You are here

Auditory Cues based Speech Enhancement Processing Technology (ACSEPT)

Award Information
Agency: Department of Defense
Branch: Air Force
Contract: FA9550-10-C-0050
Agency Tracking Number: F09B-T12-0055
Amount: $99,980.00
Phase: Phase I
Program: STTR
Solicitation Topic Code: AF09-BT12
Solicitation Number: 2009.B
Timeline
Solicitation Year: 2009
Award Year: 2010
Award Start Date (Proposal Award Date): 2010-06-01
Award End Date (Contract End Date): 2011-03-31
Small Business Information
20 Burlington Mall Road
Burlington, MA 01803
United States
DUNS: 153869896
HUBZone Owned: No
Woman Owned: No
Socially and Economically Disadvantaged: No
Principal Investigator
 Xuedong Zhang
 Senior DSP Engineer
 (781) 359-9500
 zhang@mayflowercom.com
Business Contact
 Bruce Warwick
Title: Controller & Director of Finance
Phone: (781) 359-9500
Email: warwick@mayflowercom.com
Research Institution
 BU Hearing Research Center
 David C Mountain, Ph.D.
 
44 Cummings Street
Boston, MA 2215
United States

 (617) 353-4343
 Nonprofit College or University
Abstract

Mayflower Communications has partnered with the Boston University Hearing Research Center to propose novel techniques for speech enhancement in reverberant environment based on advanced auditory modeling of higher-level processing in the human auditory brain. Speech enhancement or noise suppression has been an area of active research for well over three decades. State-of-the-art single-microphone methods have been shown to be adequate for reducing stationary background noise without introducing audible signal distortions; however, they cannot reduce non-stationary background noises, signal-correlated interferences like echoes and sound reflections in a reverberant environment or interfering background speakers, all of which can occur in real-world use scenarios. The development of the “Auditory Cues based Speech Enhancement Processing Technology” (ACSEPT) will address these and other shortcomings. It leverages existing, US Government-funded projects at Boston University as technical foundation; particularly the BU EarLab project. The EarLab is an auditory modeling environment that provides access to a wide range of models covering many aspects of the auditory function. This effort will specialize, enhance, and augment the EarLab capabilities to meet the desired auditory modeling objectives. Based on the output of the EarLab models, we will derive pertinent auditory cues based on principle of Auditory Scene Analysis which will then be used to identify, and later implement and demonstrate, novel techniques for efficient speech enhancement under conditions of noise and reverberation. BENEFIT: Dealing with speech production in high level of noise and reverberant environment affect our daily lives. Many common occurrence applications would benefit tremendously from technologies, such as ACSEPT, created to efficiently address this problem. The success of ACSEPT, based on mimicking the human brain auditory functions, could present a significant shift compared to the technologies being used today. The potential commercial applications are wide, especially if this technology could be successfully transitioned to the mobile voice communications user devices market; a viable target market for it. Today there are over 3 billion such device in use!

* Information listed above is at the time of submission. *

US Flag An Official Website of the United States Government