June 21, 2001 - Voiceware announces that it has developed ZVSR, a Digital Signal Processor(DSP) designed specifically for speech recognition application in electronic speech products, cooperating with Zaram Technology, a Korean DSP chip developer.
Using a prominent hidden Markov Model algorithm, ZVSR takes only under one minute for making word library, compared to competitors’ 2-3 months of time requirements to update words. And it reaches an accuracy of greater than 97% for speaker independent recognition and greater than 99% for speaker dependent recognition.
Based on Phoneme-like-unit algorithm, anyone who wants to make a word library, just types words to Voiceware’s word library generator software, then one can download word library to ZVSR.
HMM algorithm based recognition chips have disadvantage in price competitiveness; however, ZVSR will be sold at the low price as chief as a chip based on the Neural Network (NN) because Voiceware succeeded to optimize HMM recognition engine with price competitive DSP-core. The ZVSR will be available for shipping from this July.
About Zaram Technology (www.zaram.co.kr)
Zaram Technology is a global leader in the development and marketing of high-performance, low-power, cost-effective digital signal processing cores used in a wide range of applications such as high quality audio/sound application, wireless communications, and personal computers. Zaram Technology has developed a family of low power, low cost DSP core architectures that are suitable for multimedia, mobile computer, and wireless communications products. All DSP cores are highly modular and designed for customers developing their own high-volume, application-specific devices.
About Voiceware Co., Ltd.(www.voiceware.co.kr)
Voiceware is the leading speech technology provider that develops and markets a speech interface for network and telephony solutions, embedded devices and PC applications. Voiceware is offering VoiceText™, high quality text-to-speech technology, VoiceEz™, natural speech recognition technology and VoiceCop™, speaker verification technology to many leading companies and administrative organs like Microsoft, IBM, Lycos Asia, Hyundai Motor, Asiana Airlines, Korea Telecom, SK Telecom, SK Securities, HeyAnita Korea and Korea Meteorological Administration. Voiceware is headquartered in Seoul, Korea and has US office in San Jose, CA. For additional information, visit www.voiceware.co.kr or send e-mail to firstname.lastname@example.org