Broadcast-grade captions and dubs inside your SRT stream
SyncWords sits ahead of your transcoder on the SRT contribution path. Your MPEG transport stream comes back with captions, subtitles and dubbed audio tracks muxed inside, without re-encoding the video.
- Embedded EIA-608
- Teletext (DVB-TXT)
- DVB-SUB
- DVB-TTML
- Dubbed audio tracks
How it works
Your stream goes out the way it came in, with subtitles inside
Point your stream at SyncWords, or let SyncWords connect to you. Caller and listener workflows are both supported, so you keep the SRT setup your production already runs. No transcoding at any step.
Your stream in
Send your live SRT feed as caller or listener, with an encryption passphrase if you use one. Any framerate, built for high-bitrate HD, 4K and UHD video.
Captioned, translated, dubbed
One live service can return same-language captions, translated subtitles and dubbed audio tracks together.
Muxed into the transport stream
The Kobe Muxer, SyncWords' own MPEG-TS stack, writes each output into the stream: 608 in the video, each subtitle language on its own PID, dubs as additional audio tracks.
Back to your production
The stream returns to your environment with everything inside, so you can monitor it before it goes to your transcoder and out to air.
Formats on this path
Every broadcast caption format, one stream
The SRT path carries the full broadcast format set. Each language travels in the same transport stream, labeled so set-top boxes and players show the right selector.
Embedded EIA-608
Classic embedded captions written into the video, in the seven languages the standard supports, including English, Spanish, French, Portuguese and Italian.
Teletext (DVB-TXT)
The workhorse of European broadcast chains. Per-language page mapping, auto-assigned from page 888.
DVB-SUB
Bitmap subtitles rasterised before they leave us, so any script renders correctly on any decoder, fonts included.
DVB-TTML
The W3C-based standard the industry is heading toward, with modern styling and positioning where the chain supports it.
Service mapping
Map each stream once, then reuse it
Create a SyncWords service for each of your SRT streams. Once a service is mapped to its channel, start and stop it as often as you need without touching the settings again. Run it around an event schedule, or around the clock.
- One service per channel
- Each channel gets its own service, with its own languages, dictionaries and subtitle styling saved against it.
- Event-based or 24/7
- Start and stop a service around a match or a conference, or leave it up for a linear channel that never goes dark.
- Concurrent channels
- Run as many services at once as your schedule needs. No per-channel licence and no per-channel ceiling.
- Several languages, one stream
- Each subtitle language rides its own PID inside the same stream, so you are not opening a new output for every market.
- Audio-track selection
- Pick the program audio by track index or PID, down to specific channels within a track, so captions come from the right feed.
- Redundancy
- Dual-pipeline A/B failover with SCTE-35 pass-through, including dual independent SRT streams, with primary and backup in any AWS region.
Automation
Wire it into the systems you already run
Create and manage SRT services from your own automation, so scheduling captions sits in the same tooling as the rest of your playout.
- Reusable channel mappings
- Event-based scheduling
- Webhook events
- Start, stop and schedule without reconfiguring
curl -X POST https://api.syncwords.com/v1/live/services \
-H "Authorization: Bearer $SYNCWORDS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"name": "Channel 1 - 24/7 news",
"input": { "protocol": "srt", "mode": "listener", "port": 9000 },
"captions": { "source_language": "en" },
"subtitles": { "format": "dvb-ttml", "languages": ["es", "fr", "ar", "ja"] },
"output": { "protocol": "srt", "mode": "caller", "url": "srt://playout:9100" }
}'Create an SRT service that returns DVB-TTML subtitles in four languages.
In your chain
Runs with the stack you already have
SyncWords slots between your contribution feed and your transcoder. AWS Elemental workflows are the most common home for this path, contribution encoders from LiveU, TVU and Ateme feed it, and existing caption encoders from Evertz and Link Electronics can be driven over IP while you migrate at your own pace.
Frequently asked questions
How much latency does captioning add to an SRT stream?
Separate the video path from the captions. The video path itself adds a half-second latency in SRT workflows: about 0.3 seconds of muxing plus each SRT leg. Captions land in the stream with sub-2s caption delay in SRT, where typical captioning solutions display live captions within 4 to 8 seconds. If you prefer captions perfectly in sync with the audio, a few seconds of configurable buffer absorbs the processing time entirely. See latency, processing time and delay for how the three quantities relate.
Do you support SRT caller and listener workflows?
Both. SyncWords can call your stream or listen for it, with encryption passphrases supported, so you keep whichever SRT setup your production already runs. The captioned stream is pushed back to your environment the same way.
Can this path carry non-Latin scripts?
Yes, with the right format. DVB-SUB carries any script at all, because the subtitle is rasterised before it leaves us and the decoder never needs the font. DVB-TTML handles non-Latin scripts natively where the chain supports it. Teletext has its own character-set limits and covers part of the language set. Embedded EIA-608 carries only the seven languages the standard supports, so anything beyond that rides as a subtitle track.
How do dubbed audio tracks travel on this path?
As additional audio tracks in the same transport stream. Codec is settable per track (AAC, PCM or MP2), PIDs can be explicit or auto-assigned, and each dub's position in the PMT relative to the original audio is controllable, so downstream systems see a normally-formed multi-audio stream.
Do we need a hardware caption encoder?
No. The whole path runs in the cloud, with no transcoding and no per-channel licence. If you already own caption encoders, SyncWords can drive them over IP instead, so you migrate to the cloud path at your own pace.
What happens to SCTE-35 markers and ad insertion?
SCTE-35 markers and SSAI are preserved across every language output, so ad structure matches the original broadcast. Dual-pipeline A/B failover carries SCTE-35 through as well.
Can we feed captions from our own captioner into this path?
Yes. SyncWords can ingest captions from your own captioner or steno software over a direct EIA-608 Control-A connection, and mux them into the stream the same way as AI captions.
Put it on your own feed
Broadcast teams usually start with a proof of concept on the workflow they actually run. Bring your SRT feed and our engineers will set it up with you.