Step 4: Match Formats to Categories - AWS Elemental Live

This is version 2.20 of the AWS Elemental Live documentation. This is the latest version. For prior versions, see the Previous versions section of AWS Elemental Live and AWS Elemental Statmux documentation.

Step 4: Match Formats to Categories

There are different procedures to follow to create captions encodes in the output. The correct procedure depends on the "category" that the output captions belong to. There are five categories of captions, described in the following table.

On the list of outputs that you have created, make a note of the category that each captions option belongs to.

Format of Output Captions

Category of This Format

Ancillary+Embedded The captions in ancillary format are in the ancillary data in the stream. The embedded captions are embedded in the video.

To choose Ancillary+Embedded, choose Embedded as the Destination Type (in the procedure). AWS Elemental Live will automatically produce both Ancillary and embedded.





CFF-TT Object







RTMP CaptionInfo


RTMP CuePoint Object
SCC Sidecar



SCTE-27 Object
SMI Sidecar
SMPTE-TT Sidecar when in Archive
SMPTE-TT Stream when in MS Smooth
SRT Sidecar



TTML wrapped in ID3 data Wrapped in ID3 data





For example, your list of outputs might now look like this:

  • MS Smooth output with TTML captions (sidecar) in Czech.

  • MS Smooth output with TTML captions (sidecar) in Polish.

  • HLS output with WebVTT captions (sidecar) in Czech.

  • HLS output with WebVTT captions (sidecar) in Polish.

Captions Embedded in Video

The captions are carried inside the video encode, which is itself in an output in the output group. Only one captions asset ever exists within that video encode. That single asset might contain captions for several languages.

Captions Object

The captions are in their own "captions encode" in an output in the output group. They are not part of the video encode. However, they are in the same output as their corresponding video and audio encodes. There might be several captions encodes in the output, for example, for different languages.


The captions are each in their own output in the output group, separate from the output that contains the video and audio. Each captions output contains only one captions asset (file), meaning that it is a "captions-only" output. The output group might contain several "captions-only" outputs, for example, one for each language in the output group.

TTML Captions Wrapped in ID3 Data

The captions are converted to TTML and included in ID3 data. (The other way to produce TTML output is as a sidecar.)

SMPTE-TT in MS Smooth

The captions are handled as a separate stream in MS Smooth.

Note that SMPTE-TT captions for other package types are handled as sidecars. However, for both sidecar handling and stream handling, the procedure for setting up SMPTE-TT captions in the output is identical. AWS Elemental Live will package the SMPTE-TT captions correctly for the package type.


Here, the captions are converted into text and then overlaid on the picture directly in the video encode. Strictly speaking, once the overlay occurs, these are not really captions because they are indistinguishable from the video.