Skip to content
Playing with Chrome's speech recognition API to communicate better with deaf patients
JavaScript HTML CSS
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
CNAME
LICENSE
README.md
favicon.ico
index.html
script.js
style.css
visible-speech.png

README.md

A hacky little one-page app I built one night to play with Chrome's speech recognition API to communicate better with my deaf patients while in the hospital. People had positioned a computer with a terminal at the end of a patient's bed who had recently become deaf. They were typing in notes. I thought, there has to be something better. I tried a handful of phone apps (like Live Caption) with some success, but then discovered Speech Logger. Speech Logger's real-time voice-to-text was a better option, but the user interface was cluttered with features we didn't need. Fortunately, the nice people at Speech Logger were kind enough to document the Web Speech API to get me started.

So, I threw together this little one-page app to leverage the speech recognition generously provided by Google in their Chrome browser. The goal was to make a really simple user interface that would act as a reverse teleprompter for our patient. I had to figure out how to keep it continuously listening, but, once I got past that little hump, it worked well. Even with the built-in microphone on the laptop in the room, it has proven far more useful than hand-typing messages. I threw in a few commands to make it a little more convenient:

  • "clear" – clears the screen
  • "delete" – deletes the last segment of recognized speech
  • "bigger" or "smaller" – change the size of text
  • "new line" or "new paragraph" – you can probably guess what these do
  • "stop listening" – turns off the mic

So, if you are working with a deaf patient (we're assuming he or she has reasonable vision, can read, can speak, and – like in our case – you don't have more effective & effecient ways to communicate). Place a computer at their bedside with a reasonable microphone, open up visiblespeech.com with a recent version of Chrome, and start talking to your patient.

Give it a try at https://bmamlin.github.io/visible-speech/ (requires Chrome browser)

You can’t perform that action at this time.