IBM and Linguatec Further Develop Speech Recognition
New Version Voice Pro 11
There is new impetus in the field of speech recognition. In close cooperation, IBM and Linguatec have further developed ViaVoice technology. Voice Pro 11, the new edition of the Stiftung Warentest winner, has once again been improved and has been released just in time for SYSTEMS 2006. The recognition rate has been significantly improved by extending speaker-independent speech recognition, by optimizing the software precisely for modern PC processors and by adjusting it perfectly to the Sennheiser headset included in the package.
Extended Acoustic Models
In a nationwide effort, Linguatec collected hundreds of speech models
which were processed by the mainframe computers at the IBM
development center in Böblingen and then integrated into Voice Pro
11. In connection with the project “Superhuman Speech Initiative” IBM
has been researching speaker-independent speech recognition for
years. By 2010 voice recognition should reach a near-human
performance level. Dr. Carsten Günther, head of speech recognition
technology at the IBM development center, confirms: "The more voice
data that can be integrated, the better the recognition results. The
numerous acoustic models that have expanded Voice Pro 11
noticeably improve the recognition rate and are an important step
toward speaker independence."
Optimal Use of Modern CPU Performance
Adjusting the software to the system architecture of modern Intel and AMD processors of 1.5 GHz and higher has a double positive effect: it improves the recognition rate and at the same time increases the processing speed.
Select Hardware Support from Sennheiser
Thanks to cooperation with the audio specialist Sennheiser Communications, Linguatec was also able to make crucial improvements in the new product’s hardware: the digital headset PC- 25 USB included with Voice Pro 11 was developed and optimized specifically for the demands of modern speech recognition software. "The quality of the microphone must meet the highest standards for transmitting voice data to the software. Even hardly noticeable disruptions like static or background noise can have fatal consequences for recognition quality," explains Jesper Kock, head of development at Sennheiser Communications.
"Progress in software and hardware technology now makes recognition rates possible that we could only dream of a few years ago,” Dr. Reinhard Busch, CEO of Linguatec, describes the rapid development of speech recognition technology. “Based on the performance of the new version Voice Pro 11 we are confident that the implementation of Speech Recognition in the workplace will become just as commonplace as it is indispensable."
Linguatec GmbH is Germany’s leading producer of language technology programs for office use. Its main product lines are voice recognition, automatic translation and voice output. Linguatec is the only company ever to have received the European Information Technology Prize three times. http://www.linguatec.de
About IBM Development:
IBM Germany Development Ltd. is one of IBM’s largest innovation centers worldwide. The Böblingen language technology team works in close cooperation with the worldwide IBM language development team and IBM Research on speech recognition and speech synthesis for a wide variety of applications.
The Sennheiser Group is one of the world’s leading producers of microphones, headphones and wireless transmission systems. Numerous patents and awards, such as two German economic innovation awards, the Scientific and Engineering Award, the Emmy and the Grammy, bear witness to the company’s power of innovation. http://www.senncom.de
Frau Elisabeth Bauer
Tel.: +49-89-89 66 64-152