Live subtitles in your CMAF Ingest cloud workflow
SyncWords sits between your transcoder and your packager over DASH-IF Ingest, which covers CMAF Ingest and the older HLS transport (DASH-IF Interface 2). Subtitles arrive as a TTML track alongside your media, and your packager delivers them the way it delivers everything else.
- CMAF Ingest & HLS
- TTML and WebVTT subtitle tracks
- Full Unicode support
- No transcoding
How it works
A subtitle and dubbing track alongside your media
This is the natural position for cloud pipelines built on MediaLive and MediaPackage and their equivalents: the transcode stays yours, the packaging stays yours, and SyncWords adds the languages in between.
Your transcoder's output in
SyncWords ingests the CMAF stream your transcoder is already producing. Nothing about the encode changes.
Captioned, subtitled, and dubbed
Live ASR transcribes the program audio and translation produces each subtitle and dubbing language, with your dictionaries and glossaries applied.
TTML and AAC tracks to your packager
Each language is pushed to the packager as a TTML subtitle or AAC audio track over DASH-IF Ingest, alongside the media it belongs to.
Your packager delivers
The packager converts the TTML, typically to WebVTT, and includes it in the HLS and DASH output it already serves your CDN.
Why this position
Why teams land between transcoder and packager
Unicode where 608 stops
A long-standing reason customers choose this path: TTML carries the scripts and languages that embedded EIA-608 cannot carry natively, and the packager delivers them player-side.
No re-architecture
The pipeline you built stays intact. SyncWords is one more ingest into the packager, not a replacement for any part of your chain.
One session, every language
Captions and translated subtitles fan out from a single live session, each language its own track, all landing in the same packager.
In your chain
Built for cloud packaging workflows
AWS Elemental MediaPackage V2, Ateme, Unified, and others are supported.
Frequently asked questions
What subtitle format does SyncWords deliver on CMAF ingest?
The intermediate format is TTML, which gets convered to WebVTT, IMSC, or SMPTE-TT by the packager for delivery to players.
Can SyncWords add subtitles to a CMAF stream without republishing the media?
It depends on your packager. Append mode requires the packager to handle a track disappearing gracefully. AWS Elemental MediaPackage V2 does not, so it runs in relay mode in practice, where SyncWords relays the media along with the subtitle tracks. Unified Packager is better positioned to support append.
How much latency does this add?
The base latency of this path is one segment, and that time also counts toward absorbing caption processing, so buffer and base latency overlap rather than sum. A configurable buffer can be added on top when you want captions to land with no perceptible delay. The three quantities involved are explained on the latency page.
We use HLS between our transcoder and packager rather than CMAF Ingest. Does that work?
Yes, SyncWords supports both. The workflow is the same: subtitle tracks pushed to your packager alongside the media.
Does SyncWords re-encode the video?
No. There is no transcoding in any SyncWords path. Your transcoder's output passes through untouched, which is also why there is no quality or performance loss at 4K.
Prove it on your pipeline
Tell us your transcoder, your packager and your languages, and our engineers will set up a proof of concept on your own chain.