Loading...
Please wait, while we are loading the content...
Similar Documents
A psychoacoustical model of the auditory periphery as the front end for ASR
| Content Provider | Semantic Scholar |
|---|---|
| Author | Tchorz, Jürgen Kleinschmidt, Michael Kollmeier, Birger |
| Copyright Year | 1999 |
| Abstract | A psychoacoustical model of the auditory periphery was developed by Dau and others to predict human performance quantitatively in typical spectral and temporal masking experiments. Its main processing stages are basilar membrane filtering for spectral analysis, adaptive envelope compression in each frequency band which enhances changes in the input signal and suppresses steady‐state portions, and low‐pass filtering of envelope modulations. In the field of speech processing, the model was applied as the front end of an automatic speech recognition (ASR) system [J. Tchorz and B. Kollmeier, J. Acoust. Soc. Am. (submitted)]. Speaker‐independent, isolated‐digit recognition experiments in different types of noise were carried out to evaluate the robustness of the auditory‐based ASR system in adverse conditions. Compared to a standard MFCC front end, the auditory‐based preprocessing yielded significantly higher recognition rates in both additive and convolutive noise. Further experiments concentrated on the ques... |
| Starting Page | 1157 |
| Ending Page | 1157 |
| Page Count | 1 |
| File Format | PDF HTM / HTML |
| DOI | 10.1121/1.425500 |
| Volume Number | 105 |
| Alternate Webpage(s) | http://medi.uni-oldenburg.de/members/juergen/asa99_asr.pdf |
| Alternate Webpage(s) | https://doi.org/10.1121/1.425500 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
