Today I ran a meeting that included a colleague who it 85% deaf. When we have in-person meetings he can lip-read but this is too difficult on screen.
It would be fantastic if live speech captions were available and they'd make a huge difference to their working life and others with similar requirements.
Users will have a transcription to follow along with who said what, and for the transcript file to be viewed after the meeting. You can read more details here: https://www.microsoft.com/microsoft-365/blog/2020/07/08/reimagining-virtual-collaboration-future-work-learning/.
URL in Working on it message ^ returns "We're sorry, the page you requested cannot be found."
Ayan Mullick commented
It would be great if the transcript file could also be easily linked to a GitHub issue.
Chris Tolmie commented
The "who said what" piece of this sounds great!
David Larkin commented
The link above about the transcript file doesn't work
this would be an excellent time saving when taking minutes of meetings there is nothing currently available and i think many people would be greatful for this facility
Transcript is only minimally useful if you do not know who said what
Giridhar Krishnagiri commented
In Teams, there is a recording feature along with capturing the transcript by enabling the captions. Unfortunately, there is no way to know who spoke. I am assuming it may be fairly simple to prefix each caption with the name of the person who spoke. The same should then appear in the transcript. It would make it lot easier to then edit the transcript what they would have actually spoken knowing who has spoken.
Format of VTT meeting transcripts is super unfriendly/unuseable. Would like to see
1. script of who said what and when!
2. identify when someone different speaks
3. timestamps for new speakers
4. diff formats: csv, word, pdf
instead of garbage report that chunks up into 5 second pagakges, littered with stuff like "NOTE Confidence: 0.896272599697113d24b0b28-23ab-4278-880e-f5ce66ae4c9e
00:00:01.640 --> 00:00:07.282".
Such information ^ is totally useless to me as meeting attendee/producter/presenter.
It would be beneficial to have the ability to pull the transcription from the recording. This would allow us to have the notes from the meeting without having to store the video recording.
Nidhish Dhru commented
Provide an api so that we can intercept the live transcription and store it in a different container. It will be good to have an Azure Stream Analytics receive a continuous stream of transcription from Teams which we can store it somewhere or use if via code.
I would like to be able to download a transcript of the chat that happens during a particular meeting.
this will be very useful
Also for the live caption
Kartik Verma commented
There is already a live captioning feature that works really well. Since Teams will know who is speaking, can you pull the name and create an automatic transcript for the meeting for people to download after the meeting? It will be a great help!
Tina R. commented
If the name of each separate speaker can not be determined, it would be helpful if at least the transcription could be attributed by login.
David Larkin commented
I agree 100%
I am also profoundly deaf and rely on lipreading. Most of the meetings quite often don't have videos and this also have a huge impact on being able to participate in Engineering or CAD presentations.
Idea would be to have an option/button which would start recording/saving the next set of messages in any chat area, chat/teams etc until you pressed stop.
So say I'm speaking to a colleague/manager, they say "okay so I wanted to talk to you about this client migration" I (or they) would then click the save transcript button. We go over details, numbers, specifics, scenario's and whatnot, then end off with "sounds good" and then one (or both, idea would be it saves locally for the person who turns it on and off) turns the off button.
Then we have the option to view/download or export the saved transcript, so we can add to documentation and other things, to refer back to later.
When generating automatic transcripts of recording done via Teams, it would be useful to include the name of the speaker with each message alongside the timestamp and UUID.
I have developed a formatter that will convert the VTT transcripts to conversational text and it would be good to include an option where each item of text/speech has the name of the speaker.