search

Azure-Samples / SpeechToText-WebSockets-Javascript

SDK & Sample to do speech recognition using websockets in Javascript
MIT License
216 stars 151 forks source link

The "Interative" mode can't deal with the long wav in the SDK #47

Closed MagicRUBICK closed 7 years ago

MagicRUBICK commented 7 years ago

I used Sample.html demo to generate the transcript from a 30s wav in Interative mode, but it only return first several secs transcription. BTW, the other two mode works.

raaaar commented 7 years ago

The interactive mode is meant to be used with short utterances (2-3 seconds), as soon as there's a silence longer than a timeout interval, the recognition session is terminated. More details here: https://docs.microsoft.com/en-us/azure/cognitive-services/speech/concepts#intercative-mode