Azure Media Service delivers video, audio, and text in different protocols. When you publish your live stream using MPEG-DASH or HLS/CMAF, then along with video and audio, our service delivers the transcribed text in IMSC1.1 compatible TTML. The University of Minnesota's Zoom web conferencing includes a feature that uses automatic transcription to turn the spoken word into text, in real-time. It should be used for accessibility, but not for accommodation (learn more about accessibility vs accommodation). Note: The Automatic Live Transcription feature does not work with breakout rooms, or in the Zoom for Linux app.
Live Transcription provides robot transcription services, which enables speech to text transcription in Zoom Meetings and Zoom Webinars. This feature is enabled by the host of the meeting or webinar once it begins. Participants will not see the Live Transcript option unless the host enables it. This feature currently only supports participants speaking the English language (Click here for the full Zoom Help Center article regarding the feature). While Zoom Video Communication systems provide assistive technology solutions for UNC Asheville Zoom platform users (Zoom Meeting and Zoom Webinar), these solutions are not intended to replace purpose built ADA accommodation services or systems for any compliance needs. The UNC Asheville Human Resources or Student Academic Accessibility Offices should be contacted for any questions or concerns regarding access to more advanced accommodation services).
- This feature is not available in Zoom Meeting Breakout Rooms.
- This featuredoes notcreate subtitles in the Zoom recording, subtitles only appear during the live meeting. Hosts who want subtitles to appear in the Zoom recording need turn onZoom Audio Transcription.
- The Live Transcription feature setting has been enabled within the UNC Asheville campus-wide Zoom account for all users, no further action is required to activate this feature.
- While Zoom Video Communication systems provide assistive technology solutions for UNC Asheville Zoom platform users (Zoom Meeting and Zoom Webinar), these solutions are not intended to replace purpose built ADA accommodation services or systems for any compliance needs. The UNC Asheville Human Resources or Student Academic Accessibility Offices should be contacted for any questions or concerns regarding access to more advanced accommodation services).
Zoom Live Transcription Feature
The Zoom Live Transcription feature requires that all event participants (Host and participants) utilize Zoom meeting software version 5.0.2 of higher.
- (Click here for information related to 'Keeping Zoom updated on Windows')
- (Click here for information related to 'Keeping Zoom updated on macOS')
Hosts or Co-Hosts must perform the following steps to enable auto-transcriptions for viewing by event participants:
Step 1: After the host has started the Zoom Meeting or Zoom Webinar event, they should click onLive Transcript icon at the bottom of the Zoom Toolbar.
Step 2: When the Live Transcription menu opens, clickEnable Auto-Transcription.
The Live Transcription feature is now active and the associated text will begin appearing along the bottom third of the video display. Meeting participants will be alerted that Live Transcript is available.
Note: If the meeting is being recorded, and you have enabled Live Transcript, you will receive a copy of the transcript in the same location of your recorded meeting.
STEP 3: Hosts also have the ability to disable Save Transcript. Click to toggle offSave Captionsin the Settings tab of the Zoom web application.
Live Audio Transcription Zoom
*This feature setting has been enabled within the UNC Asheville campus-wide Zoom account for all users, (Indicated by the blue background and the slider being to the right.)
Once Live Transcription is enabled by the host, participants have the option to choose how they want to view the live transcription feature by performing the following four steps.
STEP 1: Join the meeting or webinar via the Zoom Meeting ID provided by the Host. Once Live Transcription has been started, participants will see a notification aboveLive Transcriptin the meeting controls, attendees can click theLive Transcriptbutton to select how to display the meeting transcription. Participants may change the transcript text font size by clicking on Subtitle settings and moving the slider toNormal, Medium, or Large.
Show Subtitles displays subtitles on the screen.
View Full Transcript displays subtitles in real time with both the speaker's name and a time-stamp within the transcript window. We recommend this setting as best practice.
Subtitle Settings will open a window where you can adjust the font size of both the subtitles and the chat window.
STEP 2: ClickView Full Transcript to view subtitle and speaker.
STEP 3:If you would like a copy of the transcript, clickView Full Transcriptand Save Transcript. A copy of the transcript will be downloaded to your machine.Note: The transcript saves up to the moment of clicking Save Transcript. We recommend you clickSave Transcriptjust before the meeting ends to ensure you have the transcript from the entire meeting. Also note that this option may not be available if the host has disabled this feature.
Tips for using Live Transcription with Zoom
- Use an External Microphone: An external microphone/headset provides better audio quality and, therefore, more accurately matches what is being said verbally.
