Captions and Subtitles

Captions and subtitles ensure accessibility and localization for live and on-demand content. You can preserve, convert, or embed them based on your encoding profile and format requirements.

Implement captions and subtitles by including:

• CEA-608/708

  • By default, CEA-608 data is preserved. You may also choose to convert it to WebVTT.
  • A CEA-708 channel includes CEA-608 captions for backwards compatibility.

• Embedded Plain Text Captions/Subtitles

  • VOD Only
  • Embedded plain text captions or subtitles (e.g., SRT subtitles) are converted to WebVTT.

• Timed Text Markup Language (TTML)

  • VOD Only
  • A sidecar TTML file is converted to both CEA-608 and WebVTT.
  • Fragmented TTML is supported for DASH VOD.

• Scenarist Closed Captions (SCC)

  • VOD Only
  • A sidecar SCC file is converted to CEA-608.

📘

CEA-608 only supports a limited number of tracks and has limited support for non-Latin characters, while WebVTT supports additional subtitle/caption tracks and provides full Unicode character support.

Device Support

Device support for CEA-608 and WebVTT is indicated below.

Embedded CEA-608/708 Data

CEA-608 data is preserved when embedded within a container format.

Key Information:

• CEA-608 tracks may be defined via up to 4 channels (i.e., cc1, cc2, cc3, and cc4). Each channel should be marked as either data or subtitles.

• Automatically convert a CEA-608/708 channel to WebVTT via the render_608 and render_608_buffer parameters.

  Learn more about Live Streaming..

  Learn more about VOD.

• SDI only: Configure the Live Slicer to look for CEA-608 data in the horizontal ancillary data of an SDI signal via either the ancillary_lines or the ancillary_scan parameter.

• CEA-608 data is embedded within an H.264/H.265 stream as Supplemental Enhancement Information (SEI) messages within a Network Abstraction Layer (NAL).

Embedded Plain Text Captions/Subtitles

VOD Only

Embedded plain text captions and subtitles (e.g., SRT subtitles) are converted to WebVTT when a file contains a track marked as subtitle data.

Key information:

• Subtitle data may contain Unicode characters in UTF-8 format.
• Multiple subtitle tracks may be converted to WebVTT.
• Subtitle data is not converted to CEA-608.

Subtitle Language Tracks (DASH VOD)

By default, a viewer may select from any of the subtitle language tracks associated with your VOD content. Your asset's language tracks are available even when viewing ad breaks that do not contain those language tracks. This allows the viewer to select a subtitle language track associated with the requested content at any time during the playback session. For example, a viewer can immediately select the desired subtitle language track upon requesting your content, regardless of whether they are viewing a pre-roll ad.

Sample Scenarios

If the viewer is currently watching your main content, they can only select from the subtitle language track(s) incorporated into that asset. However, if a viewer is currently watching an ad, the set of available subtitle language tracks varies according to those present in the asset and the ad currently being played back.

The following table shows how the language tracks present within an asset and the currently playing ad affect the available subtitle language tracks.

Override this behavior by setting one of the following query string parameters within the playback URL:

Enable WebVTT for HLS Live Streams

Leverage WebVTT tracks when streaming a live event or live channel via HLS by adding the following query string parameter to the playback URL:

show_vtt=1

Key information:

• This parameter is only required when streaming a live event or live channel via HLS. Streaming VOD via HLS displays subtitles regardless of whether this parameter is present.
• Playback may falter, halt, or display unexpected results when using a different subtitle track. This may occur when subtitle track order is switched or when content does not have subtitles.
• Ads and system slate do not have subtitle tracks. Therefore, do not enable WebVTT if your HLS live stream contains ads or system slate. Only enable WebVTT when your HLS live stream does not have ads and all slate has been sliced with subtitle tracks.
• Subtitles are not displayed if a viewer starts playback with content that does not have subtitle tracks and then switches to content that does have subtitle tracks within the same playback session.

