Perception of vowels with missing formant peaks.

Area of Science:

• Phonetics and Speech Science
• Psychoacoustics
• Auditory Perception

Background:

• Traditional vowel identification models emphasize formant frequencies (F1, F2, F3).
• Previous research suggests listeners may use broader spectral shape cues.
• Ito et al. (2001) demonstrated vowel identification is robust to individual formant peak attenuation.

Purpose of the Study:

• To replicate and extend Ito et al.'s findings on formant peak importance in English dialects.
• To investigate the impact of formant suppression on listener uncertainty and vowel identification accuracy.
• To examine how formant suppression affects perception of steady-state versus dynamic vowels.

Main Methods:

• Replication of experiments from Ito et al. (2001) using two English dialects.
• Quantitative analysis of participant response entropy following formant suppression.
• Experiment using synthesized vowels with dynamic formant frequency changes and peak suppression.

Main Results:

• Vowel identification remained largely possible despite formant peak suppression, consistent with prior work.
• Formant suppression increased participant response entropy, indicating heightened listener uncertainty.
• Suppression reliably altered identification of some vowels, particularly dynamic ones, but not others.

Conclusions:

• Listeners can identify vowels even when specific formant peaks are absent or attenuated.
• Formant suppression can decrease identification certainty and alter perception, especially for dynamic vowel sounds.
• Gross spectral shape cues likely play a significant role in robust vowel perception.