Getting Better Results with Zoom’s Live Auto-Transcription

Live Auto-Transcription was recently added to Zoom. This allows participants to see a text transcription of the speaker’s words, either as subtitles at the bottom of the screen, or as a transcript to the side. This isn’t a perfect solution, particularly for accessibility, but it may make live Zoom sessions easier for students to understand and can provide a searchable transcript for later use, if you choose to save it.

Auto-transcription will always struggle with specific terminology, but may also produce less than great transcripts depending on many other factors. There are a few things you can do which may help live auto-transcription to produce a clearer transcript for your session:

  • Speak slowly and clearly
    • You don’t need to put a full stop after each word, but speaking more slowly helps the system tell the difference between different words and where one word stops and another beings.
    • You can have people use the Slow Down reaction in Zoom to help you see if you need to speak more slowly for the auto-transcription.
  • Minimize Background Noise
    • The more clearly your voice stands out from the background the better Zoom will be at transcribing your spoken words.
    • Using a headset or a separate condenser microphone can help improve your sound quality.
    • Microphones also perform best when slightly out of your direct line of speaking, if possible, position your microphone near, but to the side or just below your mouth in your natural speaking position.
  • Leave a break between speakers
    • In the live transcript, Zoom will try to identify speakers, but it often needs a moment to “get its bearings”. If one speaker speaks to quickly after the other speaker Zoom may not be able to tell there was a switch.

Auto-Transcription of recorded Zoom sessions has been available for a while, and our video platform Yuja also provides auto-transcription abilities. With the Yuja auto-transcriptions, you have the ability to manually adjust the transcription to correct mistranscribed words. Another alternative instructors have shared with us is using Automatic Captions in Powerpoint. The quality of the transcription has been described as a little more reliable, but it also requires that your whole session have power point shared with participants.

Have you had particular success with transcripts from Zoom’s Auto-Transcription? Let us know what you did to achieve them.