<p>Implementations may support WebVTT as a text track format for subtitles, captions, chapter

titles, metadata, etc, for media resources. <ref spec=WEBVTT></p>

<li><dfn data-x-href="https://w3c.github.io/webvtt/#webvtt-file-using-chapter-title-text">WebVTT file using chapter title text</dfn></li>

<td><dfn data-x="attr-track-kind-chapters">Chapters</dfn>

<td>

Chapter titles, intended to be used for navigating the <span>media resource</span>.

Displayed as an interactive (potentially nested) list in the user agent's interface.

<td>

Tracks intended for use from script.

Not displayed by the user agent.

<p>Furthermore, if the element's <span>track URL</span> identifies a WebVTT resource, and the

element's <code data-x="attr-track-kind">kind</code> attribute is in the <span

data-x="attr-track-kind-chapters">chapters</span> state, then the WebVTT file must be both a

<span>WebVTT file using chapter title text</span> and a <span>WebVTT file using only nested

cues</span>. <ref spec=WEBVTT></p>

data-x="attr-track-kind-chapters">chapters</span> state and whose <code

<dt><dfn data-x="text track rules for extracting the chapter title">Rules for extracting the chapter title</dfn>

<dd>

<p>An algorithm which, when applied to the cue, returns a string that can be used in user

interfaces that use the cue as a chapter title.</p>

</dd>

<td><span data-x="attr-track-kind-chapters">Chapters</span>

<li><p><span>Perform automatic text track selection</span> for <code

data-x="dom-TextTrack-kind-chapters">chapters</code>.</p></li>

</div>

<h6>Text tracks describing chapters</h6>

<p>Chapters are segments of a <span>media resource</span> with a given title. Chapters can be

nested, in the same way that sections in a document outline can have subsections.</p>

<p>Each <span>text track cue</span> in a <span>text track</span> being used for describing

chapters has three key features: the <span>text track cue start time</span>, giving the start time

of the chapter, the <span>text track cue end time</span>, giving the end time of the chapter, and

the <span>text track rules for extracting the chapter title</span>.</p>

<div w-nodev>

<p>The <dfn>rules for constructing the chapter tree from a text track</dfn> are as follows. They

produce a potentially nested list of chapters, each of which have a start time, end time, title,

and a list of nested chapters. This algorithm discards cues that do not correctly nest within each

other, or that are out of order.</p>

<ol>

<li><p>Let <var>list</var> be a copy of the <span data-x="text track list of cues">list

of cues</span> of the <span>text track</span> being processed.</p></li>

<li><p>Remove from <var>list</var> any <span>text track cue</span> whose <span>text

track cue end time</span> is before its <span>text track cue start time</span>.</p></li>

<li><p>Let <var>output</var> be an empty list of chapters, where a chapter is a record

consisting of a start time, an end time, a title, and a (potentially empty) list of nested

chapters. For the purpose of this algorithm, each chapter also has a parent chapter.</p></li>

<li><p>Let <var>current chapter</var> be a stand-in chapter whose start time is negative

infinity, whose end time is positive infinity, and whose list of nested chapters is <var>output</var>. (This is just used to make the algorithm easier to describe.)</p></li>

<!-- while not empty... -->

<li><p><i>Loop</i>: If <var>list</var> is empty, jump to the step labeled

<i>end</i>.</p></li>

<!-- do... -->

<li><p>Let <var>current cue</var> be the first cue in <var>list</var>, and then

remove it from <var>list</var>.</p></li>

<li><p>If <var>current cue</var>'s <span>text track cue start time</span> is less than

the start time of <var>current chapter</var>, then return to the step labeled

<i>loop</i>.</p> <!-- out of order chapter -->

<li><p>While <var>current cue</var>'s <span>text track cue start time</span> is greater

than or equal to <var>current chapter</var>'s end time, let <var>current

chapter</var> be <var>current chapter</var>'s parent chapter.</p></li>

<li><p>If <var>current cue</var>'s <span>text track cue end time</span> is greater than

the end time of <var>current chapter</var>, then return to the step labeled

<i>loop</i>.</p> <!-- misnested chapter -->

<li>

<p>Create a new chapter <var>new chapter</var>, whose start time is <var>current cue</var>'s <span>text track cue start time</span>, whose end time is <var>current cue</var>'s <span>text track cue end time</span>, whose title is the result of running <var>current cue</var>'s <span>text track rules for extracting the chapter title</span>, and whose list of nested chapters is

empty.</p>

</li>

<li><p>Append <var>new chapter</var> to <var>current chapter</var>'s list of

nested chapters, and let <var>current chapter</var> be <var>new chapter</var>'s

parent.</p></li>

<li><p>Let <var>current chapter</var> be <var>new chapter</var>.</p></li>

<li><p>Return to the step labeled <i>loop</i>.</p></li>

<!-- ...end while -->

<li><p><i>End</i>: Return <var>output</var>.</p></li>

</ol>

</div>

<div class="example">

<p>The following snippet of a <span>WebVTT file</span> shows how nested chapters can be marked

up. The file describes three 50-minute chapters, "Astrophysics", "Computational Physics", and

"General Relativity". The first has three subchapters, the second has four, and the third has

two. <ref spec=WEBVTT></p>

<pre>WEBVTT

00:00:00.000 --> 00:50:00.000

Astrophysics

00:00:00.000 --> 00:10:00.000

Introduction to Astrophysics

00:10:00.000 --> 00:45:00.000

The Solar System

00:00:00.000 --> 00:10:00.000

Coursework Description

00:50:00.000 --> 01:40:00.000

Computational Physics

00:50:00.000 --> 00:55:00.000

Introduction to Programming

00:55:00.000 --> 01:30:00.000

Data Structures

01:30:00.000 --> 01:35:00.000

Answers to Last Exam

01:35:00.000 --> 01:40:00.000

Coursework Description

01:40:00.000 --> 02:30:00.000

General Relativity

01:40:00.000 --> 02:00:00.000

Tensor Algebra

02:00:00.000 --> 02:30:00.000

The General Relativistic Field Equations</pre>

</div>

<div w-nodev>

<p>For the purposes of listing chapters in the <span>media resource</span>, only <span data-x="text

track">text tracks</span> in the <span>media element</span>'s <span>list of text tracks</span>

that are <span data-x="text track showing">showing</span> and whose <span>text track kind</span> is

<code data-x="dom-TextTrack-kind-chapters">chapters</code> should be used. Such tracks must be

interpreted according to the <span>rules for constructing the chapter tree from a text

track</span>. When seeking in response to a user manipulating a chapter selection interface, user

agents should not use the <i>approximate-for-speed</i> flag.</p>

exposes metadata within the content such as subtitles or chapter titles. Such information is

therefore only exposed if the video resource uses CORS. The <code

data-x="attr-media-crossorigin">crossorigin</code> attribute allows authors to enable CORS. <ref

spec=FETCH></p>

<td> "<code data-x="attr-track-kind-subtitles">subtitles</code>";

"<code data-x="attr-track-kind-captions">captions</code>";

"<code data-x="attr-track-kind-descriptions">descriptions</code>";

"<code data-x="attr-track-kind-chapters">chapters</code>";

"<code data-x="attr-track-kind-metadata">metadata</code>"