Digital Video Broadcasting (DVB) Teletext for Live Streams

Add closed captions from one or more DVB teletext pages by converting them to WebVTT. Convert DVB teletext to WebVTT by defining the render_teletext parameter in your Live Slicer configuration file.

Key information:

• The system only processes the first DVB teletext stream that it encounters.
• Ensure that your DVB teletext stream assigns a unique name to each language.
• If the teletext stream contains multiple language tracks, use a comma to delimit each desired combination of page number and language.

Timed Text Markup Language (TTML)

VOD Only

Captions and subtitles within a sidecar TTML file are converted to both CEA-608 and WebVTT whenever possible.

Key information:

• Multiple TTML files can be provided for multiple subtitle tracks. Alternatively, a TTML file may contain a div tag for each desired subtitle track.
• Use the ttml option to define the location of one or more language-specific sidecar TTML file(s). The Slicer replaces embedded captions or subtitles with the data defined within the TTML file(s).
• TTML files consumed by the Slicer must be in UTF-8 format.
• The Slicer supports a subset of the TTML standard as indicated below.

TTML Layout and Formatting

Ensure that your TTML files conform to the following layout:

<?xml version="1.0" encoding="utf-8"?>
<tt xml:lang="en" xmlns="http://www.w3.org/2006/04/ttaf1" xmlns:tts="http://www.w3.org/2006/04/ttaf1#styling">
<head>
	<styling>
		<style id="defaultCaption" tts:fontStyle="normal" tts:textAlign="center" />
			...
	</styling>
</head>
<body style="defaultCaption" id="thebody">
	<div xml:lang="en">
		<p begin="00:00:01.420" end="00:00:02.620"><metadata ccrow="14" cccol="3" /><span tts:fontStyle="italic">THIS IS SOME TEXT...</span></p>
		...
	</div>
	...
</body>
</tt>

• Style Attributes

  The following attributes may be applied to style and p tags:

  Attribute	Supported Values
  tts:fontStyle	"normal" and "italic"
  tts:textAlign	"left", "right" and "center"

• Default Style Attributes

  The following attribute may be applied to body, div, and p tags:

  Attribute	Supported Values
  style	ID of a style to use as the default style
  tts:textAlign	"left", "right" and "center"

• Language Code

  The following attribute may be applied to div tags:

  Attribute	Supported Values
  xml:lang	The RFC 5646 language code for the track

• Timing

  p tags must contain begin and end attributes to indicate when captions/subtitles should be shown.

  Attribute	Supported Values
  begin	hh:mm:ss.sss formatted timestamp
  end	hh:mm:ss.sss formatted timestamp

• Spans

  span tags may be used to set the tts:fontStyle attribute for a portion of text within a p element.

  Example:

  XML

  <span tts:fontStyle="italic">THIS IS SOME TEXT...</span>

• Positioning

  p tags may contain a metadata child tag to indicate the CEA-608 column and row position.

  Attribute	Supported Values
  ccrow	0-based row number. Must be 0-3 or 11-14
  cccol	0-based column number. Must be 0-31

CEA-608 Conversion

All TTML tracks are converted to WebVTT format, but only one track is converted to CEA-608. The Slicer chooses the track with the fewest characters that cannot be converted to the CEA-608 default character set.

Key information:

• CEA-608 captions are restricted to 32 columns. Any captions that extend past column 31 do not render correctly. The Slicer does not auto-wrap when converting to CEA-608.

• CEA-608 captions are restricted to 15 rows. Any captions that extend past row 15 do not render correctly.

• CEA-608 line numbers are 1-based, while TTML row numbers are 0-based.

  Example: TTML row 14 refers to CEA-608 line 15.

