Learn how to use Mayo Clinic Connect
Request an Appointment
Has anyone tried a speech-to-text app? I have downloaded 3 of them and they all have a "wrong word" issue. Bill
Interested in more discussions like this? Go to the Hearing Loss group.
I was able to download Live Transcribe on my Android – Galaxy S8. It's ok, but I have difficulty trying to determine what I the other person said from what I said.
Jump to this post
Live Transcribe reels all the conversation out in a single paragraph…no break between speakers. That does make it difficult to recreate after a meeting's over, especially one that lasts for hours and is larded with lots of technical stuff. Live Transcribe is better than anything else about unusual words, but some technical stuff is very different from what what actually said. Much less bad than Otter.
@lindleys and @joyces, not separating the speaker is a drawback to using speech to text. CART (real person doing the transcriptions) will not have that problem. When I have a remote meeting, usually by using Zoom, I delegate who speaks so I will say their name before the person speaks. The name shows up in the transcript. Also, before someone speaks, we remind them to say their name first. It's a duplication of the persons name, but it prevents the problem you describe. Some meeting platforms know who is speaking and will automatically insert that into the transcript.
Tony in Michigan
I agree with you about Live Transcribe which I have been using since it first came out. It’s been invaluable with everyone wearing masks. I tried Otter and it was okay but I uninstalled it as it was not being used. I am going to give it a try again. I use TextEdit in desperation if the WiFi is bad in a certain area but it’s mediocre at best.
I never found it difficult to separate my comments from the other speaker even in the few meetings I have used it in. But my use of it is 90% personal and I never thought of that option like CART which is awesome. I am going to post another review on Live Transcribe in the play store and mention the possibility of separating conversations as a future upgrade.
Thank you Gaudellet University.
Lots of comments on Live Transcribe not identifying who is speaking. I think Otter does that, and I know that http://www.Streamer.center can automatically identify who is speaking. For a live conversation it shows me the name and a picture of the person that's speaking. In the printed transcript it doesn't show the picture, but it does list the name along side the text of what they said. With Streamer I can also tag the transcript with notes.
Not identifying speakers isn't an issue of speech-to-text software, but rather an issue with the Live Transcribe app. I think that in the printed transcript Otter separates speakers (but not during the actual conversation), and I know that http://www.Streamer.center identifies speakers both during the actual conversation and in the printed transcript. During the live conversation it also pops up a picture of the person that is talking at that moment.
… and for those of you using Zoom, both Rev and Streamer now offer live captions for Zoom meetings and webinars. For Rev the speakers aren't identified, whereas with Streamer they are. Here's a link to the Streamer video showing you how it works. https://streamerlink.cc/Streamer-Zoom
@livingat45north, I just watched the video. At the beginning of the video, they show "Captioning is autogenerated by Streamer". So, it appears that Streamer can predict what will be said since the captions appear before the dialect. Am I missing something by their claim that "Captioning is autogenerated"?
Tony in Michigan
Yes, your confusing live captioning which is shown as a person is speaking, with captioning that is displayed when viewing a previously recorded video. Streamer does both. If you view the closed captioning on a pre-recorded video (which is what you're doing), during the live event the words were generated by Streamer along with timestamps that state when the words should appear in that recorded video, In YouTube, that type of file format is called an SRT file. In the second half of that video they show how the live captioning appears, about a half-second after the person says the word.
Hey, Bill. Unfortunately, all of the transcription apps are prone to errors, but they're still pretty great. I use both otter.ai and Google Live Transcribe regularly for conversation. I have given up on using them at doctor's visits, because of inaccuracies. I find that old school tech works best for that, so I bring a small dry erase whiteboard with me. Then the doctor can write down whatever they want to say to me. A lot of doctors offices have those on hand as a matter of course.
One thing that I don't think was mentioned is Google Meet. Meet will let you make video calls with captioning without having to pay a fee (like incorporating Streamer into Zoom). All you need is a free gmail account and then you can start a Meet call from your gmail.
Streamer is now directly integrated with Zoom, so you can use it to caption your meetings and webinars. Here's the info: https://streamerlink.cc/Streamer-Zoom-Letter
The subscription to Streamer for 99/yr. isn't bad. I meet with several groups several times a week on Zoom and someone else has the 14.99/mo. Zoom subscription. Does anyone know if I would have to have a Zoom subscription in order to subscribe to the Streamer for captioning? @sag
@sag I just got done evaluating Streamer for Zoom for my HLAA Chapter meetings. Streamer's captioning works well. But the one big drawback to using it is that everyone who wants to access it on your call with you must download Google Chrome browser in order to access the microphone. With many older, non-techie people in our group that was thumbs down for it. Maybe if we can spend the next several months educating our members on how to download and use Chrome we might come bask to it.
Streamer works with the free version of Zoom, you do not need to pay for a Zoom subscription.
Connect with thousands of patients and caregivers for support and answers.
Already have an account? Sign In