Lingopal is expanding its live and on-demand localisation platform to more than 130 languages and adding a measurement layer called Signal. The module is designed to compare translation and dubbing output across languages using metrics including P.808 mean opinion score, text-to-speech naturalness, background noise and voice complexity and clarity.
The broader platform provides low-latency speech-to-speech translation, captions, subtitles and media-library localisation. Lingopal says live dubbing can operate below ten seconds of latency and that future work will use video analysis to generate synthetic commentary tailored to a programme or audience.
For broadcasters, Signal is the more consequential change because multilingual AI needs repeatable quality control rather than a simple language count. The announcement describes the metrics but does not publish methodology, benchmark datasets or independent results, so teams will still need programme-specific listening tests and human linguistic review.
