Speech Technology Magazine

 

LumenVox Improves Speech Engine Detection of Human Voices in Noisy Environments

LumenVox released new Voice Activity Detection (VAD) technology that improves the LumenVox Speech Engine's ability to decipher a human voice from any other sound, such as music, background noise, car noise, and traffic.
Posted Aug 1, 2007
Page1 of 1
Bookmark and Share

LumenVox released new Voice Activity Detection (VAD) technology that improves the LumenVox Speech Engine's ability to decipher a human voice from any other sound, such as music, background noise, car noise, and traffic.

To test the new VAD, LumenVox deliberately chose calls where the previous VAD could not differentiate the speaker from the background noise. These were unusually noisy calls that included various types of loud background noise. The new VAD correctly ignored the noise 92 percent of the time.

"Quality VAD is essential to good recognition accuracy, which is why we are committed to regularly improving this technology," said LumenVox’s chief technology officer, Jim Blake. "This release not only addresses the particular needs of mobile users that make up a large majority of telephone-based speech recognition users, but really any caller that uses speech recognition in a challenging environment."

Human speech has unique qualities that make it distinguishable from other noise. VAD listens to incoming audio and listens for these qualities, such as energy level, frequency and duration. The VAD settings are configurable, so developers can customize them. For instance, one setting controls how long the delay should be between VAD no longer detecting speech and the engine processing the utterance.

Page1 of 1