Aculab to Unveil Latest Text-to-Speech Software

Aculab, a provider of enabling technology for the communications market, will unveil and demonstrate at SpeechTek 2002, the latest version of its host based text to speech (TTS) software. Aculab TTS V3.0 now supports the W3C standard speech synthesis mark-up language (SSML) as well as offering new languages, new voices and a talking lexicon tool. Aculab TTS offers improved accuracy and pronunciation and features full release versions of Brazilian Portuguese and Italian as well as new British English male and Spanish female voices. Aculab TTS also brings benefits such as non-declarative intonation, allowing interrogative statements to be produced, as product manager at Aculab, Ian Colville points out. "With TTS V3.0, the right intonation can be given to a phrase to make it sound like a question rather than a statement - for example - 'You want a ticket to Denver?'" he says. "You can also have phonetic transcription, which allows you to add a tag to the text to provide the correct pronunciation of an ambiguous word," he explains. "For example - that the name 'Smyth" should be pronounced "Smith" - and not "Smythe". Aculab's new TTS V3.0 brings improved text normalisation and enhanced email pre-processing. "For example, the sequence of numbers '10-4' can be stated as 'ten four', 'ten minus four', 'fourth of October' or 'tenth of April' with the addition of simple tags," points out Ian Colville. Aculab's support of SSML also allows systems integrators and developers to change voice or language within the text, as well as control pitch, speech rate, intonation and emphasis. "These features are particularly useful when combining synthesised speech with pre-recorded speech, or customising voices for specific target audiences," says Ian Colville. Aculab's TTS V3.0 also automatically provides the prosodic phrases for each sentence, whatever the language, creating intonation contours and assigning an appropriate duration to each sound. Aculab TTS uses client/server architecture and works in conjunction with Aculab's Prosody digital signal processor (DSP) resource cards, which are used for audio replay and other simultaneous DSP resource functions such as echo cancellation. As a result, developers can enhance the performance of their telephony applications, combining TTS with speech recognition and separate speech processing algorithms through the generic Aculab application programming interface (API). "Because Aculab's unique method of software distribution allows developers to download software with a cost free licence from our website, they can readily achieve targets in an extremely cost effective manner," Ian Colville concluded. If you would like further information on Aculab TTS, Aculab representatives will be available at SpeechTek 30-31st October 2002, on stand number 205, New York Marriott Marquis Hotel, New York.