I completed the evaluation of Microsoft Speech SDK whether it fulfills our needs to build mobile ASR/TTS applications. The handicap of this SDK is that speech recognition and text-to-speech transformation occurs on the server and not on the mobile device. This solution requires constant WLAN connection with the Microsoft Speech Server and the application built is an ASP.NET application running on IIS and being accessed through Pocket Internet Explorer with a special plug- in.
I have sketched the system architecture in this UML diagram to give a review how the solution can be applied:
No comments:
Post a Comment