Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Tracking: Web Chat + new Cognitive Services Speech SDK #1336

Closed
17 of 18 tasks
compulim opened this issue Nov 13, 2018 · 3 comments · Fixed by #1442
Closed
17 of 18 tasks

Tracking: Web Chat + new Cognitive Services Speech SDK #1336

compulim opened this issue Nov 13, 2018 · 3 comments · Fixed by #1442
Assignees
Labels
front-burner p0 Must Fix. Release-blocker
Milestone

Comments

@compulim
Copy link
Contributor

compulim commented Nov 13, 2018

This issue is for tracking the status on the work around Speech SDK.

Background

Web Chat use W3C Web Speech API to provide speech recognition and synthesis capability. Cognitive Services Speech Services has published their own API to provide speech recognition and synthesis capability.

This work is to create an adapter that turns Speech SDK into a W3C Web Speech API compliant so Web Chat (and other web apps) can consume.

For the deprecated Bing Speech, we have a version of the adapter. But that adapter no longer works with the new Speech Services. And we need to build a new one.

To-do

The to-do list below may expand from time to time to give more details on the actual work.

  • Speech recognition
  • Speech synthesis
    • Support pause() and resume()
  • Switch between Bing Speech and Speech Services
  • Integrate into Web Chat bundle

FAQs

We will update this issue when we have more information.

When will this feature land?

We are targeting 12/15 (version 4.2). But there are multiple risks related to this work. We are aware that this feature has very high demand and we are working with Cognitive Services Speech team to land this feature as soon as possible.

Will this feature support Custom Speech (CRIS)?

We want to support both Speech Service and Custom Speech Service. But information on our hands are limited. If we are not able to launch the first adapter with CRIS, we will schedule it as a point release (patch release).

Why I cannot use speech in Web Chat now?

Cognitive Services team deprecated Bing Speech and no new subscriptions can be created. Web Chat currently only support Bing Speech but not Speech Services. If you already have a Bing Speech subscription, you can continue to use speech in Web Chat until 2019 October. If you do not have a Bing Speech subscription, you will need to wait until this work is done.

What if my question is not in this FAQs?

Please post it in the comment below. We will try our best to answer it and put it up here for readability.

References

@compulim compulim added the 4.2 label Nov 13, 2018
@compulim compulim added this to the v4.2 milestone Nov 13, 2018
@compulim compulim changed the title Tracking: Web Chat + new Speech Services SDK Tracking: Web Chat + new Cognitive Services Speech SDK Nov 13, 2018
@andersponders
Copy link

andersponders commented Nov 15, 2018

i would like to help with this, have you guys found the source for the new SDK? when are they gonna throw it on NPM?

I am unable to find anywhere where i can get a subscription key to the old API, so it looks like i can't have voice in my azure bot until this issue is fixed, so I would like to help get it done asap

@compulim
Copy link
Contributor Author

compulim commented Nov 27, 2018

Their new SDK is on a ZIP file now, and it's a Webpack bundle. I am working on it in another repo and progress is pretty good. Speech recognition works as expected (plus automated tests), and now working on speech synthesis.

I am targeting 12/15 for a development build. I probably can't release as a production build until they release their non-bundled code on NPM.

I have updated the FAQs above based on your question. I am sorry to say you will need to wait until this work is done.

@corinagum
Copy link
Contributor

This bug is to bring the speech adapter for v4 on par with current v3. On par is on track for 4.2.
Other features in addition to this will be added to a bug tracked for 4.3. You can view these items on #1400

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
front-burner p0 Must Fix. Release-blocker
Projects
None yet
Development

Successfully merging a pull request may close this issue.

4 participants