Fonix Launches VoiceSync 1.0

Fonix Speech today released its VoiceSync 1.0 speech recognition technology that accurately interprets phonetic data and times that information to create “human-like” lip and facial movements.

Fonix VoiceSync 1.0 recognizes phonemes in live or recorded speech and aligns them with their location in the speech audio using a phonetic engine. The technology then interprets text of the utterance to provide  phonemes and their locations or generates phonetic estimates of free-form speech, depending on the mode of operation. The corresponding phonetic data is used by developers of animation products to more closely match lip, tongue, and facial animations to natural human movements.

Fonix VoiceSync is already being used by large technology companies like Autodesk, Epic and OC3, and was recently licensed to Microsoft for “South Park Let’s Go Tower Defense Play!” video game on Xbox LIVE Arcade. 

 Fonix VoiceSync requires a small memory footprint and computing power (starting at 20 MIPS).

“The need for more accurate speech recognition technology in the gaming industry is growing with the birth of more efficient complementary development technologies,” said D. Lynn Shepherd, senior vice president of engineering and product development at Fonix. “Fonix VoiceSync 1.0 represents a new plateau in accurate speech recognition. Beyond gaming, the technology has direct applications in animating avatars for online chat or creating animation features, movies, TV shows, and more.”

Fonix VoiceSync 1.0 is noise-tolerant and insensitive to environment acoustics for a pure data translation. The program operates very close to real time (70-100 ms delay) and has no vocabulary limitations. The software supports languages including U.S. and U.K. English, Canadian and European French, German, Italian, Japanese, Korean, and Castilian and Latin American Spanish. Supported platforms include Windows 32/64, Linux 32/64, Microsoft Xbox 360, Sony Playstation 3, and Nintendo Wii.