Fragmented TTML for DASH VOD

The Slicer outputs a fragmented TTML for DASH VOD for a TTML document or a DVB-TTML stream when the following conditions are met:

• Pass one of the following set of options to the Slicer:

  • TTML Document:

    -ttml {Path} -output_ttml

  • DVB-TTML Stream:

    render_dvb_ttml {PID 1}:{Language 1},{PID 2}:{Language 2},{PID n}:{Language n}

• Set show_dash_subtitles=imsc within the playback URL's query string.

Sample manifest request:

https://content.uplynk.com/1234567890abcdefghijklmnopqrstuv.mpd?show_dash_subtitles=imsc

Scenarist Closed Captions (SCC)

VOD Only

Captions and subtitles within a sidecar SCC file are converted to CEA-608. They are also converted to WebVTT when you pass the render_608 parameter. Use the scc parameter to define the location of a sidecar SCC file. The Slicer replaces embedded captions or subtitles with the data defined within the SCC file.

WebVTT Codec Initialization for DASH Live Streaming

The WebVTT codec is automatically initialized for VOD playback of an asset that uses that codec. This occurs regardless of whether a pre-roll ad contains that codec.

The WebVTT codec may not be initialized when the initial request is for a pre-roll ad that does not contain the WebVTT codec. This prevents the codec from working within that playback session.

Most DASH players initialize the WebVTT codec at the start of playback. This poses a challenge when the initial request is for a pre-roll ad that does not contain the WebVTT codec. If the player does not detect the WebVTT codec at the start of playback, it may not initialize it regardless of whether the main content uses it.

Ensure that the WebVTT codec is always initialized by playing a short clip that contains it at the start of playback. This clip, known as a codec initialization clip, should be nearly undetectable during playback. Add this clip by including one of the following parameters within the playback URL:

Tutorial

Goal: Learn how to add closed captioning support to on-demand content.

Key Steps

1. Create a TTML file.
2. Add style information.
3. Slice the file.

Step 1: Create a TTML File

There are many methods for creating a TTML file. In this example, we use an online tool from Microsoft called HTML5 Video Caption Maker (VCM).

Because it is an online tool, VCM has the limitation that your content must be hosted via a web server. VCM prompts you for the URL of your content.

1. Copy and paste the following URL into the text field labeled "Enter URL of video file:":

   http://ftp.nluug.nl/ftp/graphics/blender/apricot/trailer/sintel_trailer-480p.mp4

2. Select the Load button.

   The trailer is short and only has a few lines of dialog.

3. Begin by selecting Play.

4. Once you hear a line of dialog, select Pause.

5. Enter the dialog in the textarea below the video (yellow highlight added for emphasis).

6. Select Save & Continue. The video continues playing.

7. Wait for a line of dialog, then select Pause.

8. Enter the dialog in the text area below the video.

9. Repeat these steps until you are satisfied with the captions.

10. Find the TTML radio button and select it. This displays the markup for your TTML file.

11. Copy this text and save it as a local text file on your computer.

The same TTML used in the example can also be copied below:

<?xml version="1.0" encoding="UTF-8"?>
<tt xmlns="http://www.w3.org/ns/ttml" xml:lang="en" >
<body>
    <div>
        <p begin="0.000s" end="14.780s">What brings you to the land of the gatekeepers?</p>
        <p begin="14.780s" end="20.193s">I'm searching for someone.</p>
        <p begin="20.193s" end="39.444s">A dangerous quest for a lone hunter.</p>
        <p begin="39.444s" end="44.371s">I've been alone for as long as I can remember.</p>
    </div>
</body>
</tt>

Step 2: Add Style Information

The default caption styling varies widely by platform. To give your captions uniformity, add minimal style information to your TTML document. The document below includes the style additions.

Step 3: Slice File with TTML

Create the VOD asset using the command line slicer. Use the -ttml option to tell the slicer where to locate your TTML caption file. The following is an example slicer command with the option:

/path/to/slicer /path/to/video/sintel_trailer-480p.mp4 -u username -apikey APIKey -ttml /path/to/ttml/sintel_trailer-480p.mp4.ttml