Application of monaural fusion technique for exploration of potential interaction between channels of phonetic analysis

Authors

  • D.C. Stevenson Dept. of Linguistics, Alberta Univ., Edmonton, Alta., Canada
  • A.J. Rozsypal Dept. of Linguistics, Alberta Univ., Edmonton, Alta., Canada
  • J.T. Hogan Dept. of Linguistics, Alberta Univ., Edmonton, Alta., Canada

Keywords:

speech recognition, monaural fusion, phonetic analysis, speech tokens, monaural presentation, vowels, stop consonants

Abstract

Interaction of detectors in speech recognition has been investigated using a monaural superposition technique which permits simultaneous probing of individual detectors. Waveforms of two natural speech tokens of similar temporal patterns are mixed after careful period-by-period alignment. A stimulus continuum is generated by varying the relative amplitude of the two components. In monaural presentation certain speech stimulus types fuses in perception so that only one of the two components of the mixed stimulus is perceptible. For all tested consonant pairs monaural fusion takes place. In the case of the vowels tested, fusion has not been observed. Instead, both component vowels are perceived simultaneously even for extreme level differences of 26 dB. This suggests that detectors responsible for recognition of vowels are essentially independent, while detectors for stop consonants interact in an inhibitory manner

Additional Files

Published

1987-04-01

How to Cite

1.
Stevenson D, Rozsypal A, Hogan J. Application of monaural fusion technique for exploration of potential interaction between channels of phonetic analysis. Canadian Acoustics [Internet]. 1987 Apr. 1 [cited 2024 Mar. 28];15(2):5-13. Available from: https://jcaa.caa-aca.ca/index.php/jcaa/article/view/574

Issue

Section

Technical Articles