Text tracks
Overview
The rx-player allows to display text tracks - such as subtitles or closed captions - directly over your content:
Adding text tracks to contents can be done by two means:
- by using a Manifest declaring those text tracks
- by manually adding text track(s) when you call the
loadVideo
API
You can then choose the right track through the different text track-related API, all documented in the general API documentation.
Supported text track formats
The rx-player supports the following formats:
- TTML (TTML1, EBU-TT and IMSC1)
- WebVTT
- SAMI
- SRT
- TTML embedded in an MP4 file
- WebVTT embedded in an MP4 file
Text tracks indicated in a manifest
Each streaming technology supported by the Manifest defines a way to add text track directly in their Manifests files.
This chapter explains what is supported by the RxPlayer.
In DASH
In DASH, text tracks are defined by AdaptationSet
elements, which have a
contentType
attribute equal to text
.
Those AdaptationSet
can also define a lang
, codecs
and mimeType
,
which are then exploited by the RxPlayer.
The lang
attribute is used to know the language the track is in.
The RxPlayer understands the following standards:
- ISO 639-1 (2 letters)
- ISO 639-2 (3 letters)
- ISO 639-3 (3 letters)
More complex combined formats are also understood, as long as it begins by one of the understood standards, followed by a dash ("-").
For example “en-US” is translated into just “en”, and then inferred to be english.
The mimeType
attribute is used to know in which format the track is in.
The RxPlayer understands the following ones:
application/ttml+xml
: TTML in plain texttext/vtt
: WebVTT in plain textapplication/x-sami
: SAMI in plain textapplication/mp4
: Text track embedded in an MP4 container.text/plain
: Generic plain text mimeType
For the last two, the codecs
attribute of the AdaptationSet
will be
exploited to know the exact format.
The rx-player uses the codecs
attribute for text tracks in only two cases:
- the
mimeType
is equal toapplication/mp4
- the
mimeType
is equal totext/plain
For the first case, both WebVTT and TTML can be embedded in an MP4 file. To know
which one we’re dealing with, the codecs
attribute should be equal to:
stpp
for TTMLwvtt
for WebVTT
For the second case ("text/plain"
), this is specifically to support plain
SubRip (SRT) subtitles. To use them you need to set codecs
simply to
srt
.
To know if we’re dealing with a closed caption text track, the RxPlayer uses the DVB-DASH specification.
That is, an AdaptationSet
is inferred to be a closed caption for the hard of
hearing, if it contains an Accessibility
descriptor with the following
attributes:
SchemeIdUri
set tourn:tva:metadata:cs:AudioPurposeCS:2007
value
set to2
In Microsoft Smooth Streaming
In Smooth Manifests, a StreamIndex
is inferred to be for a text track if its
Type
attribute is equal to text
.
The FourCC
attribute is used to infer the format of the text track. Only
TTML
is understood, and is translated to be a TTML track embedded in an MP4
container.
Adding support for other formats is very simple, open an issue if you want us to add a standardized FourCC code for another supported format.
The Language
is used to know the language the track is in. The rules are the
same than for DASH:
The rx-player understand the following standards:
- ISO 639-1 (2 letters)
- ISO 639-2 (3 letters)
- ISO 639-3 (3 letters) More complex combined formats are also understood, as long as it begins by one of the understood standards, followed by a dash ("-").
For example “en-US” is translated into just “en”, and then inferred to be english.
The Subtype
attribute is used to know if the language is a closed caption or
not.
At the moment, the RxPlayer infers the track to be a closed caption only if its
value is DESC
.
Text tracks added manually
It is also possible to add a supplementary text track dynamically, by using the
TextTrackRenderer
tool. You can read its documentation
here.
Text track display modes
There is two ways the text track can be displayed:
-
"native"
: The text track is displayed in<track>
elements, which are directly in the linkedvideoElement
. -
"html"
: The text track is displayed in a separate<div>
element.
The second ones allows for a better track stylisation. The distinction between those two is pretty simple and is explained here, in the loadVideo options documentation.