Audiosieve

Audiosieve ^TM brings a radical new approach to voice interfaces that allows machines to hear at a human level. It is a unique software technology that separates voice from ambient sounds in everyday noisy environments.

Voice based computing is fast becoming a part of everyday life. VoIP systems provide a low cost solution for voice communication and conferencing. Increasingly effective speech recognition software lets people use voice to browse the web, navigate menus, send email, and request driving directions using a growing array of devices.

Currently, however, users of voice based applications are limited to quiet offices and often must wear microphone headsets to reduce noise. Environments such as a car, restaurant, airport, or hotel lobby contain loud ambient sounds, reverberations and echoes. Those effects significantly reduce voice intelligibility in communication systes. They also confuse speech recognition engines and sharply degrade their accuracy.

Audiosieve alleviates this problem by using smart machine learning algorithms. These algorithms are alble to extract the voice one wants to focus on from a noisy environment, and suppress all other sounds. They are based on a novel, proprietary class of adaptive voice processing techniques. Audiosieve automatically adapts to the acoustic environment and the microphone array configuration. It can use arrays with any number of microphones. Audiosieve significantly outperforms current noise suppression techniques.

Contact:    global (at) goldenmetallic.com