VideoSuite

Real speaker: one audio track per microphone

In VideoSuite the transcript knows who is speaking because each person has their own track, not because an algorithm guesses. How it works and when guessing is still needed.

Most note assistants receive mixed audio and have to work out who is speaking (diarisation). VideoSuite, recording its own room, keeps one microphone per track: the owner of the track is the real name of the person who joined. So in live meetings the attribution of a sentence comes from the system, not from an estimate.

Who it is for

For anyone who needs to know for certain who said what: a committee, a client consultation, an interview.

What it does, step by step

  • Each microphone is recorded on its own track, in parallel with the composed recording of the meeting.
  • The transcript is aligned with word-by-word timestamps.
  • The host picks the meeting language when creating it (Spanish, English, Portuguese, French, German, Italian or Turkish); if nobody picks one, the transcription engine detects it and the report comes out in that language.
  • The viewer shows lanes per speaker, jump by minute and highlighting of the word being spoken.
  • When several people share one device, or when an external recording is imported, diarisation is used, and the host can assign a name to each voice.

Honest limits

  • With a shared device or imported recordings the voice has to be guessed, as in other tools.
  • If someone joins with an inexact name, the transcript carries that name.

Which plans include it

What each plan includes, with prices and the own-brand extra, is in the plans section of the site: VideoSuite's features are described here as they work today in production, and some (such as own brand or own domain) depend on the plan.

See all articles