| Autor: |
Donai, Jeremy J. |
| Přispěvatelé: |
Dembowski, James, Hicks, Candace B., Lee, Sue Ann, Shahriar, Muneem, Paschall, Dwayne |
| Jazyk: |
angličtina |
| Rok vydání: |
2013 |
| Předmět: |
|
| Zdroj: |
IndraStra Global. |
| ISSN: |
2381-3652 |
| Popis: |
The traditional bandwidth presumed to be required for accurately identifying speech, vowels in particular, focuses primarily on frequencies below 3000-4000 Hz. Early research showed that normal-hearing listeners were unable to identify vowel category above chance level accuracy for naturally produced vowels filtered to remove information below 3500 Hz. Recent evidence has highlighted the importance of frequency information well above this 100-3000 Hz region in the recognition of speech. The experiments in this dissertation project were undertaken to evaluate the perceptual utility of high frequency information (above 3000 Hz) in naturally produced vowels. It was of interest to determine if listeners could utilize information in this high frequency spectral region to perform perceptual tasks including vowel category identification, determination of gender (i.e., male vs. female), and mode of production (i.e., spoken vs. sung). Results of the experiments provided evidence that high frequency information can be used in the perceptual processing of speech in absence of low frequency spectral information. Implications of these findings are discussed. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
