LumenVox Releases v6.5 of Their Speech Recognition Engine

Advanced Acoustic Models Improve Accuracy and Reduce False Positives

San Diego, CA (APRIL 25, 2006) — LumenVox, an innovator of speech recognition technology, announced today the general availability of version 6.5 of their award–winning Speech Engine. Driven by the commitment to meet or exceed competitors' performance, version 6.5 introduces new acoustic models that benefit developers requiring high–accuracy speech recognition performance in an affordable, full–featured speech recognition software solution.

A considerable reduction in false positives and accuracy improvements contribute to an overall performance increase from the previous version, and solidly lands LumenVox on equal footing with other leading speech recognition engines. "Our development focus on recognition accuracy stems from a combination of industry and customer demand and a continuing commitment to innovation and precision in the area of speech recognition," said Jim Blake, CTO of LumenVox.

Further 6.5 improvements include a new distributed licensing system that permits a license server to reside on any available server, instead of being co–located with the speech server, increasing flexibility and fault–tolerance. Resellers also benefit by allowing specific field system version upgrades on an individual basis instead of only as a group.

In addition, the LumenVox Speech Tuner — a complete maintenance tool designed to perform tuning and transcription, as well as instant parameter, grammar and version upgrade testing of any speech application — will now benefit from faster batch test speed and support for large grammar testing.