As the RxPlayer manages multiple type of streaming technologies, which can use their own definition and terminology, we had to find a compromise and use our own terminology, which try to take the best from these.
We here define various terms used in the documentation which might not be obvious right along.
Simply put, what we call an "Adaptation" is just an audio, video or text track.
More technically, it is an element of a Period (and by extension of the Manifest) which represents a single type of media.
An adaptation can be for example any of those things:
- A video track
- A french audio track
- An italian text track
- A thumbnail track
- ...
Many Streaming Technology have this concept even though their name can change, an Adaptation is equivalent to:
- DASH's AdaptationSet
- Microsoft Smooth Streaming's StreamIndex
Note: There is minor differences between the RxPlayer's Adaptation and DASH'
AdaptationSet. Namely multiple AdaptationSets can be merged into a single Adaptation
in very specific cases. You can find more infos on it
here.
When we talk about the "buffer" in the RxPlayer, we most likely refer to the structures in the browser holding media data, waiting to be decoded.
Several layers of buffers can be defined in the browser-side to allow to have a smooth playback, fast seeking etc.
RxPlayer's buffer types describe a single "type" of media.
Example of such types are:
- "video": which represents only the video content
- "audio": the audio content without the video
- "text": the subtitles, for example
Depending on the context, a chunk can be either a sub-part of a Media Segment or the Media segment itself.
An initialization segment is a specific type of media segment, which includes metadata necessary to initialize the browser's internal decoder.
Those are sometimes needed before we can actually begin to push any "real" media segment from the corresponding Representation.
As such, when one is needed, the initialization segment is the first segment downloaded for a given Representation.
The Manifest is the generic name for the document which describes the content you want to play.
This is equivalent to the DASH's Media Presentation Description (or MPD), the Microsoft Smooth Streaming's Manifest and the HLS' Master Playlist.
Such document can describe for example:
- multiple qualities for the same video or audio tracks
- multiple audio tracks in different languages
- presence of subtitles
Note that this concept is only used in Streaming technologies. You won't interact with a Manifest if you're directly playing a MP4 or webM file.
A media segment (or simply segment), is a small chunk of media data.
In many streaming technologies, a content is separated into multiple chunks of small duration (usually between 2 and 10 seconds).
This allows, for reasons to long to detail here, to easily implements many features:
- live streaming,
- language switching
- adaptive streaming
When you play a content with the RxPlayer, it will most of the time download media segments of different types (audio, video, text...) progressively rather than the whole content at a single time.
A Period is a sub-element of the the Manifest which defines what the content will be from a starting time to an ending time.
It contains its own Adaptations and Representations. Having multiple Periods in the same Manifest allows for example to define multiple programs on the same content, each with its own tracks and quality characteristics.
Depending on the transport used, they correspond to different concepts:
- for DASH contents, it is more or less the same thing than an MPD's
<Period>element - for "local" contents, it corresponds to a single object from the
periodsarray. - for "MetaPlaylist" contents, it corresponds to all the Period elements we retrieved
after parsing the corresponding Manifest from the elements of the
contentsarray. - any other transport will have a single Period, describing the whole content.
--
As an example, let's take a manifest describing a live content with chronologically:
- an english TV Show
- an old italian film with subtitles
- an American blockbuster with closed captions.
Let's say that those sub-contents are drastically different:
- they are all in different languages
- the american blockbuster has more available video bitrates than the old italian one
Because the available tracks and available qualities are different from sub-content to sub-content, we cannot just give a single list of Adaptations valid for all of them. They have to be in some way separated in the Manifest object.
That's a case where Periods will be used. Here is a visual representation of how the Periods would be divided here:
Period 1 Period 2 Period 3
08h05 09h00 10h30 now
|==================|===========================|====================|
TV Show Italian Film American Blockbuster
Each of these Periods will be linked to different audio, video and text Adaptations, themselves linked to different Representations.
A Representation is an element of an Adaptation, and by extension of the Manifest) that describes an interchangeable way to represent the parent Adaptation.
For example, a video Adaptation can have several Representations, each having its own resolution (width and height) or its own bitrate (amount of bytes per seconds of content). The idea behind a Representation is that it can be changed by any other one in the same Adaptation as the content plays.
This is most often implemented to allow multiple bitrates for the same Adaptation, to be more flexible to poor network (low bandwidth) or computing (slow computer) conditions.
A Representation has its equivalent in multiple Streaming technologies. It is roughly the same as:
- DASH's Representation
- Microsoft Smooth Streaming's QualityIndex
- HLS' variant (the notion of variant is actually a little more complex, so here it's not an exact comparison)