api.SpeechRecognition - Edge does not actually support SpeechRecognition Web Speech API

[mbforbes]

MDN URL

https://developer.mozilla.org/en-US/docs/Web/API/SpeechRecognition

What specific section or headline is this issue about?

Browser compatibility

What information was incorrect, unhelpful, or incomplete?

Microsoft Edge does not actually support the SpeechRecognition Web Speech API. It's listed that it is "Implemented with the vendor prefix: webkit." While using it does not generate any errors, the implementation does nothing. No results are ever returned.

This may be an issue of semantics. Does implementing with a no-op count as implementing? :-)

What did you expect to see?

The Edge column should either have a "No" or a warning sign because the implementation is broken.

Do you have any supporting links, references, or citations?

Two users have tested on two different devices, each on two separate pages (4 total tests). For all of them, Chrome's speech recognition works correctly, and Edge never returns any results.

n=4 isn't huge, but it's been consistent. I also believe this issue is known by Edge.

Here's an example test page this can be used to verify SpeechRecognition on the Web Speech API: https://www.google.com/intl/en/chrome/demos/speech.html

Do you have anything more you want to share?

No response

MDN metadata

Page report details

• Folder: en-us/web/api/speechrecognition
• MDN URL: https://developer.mozilla.org/en-US/docs/Web/API/SpeechRecognition
• GitHub URL: https://github.com/mdn/content/blob/main/files/en-us/web/api/speechrecognition/index.md
• Last commit: https://github.com/mdn/content/commit/6b1e3eebf22abf1b73bb219581335b1147b75d7a
• Document last modified: 2023-03-12T09:29:44.000Z

[bsmth]

Hi @mbforbes thanks a lot for raising this. I've moved it over to the browser-compat-data repo which is where the source of this compatibility data is managed.

Just to check, there are some demos to try to reproduce this. I did the following:

git clone https://github.com/mdn/dom-examples.git
cd dom-examples/web-speech-api/speech-color-changer
http-server -c-1
# Available on:
#   http://127.0.0.1:8080

This works on Chrome, but fails on Edge with "Language not supported" error. You may want to try the other examples in that repo, but that doesn't look promising to me.

See https://github.com/mdn/dom-examples/blob/40cb0f10d859384a6160861ccbbf78f337fed6a0/web-speech-api/speech-color-changer/script.js#L17

[mbforbes]

Hey @bsmth thank you for the reply, and apologies for my newb raising of this issue in the wrong repo 😅 I appreciate you looking into it already!

[bsmth]

Hey @bsmth thank you for the reply, and apologies for my newb raising of this issue in the wrong repo 😅 I appreciate you looking into it already!

Not a problem, I hope the info helps investigate / narrow it down.

[gabrieljbaker]

@mbforbes , I've been trying to get recognition to work in Edge for a while - definitely confirmed that it does not work. Wild that Safari has this and Edge doesn't.

[mbforbes]

So, with a higher sample size, the plot has thickened. I have confirmed reports of Edge working for some folks. But for others, it doesn't work at all. Anecdotally, it's about 1:4 working vs not, but the n is too small. No idea when vs not.

[bsmth]

So, with a higher sample size, the plot has thickened. I have confirmed reports of Edge working for some folks. But for others, it doesn't work at all. Anecdotally, it's about 1:4 working vs not, but the n is too small. No idea when vs not.

For those that report it working, could you get some information on operating system and browser version? That would really help figure out what's happening here.

[gabrieljbaker]

@bsmth and @mbforbes , here's some color to it:

[mobilepadawan]

Does anyone know if there is exists an estimated date to start supporting Edge on Mac OS over ARM processor?

Chrome is working fine in this O.S. and processor.

[bsmth]

here's some color to it

Thanks for picking that up. Here's a link to the post in question for convenience: https://twitter.com/khmyznikov/status/1772573147755073667

I think we could add a note for Edge entry that this depends on Azure Speech component & is not yet supported on ARM machines, it would be great to link to a bug tracker or some other official docs for this.