・High Quality Speech Synthesis System Using Speech Rate Conversion for Stock-price Bulletins [NAB2007]
NHK has been developing a high quality speech synthesis system that realizes natural-sounding speech through the use of a massive speech database, consisting of recordings of an announcer reading from past news programs. The system selects the combination with the smoothest acoustic flow of words and phrases to compose an entire sentence based on the database. A prototype voice synthesizer using numerical speech synthesis and automatic speech rate conversion is currently used for stock-price bulletins in the experimental broadcasting of digital terrestrial radio.
・Real-time Closed-captioning using Speech Recognition [ABU2007]
There is a great need for more TV programs to be closed-captioned to help hearing impaired and elderly people watch TV. Automatic speech recognition techniques are expected to contribute to providing text from speech in real-time. NHK has been using speech recognition for closed-captioning of some of its news, sports and other live TV programs. To efficiently expand closed-captioning, a new hybrid speech recognition system that switches input speech between the original program sound and the rephrased speech with fewer correction operators is under study.