- Speak Slowly: Speaking slowly enables a more accurate match to what is being said verbally and allows others to follow the automatic transcript or subtitles.
- Take Turns/Pause Before Speaking: Zoom's active speaker tracking is used to identify the speaker. When another individual begins speaking, it can take a second or two before Zoom updates the transcript to identify the new speaker.
- Monitor the Participants Panel for the Go Slower function: Zoom's Participants panel displays all attendees and shows any nonverbal feedback the attendees provide. When using automatic live transcription, consider keeping the Participants panel open to see if attendees are indicating they need speakers toGo Slower.
Live Audio Transcription Companies
STEP 4: Live captioning settings are available for use that allow for font size and Chat display size changes, as well as selection of Screen Reader alerts.
- All event participants should use the Zoom Desktop App or Zoom Mobile App if possible to ensure that their respective names are accurate.
Per the Zoom Help Center article, the following are some issues (but not limited to) that may impact the accuracy of Zoom's live transcription feature:
- Background noise
- Volume and clarity of the speaker's voice
- Speaker's proficiency with the English language
- Lexicons and dialects specific to a geography or community.
(Click here for information regarding Office of Academic Accessibility accessibility support services)
Additional related knowledge base articles:
- Audio Transcriptions for Zoom Cloud Recordings
Resources for assistance or followup questions and concerns.
Contact UNC Asheville's Information TechnologyIT Service Deskfor assistance in implementing the above described services and support. Email: [email protected] or by calling 828.251.6445
Contact UNC Asheville's Center for Teaching and Learning unit for pedagogy and teaching strategies regarding this knowledge base article. Email: [email protected] or by calling 828.251.6540
Use the Zoom Video Communication systemknowledge basefor Zoom meeting specific support questions or concerns. www.zoom.us/support
Azure Media Service delivers video, audio, and text in different protocols. When you publish your live stream using MPEG-DASH or HLS/CMAF, then along with video and audio, our service delivers the transcribed text in IMSC1.1 compatible TTML. The delivery is packaged into MPEG-4 Part 30 (ISO/IEC 14496-30) fragments. If using delivery via HLS/TS, then text is delivered as chunked VTT.
Additional charges apply when live transcription is turned on. Please review the pricing information in the Live Video section of the Media Services pricing page.
This article describes how to enable live transcription when streaming a Live Event with Azure Media Services. Before you continue, make sure you're familiar with the use of Media Services v3 REST APIs (see this tutorial for details). You should also be familiar with the live streaming concept. It's recommended to complete the Stream live with Media Services tutorial.
Live transcription preview regions and languages
Live transcription is available in the following regions:
- Southeast Asia
- West Europe
- North Europe
- East US
- Central US
- South Central US
- West US 2
- Brazil South
This is the list of available languages that can be transcribed, use the language code in the API.
|English (United Kingdom)||en-GB|
|English (New Zealand)||en-NZ|
|English (United States)||en-US|
Create the live event with live transcription
To create a live event with the transcription turned on, send the PUT operation with the 2019-05-01-preview API version, for example:
The operation has the following body (where a pass-through Live Event is created with RTMP as the ingest protocol). Note the addition of a transcriptions property.
Start or stop transcription after the live event has started
You can start and stop live transcription while the live event is in running state. For more information about starting and stopping live events, read the Long-running operations section at Develop with Media Services v3 APIs.
To turn on live transcriptions or to update the transcription language, patch the live event to include a “transcriptions” property. To turn off live transcriptions, remove the “transcriptions” property from the live event object.
Turning the transcription on or off more than once during the live event is not a supported scenario.
This is the sample call to turn on live transcriptions.
Transcription delivery and playback
Review the Dynamic packaging overview article of how our service uses dynamic packaging to deliver video, audio, and text in different protocols. When you publish your live stream using MPEG-DASH or HLS/CMAF, then along with video and audio, our service delivers the transcribed text in IMSC1.1 compatible TTML. This delivery is packaged into MPEG-4 Part 30 (ISO/IEC 14496-30) fragments. If using delivery via HLS/TS, then the text is delivered as chunked VTT. You can use a web player such as the Azure Media Player to play the stream.
If using Azure Media Player, use version 2.3.3 or later.
For preview, the following are known issues with live transcription:
- Apps need to use the preview APIs, described in the Media Services v3 OpenAPI Specification.
- Digital rights management (DRM) protection does not apply to the text track, only AES envelope encryption is possible.