() will replace any previous association for given key. Since Intervals are immutable objects, modification of a cue must be be done by generating a new Interval and replacing the association using .addCue() with the same key.

The keyspace is designed by the programmer. In this regard, the Sequencer is essentially an associative array for (Interval, data) tuples. Often, application specific datamodels include unique keys of some sort, and these may be used directly with the sequencer. These application specific keys are then reported back to application code by correctly timed Sequencer events. Intervals define when keys are active. So, when the current position of the timing object enters an Interval, the associated key becomes active.

s.addCue("key", new Interval(12.1, 24.22), {value:42});


  • param: {string} [key] unique key identifying (Interval, data).

Removes existing association (if any) between key and (Interval, data).


Batch processing

AddCue and removeCue operations are processed as a new task on the JavaScript event queue, so multiple operations may be processed in one batch. On the other hand, this means that the effects of addCue are not visible immediately after the operation.

cueList.forEach(function (cue) {
    s.addCue(cue.key, cue.interval, cue.data);

The Sequencer supports a number of efficient search operations on its collection of SequencerCues.


s.keys().forEach(function (key){});


  • param: {string} [key] unique key identifying an Interval.
  • returns: {boolean} True if Interval is defined for key
if (s.hasCue("key1")) {}


s.getCues().forEach(function (cue){});


  • param: {string} [key] unique key identifying an Interval.
  • returns: {object} SequencerCue if exists for key else null
var cue = s.getCue("key1");

Active Cues

The Sequencer maintains a list of Active SequencerCues. A SequencerCue is active if cue.interval.low <= timingobject.query().position <= cue.interval.high. In other words, if the position of the timing object is inside the Interval of a cue, that cue is said to be active. More generally, for timed media, the union of active cues may define the state of media, at any given point in time.






if (s.isActive("key1") {};


  • param: {float} [searchPoint] return all SequencerCues, where cue Interval cover given search point.
  • returns: {list} list of SequencerCues

The cost of this operation is linear O(N), with N being the number of SequencerCues managed by the Sequencer.



  • param: {Interval} [searchInterval] search Interval
  • returns: {list} list of all SequencerCues, where cue Interval overlaps with given search Interval.

The cost of this operation is logarithmic O(logN), with N being the number of SequencerCues managed by the Sequencer.

s.getCuesByInterval(new Interval(4.0, 8.0)).forEach(function(cue){});


  • param: {Interval} [searchInterval] search Interval
  • returns: {list} list of all SequencerCues, where cue Interval is covered by given search Interval.
s.getCuesCoveredByInterval(new Interval(4.0, 8.0)).forEach(function(cue){});

The cost of this operation is logarithmic O(logN), with N being the number of SequencerCues managed by the Sequencer.


The Sequencer supports three event types; "change"|"remove"|"events".

  • “change” : a cue was inactive and became active, or remained active.
  • “remove” : a cue was active and became inactive
  • “events” : a batch (list) of events including both “change” and “remove” events.

The programmer should likely choose to handle events in batch mode using “events”, or handle events individually using “change” and “remove” events.

initial events

The classical pattern for programming towards an event source typically involves two steps

  1. fetch the current state from the event source
  2. register event handlers for listening to subsequent changes to the state of the event source

The Sequencer event API simplifies this process for the programmer by delivering current state (active cues) as events on handler callback, just after an event handler is registered, but before any subsequent events. So, registering a handler or event types “change” or “events” will cause a batch of initial “change” events corresponding to active cues. This is equivalent to current state being empty initially, but then changing quickly. This implies that current state based on active cues can always be built the same way, through a single event handler. In this context, just after means that the events will be dispatched to the JaveScript task queue during .on() call, and consequently not be processed until after the current task has completed.

Read more about initial events in Initial Events Background.

enter and exit flags

Events include two boolean flags enter and exit.

  • enter : true if the cue was inactive (or non-existent) and became active, else false
  • exit : true if the cue was active and became inactive (or non-existent), else false

Event flags “enter” and “exit” relate to changes to the set of active cues. Enter and exit flags may both be false if a cue changed but remained active. Enter and exit flags can never both be true.


Events also expose the cause of the event {“init”, “add”, “update”, “remove”, “motion”}.

  • “init” : initial events corresponding to active cues
  • “add” : cue added to the sequencer
  • “update” : cue updated (i.e. replaced)
  • “remove” : cue removed from sequencer
  • “motion” : cue event due to motion of the timing object (e.g. playback).


These are the legal permutations expressed as tuples (type, cause, enter, exit)

  • (“change”, “init” , true, false) : initial events (active cues)
  • (“change”, “add” , true, false) : added cue became active
  • (“change”, “update”, true, false) : inactive cue updated and became active
  • (“change”, “update”, true, false) : inactive cue became active during playback
  • (“change”, “update”, false, false) : active cue updated and remained active
  • (“remove”, “update”, false, true ) : active cue updated and became inactive
  • (“remove”, “remove”, false, true ) : active cue removed
  • (“remove”, “motion”, false, true ) : active cue became inactive during playback

data changes

When using the sequencer, it is very convenient that all events relative to an active cue are made available through “change” and “remove” events of the sequencer. In particular, data changes on active cues often need to be reflected in visualization. This is achieved by exposing events for active cues that remain active (“change”, “update”, false, false). To exploit this feature, always refresh cues after data changes addCue(key, interval, data), whether the interval is changed or not.

singular points

Intervals that are singular points will emit both “enter” and “exit” events during playback. If the timing object is paused precisely within a singular Interval, only the “enter” event is emitted, just like a non-singular Interval. The “exit” event will then be emitted as the position is later changed.

event handler

var handler = function (e) {};

Events “change”, “remove” and “events” provide SequencerEArgs as event parameter.

.on(type, handler, ctx)

  • param: {string} [type] event type
  • param: {function} [handler] event handler
  • param: optional {object} [ctx] this === ctx in event handler, if ctxt is provided, else this === Sequencer instance.
this.handler = function (e) {};
// register callback
s.on("change", this.handler, this);

// callback invocation from sequencer
handler.call(ctx, e);

.off(type, handler)

  • param: {string} [type] event type
  • param: {function} [handler] event handler Remove handler from Sequencer.
var handler = function(e) {};
s.on("change", handler);
s.off("change", handler);

event delay

Note that event delay is not a direct measure of the timeliness of the Sequencer. This is because dueTs is derived from the timestamp of the underlying timing object. In particular, whenever a timing object connected to an online timing resource is updated, the effects will suffer network delay before clients (Sequencers) are notified. The Sequencer is aware of the distributed nature of the timing object, and takes such delays into account. In short, the Sequencer replays events that should ideally have been emitted earlier, were it not for the network delay. This effect can only be observed for a brief moment following “change” events from the timing object. In these cases, dueTs effectively means the time when the event would have been emitted if network delay was zero. This behavior of the Sequencer also ensures consistent behaviour between distributed Sequencers (provided that they are working on the same collection of timed data).

event ordering.

If multiple Interval endpoints are bound to the same point on the timeline, playback across this point implies that multiple events should be emitted at the same time. In this case, events will be emitted according to the following ordering, given that direction of the timing object is forwards. If direction is backwards, the ordering is reversed.

  • > exit Interval with > exit-endpoint
  • [ enter Interval with [ enter-endpoint
  • [ enter singular Interval
  • ] exit singular Interval
  • ] exit Interval with ] exit-endpoint
  • < enter Intervals with < enter-endpoint