SRI International Unveils EduSpeak 2.0
SRI International, an independent research institute based in Silicon Valley, announced the release of EduSpeak 2.0, a speech recognition toolkit designed for use in educational software by developers of applications for language learning, reading development, assessment and corporate training.
EduSpeak 2.0 is available for licensing as a software development toolkit for product-based organizations. The toolkit is also available for evaluation.
EduSpeak 2.0 offers:
- · Acoustic models for both adults and children, and for native and non-native English speakers
- · Patented human-calibrated pronunciation scoring at the paragraph, sentence, or word level
- · Capabilities enabling developers to add new words to dictionaries
- · Support for English, Spanish, French and German languages
- · Supports Macromedia Director, Java, and C/C++ developer environments, and any application or script that supports ActiveX, including browser-based JavaScript, Visual Basic applications, and Visual C++ applications.
"EduSpeak 2.0 is designed for developers of education software who want to make their programs more interactive with speech recognition and want the integration to happen in days - not weeks," said Ray Perrault, acting director of SRI's Speech Technology and Research (STAR) Laboratory. "Decades of SRI research in speech recognition and investment in software engineering have led to an easy-to-use, accurate tool that developers and end-users will both appreciate."
The EduSpeak speech recognition engine is based on years of development by SRI's Speech Technology and Research (STAR) Laboratory. The STAR Lab carries out research in speech recognition techniques, including work in the modeling of non-lexical aspects of speech. The staff, which includes engineers, computer scientists and linguists, is active in both technology creation and technology transfer.