Speech middleware has a human touch

Toshiba has developed multilanguage speech synthesis middleware it claims delivers synthesised speech with a clarity and natural intonation that closely approximates the human voice.

Toshiba has developed multilanguage speech synthesis middleware it claims delivers synthesised speech with a clarity and natural intonation that closely approximates the human voice.

The middleware supports nine languages from launch; Japanese, English, American English, French, German, Spanish, Italian, Dutch and Mandarin.

The middleware does not require a large memory capacity or high processing capability, and is ideally suited for integration in such applications as car navigation systems, mobile phones and PDAs.

The market for text-to-speech software is growing in such areas as computing, car navigation and integration in home appliances.

In Japan, most car navigation systems now provide spoken directions, advice and warnings.

Until recently, such systems have relied on recorded narration, which is less efficient in terms of production and requires provision of speech content by professional actors.

The world market for car navigation systems is expanding rapidly, and Toshiba's new middleware is expected to prove particularly attractive to makers in Europe, who can use the same technology in systems destined for different countries.

Toshiba plans to commercialise the middleware by the end of fiscal year 2002.

Toshiba is also creating a new UK-based research team, focusing on the development of multi-language, text-to-speech synthesis, and improved speech recognition technology based at the company's Cambridge Research Laboratory.

The research team will be focused on driving forward improved speech recognition technology and text-to-speech synthesis systems in European languages, and contributing to the development of the multi-language speech-synthesis middleware announced by Toshiba today.

Toshiba's goal is to deliver a versatile system that allows equipment makers to easily integrate speech synthesis and recognition of multiple languages into their products.

The new nine-person team will officially open for business in October 2002, led by Dr Kate Knill, previously Manager of Languages Development at Nuance Communications, where she was in charge of the development of over 20 languages for speech recognition and speaker verification.

Toshiba is determined to reinforce its leadership in speech research on a global basis with an international network of R and D.

Until now speech research has only been promoted at Toshiba's corporate Research and Development Centre in Japan and the Toshiba R and D centre in China.

Back to top