Custom Keyword continuous recognition option using Azure Cognitive Speech Services sdk for Javascript

Question

Using the Custom Keyword Recognizer provided by Microsoft Cognitive Speech Services, I would like to setup continuous recognition on the browser using the microsoft-cognitiveservices-speech-sdk npm package. Is there a way to setup continuous keyword recognition. As of right now, only the single time recognition option is available as part of the sdk. Once the keyword is recognized, it would trigger Speech to Text services to process further speech. Post performing the required action, the keyword recognition would once again take over. Is there a way to accomplish this without using Custom Commands, which provides a continuous keyword recognition?

Answer 1

You can use this only on devices SDK startKeywordRecognitionAsync and not in other SDKs.

This starts speech recognition with keyword spotting, until stopKeywordRecognitionAsync() is called

Not sure whether this will be applicable to your scenario.

At this point, the devices SDK only works with Roobo dev kits and Azure Kinect DK.

Coming to your requirement, you can always re-arm keyword recognizer by calling it again (probably a loop sort of or alternate mechanism that may meet your requirement)

Answer 2

You can use this only on devices SDK startKeywordRecognitionAsync and not in other SDKs.

This starts speech recognition with keyword spotting, until stopKeywordRecognitionAsync() is called