To signal end-of-speech, client applications send an audio chunk message with a zero-length body.
Speech Service interprets this message as the end of the incoming audio stream.
Coverage remained the same at 76.923% when pulling 8e096b53d291fceb857f880eb02040c9f1ff3c32 on pvaillant:fix-audio-stream-end into f85d94af8f393066fdb3ae3d3ed399bd2ea674f9 on noopkat:master.
Coverage remained the same at 76.923% when pulling 8e096b53d291fceb857f880eb02040c9f1ff3c32 on pvaillant:fix-audio-stream-end into f85d94af8f393066fdb3ae3d3ed399bd2ea674f9 on noopkat:master.
Fix #24
To signal end-of-speech, client applications send an audio chunk message with a zero-length body. Speech Service interprets this message as the end of the incoming audio stream.
SEE https://docs.microsoft.com/en-us/azure/cognitive-services/speech/api-reference-rest/websocketprotocol#client-end-of-speech-detection