Add Custom Speech integration to SDK and sample

[mikebranstein]

The SDK did not previously allow for testing of the Custom Speech service using the Speech WebSocket protocol. This adds an option to select the Bing Speech API or the Custom Speech API. It uses the new Custom Speech API endpoint URLs, published after //BUILD 2018.

[msftclas]

All CLA requirements met.

[compulim]

If there is no changes in package.json, there should be no changes in package-lock.json.

If their package-lock.json is outdated for some reasons, should split this change into another PR focusing on that topic.

The PR looks great IMO.

[mikebranstein]

Thank you, @compulim. I've corrected the package-lock.json issue, removing my unintended changes.

[mattiagecchele]

Can you close this PR ASAP? We need the custom endpoint

[unsafecode]

+1

[tjoudeh]

Hello @mikebranstein @compulim is there any reason this PR is not completed yet? It is blocking using the new unified speech service in bot framework WebChat, right? I hope you can follow up on this internally with the team to look into it and complete it.

[compulim]

@tjoudeh I am not the owner of this repo and not a member of Cognitive Services team, so I cannot approve/merge this PR. 😔 I also hope to see it coming in asap.

[mikebranstein]

@tjoudeh I am also not the owner, nor Microsoft. I work for a MSFT partner that has been using this extensively, so I'd love to see the changes merged.

Perhaps @RaviPidaparthi could help us find the owner.

[msimecek]

@zhouwangzw Could you please check?

[npacucci]

Any news??? @WSturman. It's possible to merge this?

[vippatel90]

@WSturman , @mikebranstein any news?? why this is not being merged?

[mikebranstein]

@viananth I am not the owner of this repo, and not a Microsoft employee. I cannot merge, unfortunately. @WSturman and @wolfma61 - perhaps you can facilitate this merge. It looks like you've made the most recent commits.

[EricDahlvang]

Can we get this merged please?

[WSturman]

This sample is for Bing Speech, which is retiring. We have a JavaScript sample for the new Speech Service at https://github.com/Azure-Samples/cognitive-services-speech-sdk/tree/master/samples/js/browser.

[Kartik1607]

@WSturman Hi. I was wondering how to implement the endpoint from cris.ai for my custom speech. There is method SpeechConfig.fromEndpoint which accepts a url and subscription key, but there is no mention from where we can get that url. Is it the websocket URL, or REST url?

[erhopf]

@Kartik1607 - I think you're looking for the REST endpoints listed in the Swagger Spec: https://westus.cris.ai/swagger/ui/index. Please select v2 from the drop-down, as the details you're looking for may not be available in Beta.