Microsoft Teams offers both Live Captions and transcription. Each works differently. This blog will help you understand them and which to use. Microsoft Teams also supports CART (Communication Access Realtime Translation), human-generated live captioning.
Microsoft Teams Live Captions
Microsoft Live Captions shows, as the name suggests, real-time speech to text in a meeting speech at the bottom of the video window. Live captions use the Microsoft Automatic Speech Recognition (ASR) technology service, which generates Teams meeting captions. Microsoft ASR involves no human intervention, and no one can access the meeting audio or any other meeting information at any time. Caption data can’t be downloaded and is permanently deleted after the meeting is over.
A meeting organizer or caller can enable it by going to the meeting controls, selecting the three dots, and then “Turning on live captions.”
Desktop users can change the language. This is not doing a translation but instead telling the speech to text engine which language is being used to better interpret and display it in the correct language. When you change the spoken language setting, it affects everyone. The captions and transcript language will change for all meeting participants. Users on mobile will view captions in the supported languages, but only desktop users will be able to switch the spoken language.
A large number of languages are supported: English (Canada), English (India), English (United Kingdom), English (Australia), English (New Zealand), German (Germany), Portuguese (Brazil), Dutch (Netherlands), Dutch (Belgium), French (France), Spanish (Spain), Japanese (Japan), French (Canada), Chinese (Cantonese, Traditional), Chinese (Mandarin, Simplified), Hindi (India), Italian (Italy), Korean (Korea), Spanish (Mexico), Swedish (Sweden), Polish (Poland), Arabic (United Arab Emirates), Arabic (Saudi Arabia), Danish (Denmark), Finnish (Finland), Norwegian (Norway), Russian (Russia).
Admins can turn Live captioning on or off with Set-CsTeamsMeetingPolicy
Microsoft Teams CART (Communication Access Realtime Translation) Captioning or captioning by humans
Communication access real-time translation (CART) is a service in which a trained CART captioner listens to the speech and instantaneously translates all speech to text. The CART captioner uses a special phonetic keyboard or stenography method to produce an accurate translation that is broadcast to the recipients on a screen, laptop, or another device. CART is sometimes called open captioning, real-time stenography, or real-time captioning. It can be helpful if you need a higher accuracy than computer-generated captions or working in an area with specialized language that computer-generated captioning may struggle with. CART is available in all tenants, including GCC.
As a meeting organizer, if your wish to use a CART service, you will need to provide the CART captioner with a CART caption URL for the meeting, which allows a CART captioning solution or software to connect to and provide CART captions in a Microsoft Teams meeting. The CART URL should be supplied to the CART captioner before the meeting, so be set up and ready. It can also be done during the meeting if required. In both cases, turning on CART and getting the URL is found in the meeting options.
Presenters can switch between Microsoft Live captions and CART during a meeting if required.
Microsoft Teams meeting and call Transciptions
Transcriptions are different to live captioning. The text appears alongside the meeting video or audio in real-time, including the speaker’s name (unless they chose to hide their name by personal policy) and a timestamp. It is referencable throughout the meeting and downloadable after the meeting. Transcription is available in both meetings and calls. Transcription is currently only supported for users who set their language to or speak English in Teams meetings.
Teams live transcription files are stored in the meeting organizer’s Exchange Online account, and only the organizer and tenant admin has permission to delete it. If enabled, a copy of the transcript is stored with the meeting recording, which allows Search, CC, and transcripts on the meeting recording.
Transciptions / Closed captions in Teams Meeting and Call Recordings
Captions for Teams meeting recordings will be available during playback only if the user had transcription turned on at the time of recording. Admins must turn on recording transcription to ensure their users have the option to record meetings with transcription.
Closed captions aren’t fully supported if the Teams Meeting Recording is moved or copied from its original location on OneDrive for Business or SharePoint.
Admin Control over Teams Transciptions
Transcriptions are controlled by a combination of a per-organizer and per-user policy. This setting controls whether captions and transcription features are available in a meeting and during playback of meeting recordings. The person who started the recording will also need transcription enabled for meeting recordings to have a transcription. Transcription is disabled by default on the Teams global policy.
Teams transcription is a “user convenience” feature, not a compliance feature. Admins can’t force it on for every call or easily centrally collect and review transcription for every call. For that use case, you would use a third party certified compliance recording solution. These can record and/or transcribe all calls and meetings, as well as more advanced things like sentiment analysis.
There is no API access to Teams transcriptions. Transcription is not available in GCC-High and DoD environments.