MINIMAL SPECTRAL CONTRAST OF FORMANT PEAKS FOR VOWEL RECOGNITION AS AFUNCTION OF SPECTRAL SLOPE

In four experiments we investigated whether listeners can locate the f ormants of vowels not only from peaks, but also from spectral ''should ers''-features that give rise to zero crossings in the third, but not the first, differential of the excitation pattern-as hypothesized by A ssmann and Summerfield (1989). Stimuli were steady-state approximation s to the vowels [a, i, e, u, o] created by summing the first 45 harmon ics of a fundamental of 100 Hz. Thirty-nine harmonics had equal amplit udes; the other 6 formed three pairs that were raised in level to defi ne three ''formants.'' An adaptive psychophysical procedure determined the minimal difference in level between the 6 harmonics and the remai ning 39 at which the vowels were identifiably different from one anoth er. These thresholds were measured through simulated communication cha nnels, giving overall slopes to the excitation patterns of the five vo wels that ranged from - 1 dB/erb to + 2 dB/erb. Excitation patterns of the threshold stimuli were computed, and the locations of formants we re estimated from zero crossings in the first and third differentials. With the more steeply sloping communication channels, some formants o f some vowels were represented as shoulders rather than peaks, confirm ing the predictions of Assmann and Summerfield's models. We discuss th e limitations of the excitation pattern model and the related issue of whether the location of formants can be computed from spectral should ers in auditory analysis.