(1993) Edwards, Thomas G.; Shamma, Shihab; Electrical Engineering; Digital Repository at the University; University of Maryland (College Park, Md)
Computer models of cochlear processing take exceedingly long times to run,
even for short data sets. A data acquisition system was developed for a new
switched-capacitor VLSI cochlea model chip, in order to rapidly perform cochleaI
processing on digitzed speech samples. The system is capable of processing very
long speech samples. Processing is in near-real-time, though it, takes about 2
minutes per second of speech to write the large amount of data to a hard drive.
Software has also been developed to convert the output data into a form readable
by the ESPS digital signal processing package from Entropic Speech, Inc.
