Add voice functions (speech synthesis and speech recognition) based on Microsoft's Embedded System)

Based on Windows Embedded standard and Windows Embedded XP, if you need to add the speech recognition and speech reading functions, you need the support of the following components.

Speech Control Panel:

You can add a voice control icon to the control panel. You can use this function to select or configure Speech Recognition (Sr-Speech Recognition) or text-to-speech engine (TTS-text speech conversion engine ). These settings depend on the language, sound output, and sound quality that you can select, as well as the microphone input that you can select. You can create and update user configuration files in the speech attributes. Through the user configuration file, the speech recognition engine can better recognize the habits of a speaker and background sounds in a specific environment.

Text to Speech core:

As a core component of text speech conversion, this component provides a public and basic file for text speech conversion and speech recognition. It is impossible to directly access these dynamic link libraries without voice or speech from any system.

Note: The SAPI component files cannot be directly tested. Because these files require voice services, successful testing features need to include relevant speeches to confirm their existence and current status.

Speech API English Text to Speech engine:

This component contains a device driverProgramTo convert the Text to Speech. This dynamic link library (DLL) cannot be directly accessed by applications, but can be obtained through the voice API (SAPI module. By default, this component provides Microsoft support in English and contains at least one voice. In general, each language requires its own DLL and support files. In addition, at least one reading sound is required for each language. Manufacturers can provide their own language format or conform to the existing Microsoft mode.

Speech API English Text to Speech voice (SAM ):

This component contains a program that inputs or stores the converted text language. You can use this program to describe text in speech. Microsoft provides a default sound-Microsoft Sam. You can purchase audio from independent voice engine manufacturers to more third-party manufacturers and load the audio into the speech recognition system during installation.

Speech API core:

This component provides the voice Application Programming Interface (SAPI module) provided by basic voice functions ). The SAPI. dll file is an integral part of the component and must depend on all the voice function components. The other file of the group, sapisvr.exe, is used for speech recognition. This function cannot be installed after the initial installation of the operating system.

Speech text services V1.1:

This component is a Windows input service based on the text service framework API. Combined with the voice API (SAPI module) and the speech recognition engine, this component provides the dictation and text conversion speech services.

Text services framework:

The text service framework (TSF) is a system service based on the Component Object Model (COM). It provides a simple and scalable framework, to provide advanced text input and natural language technologies. TSF can be enabled in an application or as a TSF text service. TSF supports multi-language text services and provides text services, such as keyboard input services, handwriting recognition and speech recognition.

TSF is a device-independent, language-neutral, and scalable system that provides text and language technologies. TSF applications can use any text service to support TSF.

 

After the following components are added, you can develop speech recognition and speech synthesis applications based on embedded systems.

 

 

 

If you develop a speech recognition engine,

See:

 

Http://blog.csdn.net/yincheng01/archive/2009/02/11/3878807.aspx

If you develop a speech synthesis engine,

See:

Http://blog.csdn.net/yincheng01/archive/2009/02/11/3879130.aspx

 

 

If you develop an embedded speech recognition application, see,

Http://blog.csdn.net/yincheng01/archive/2008/12/13/3511673.aspx

 

If you develop an embedded speech reading application, please take the test,

Http://blog.csdn.net/yincheng01/archive/2008/12/13/3511702.aspx

 

 

 

I provide voice technology project consulting services, please contact MSN: yincheng01@163.com