BaiduASRAndTTS

Constructure

Interface and Class

• SpeechModel: A class contains almost all the parameters the Baidu API need to achieve ASR and TTS.

• WavInfo: A struct contains the basic element of one Wave object.

• AutomaticSpeechRecognition: Class does the 'Get access token' and 'recognition' work, mainly send HTTP post request.

• ClassUtils: Some utility methods grouped together.

• ISpeechRecorder: Interface, define 3 methods to be implemented: SetFileName, StartRec, StopRec.

• DirectRecorder and NAudioRecorder: Two different implementation of ISpeechRecorder.

• Form: Handle most UI controls' action, response to button action, and other event.

UI

Features

1. Select a file to recognize.
2. Speech---> Text: Record from user and translate speech to text.
3. Text---> Speech: Read the text in textBox.
4. Pause and Stop.
5. Drag and drop text file into the text area.

Note

• The key and secret key has been erased from the source code, so you need to add your own when the token expired.
• The token.dat in bin/Debug, contains the generated token(first line) and the time it was generated(second line).
• When using the DirectRecorder version to do record work, set Exception Settings(Ctrl+Alt+E)--->Managed Debugging Assistants--->LoaderLock(Uncheck this).

Reference

• http://yuyin.baidu.com/docs
• http://blog.csdn.net/u013810234/article/details/54604355