Spectral Enhancement of Acoustic Signals to Provide Improved Recognition of Speech
Inventors: Keith Kluender, Rick Jenison
The Wisconsin Alumni Research Foundation (WARF) is seeking commercial partners interested in developing a method for enhancing an auditory signal to allow hearing-impaired individuals to extract speech information accurately.
Individuals with normal hearing are able to recover speech information even when spatial and temporal overlaps of adjacent speech activities occur. During signal amplification for the hearing-impaired, critical information for recovering speech information is lost, making it extremely difficult for hearing-impaired individuals to extract speech information accurately.
UW-Madison researchers have developed a method for enhancing an auditory signal. The process enhances spectral differences between sounds, minimizing loss of information during amplification of the auditory signal. This is accomplished by dividing the input auditory signal into a plurality of spectral channels, modifying these signals and then combining them to form an enhanced output auditory signal.
- Makes sounds, particularly speech sounds, more distinguishable
- Can be incorporated into a hearing aid device or a speech recognition system
- Several methods can be used to filter the signal into multiple bands, including a bank of gammatone filters and a large bank of filters called Quadrature Mirror Filters.
For More Information About the Inventors