Bitmovin has moved a hackathon project for automated live translation into its Live Encoder product. The integration sends a live feed to SyncWords over SRT and receives translated audio and subtitle streams for packaging and distribution.
The workflow can produce multiple audio languages and WebVTT or IMSC TTML subtitle tracks from one source. Bitmovin says DVB subtitle output is in production for German, English, French, Italian, Portuguese and Spanish, while DVB teletext remains a more complicated path because of character-set and regional constraints.
Technical significance
Live localisation is not a single AI operation: audio must be transcribed, translated, synthesised, timed and signalled in manifests without breaking continuity. Integrating those tracks directly into the encoder can simplify orchestration, but delay, voice quality and error handling remain visible to the audience.
What to watch
The source does not publish end-to-end latency, word-error rates or behaviour during speaker overlap and poor audio. Broadcasters should also examine editorial supervision, disclosure, voice rights, profanity handling and a fallback plan when a language service fails during a live event.
