Workshop on the Auditory Basis of Speech Perception Keele University, UK July 15-19, 1996

Vowel Segregation with Amplitude Modulation Maps: A Re-Evaluation of Place and Place-Time Models

Georg F. Meyer (1), Frédéric Berthommier (2)

(1) Department of Computer Science, Keele University, Keele, Staffs., UK
(2) Institut de la Communication Parlee, INPG, Grenoble, France

Two fundamentally different models explaining perceptual segregation of simultaneous vowels are investigated: a low-level segregation model based on amplitude modulation maps and a high-level pattern matching model based on the place representation seen in an auditory model. A third model combines the two models and allows a switch between them. Model predictions are compared with human performance and used as a basis to re-examine the relationship between signal representations and pattern matching strategies. Neither the primitive grouping model nor the pattern matching stage alone are able to fully explain human performance while a combination of the two models matches human data well.

Full Paper

Bibliographic reference.  Meyer, Georg F. / Berthommier, Frédéric (1996): "Vowel segregation with amplitude modulation maps: a re-evaluation of place and place-time models", In ABSP-1996, 212-215.