Identification and Discrimination of Stop Consonants: Formants Versus Spectral Peaks

Identification and discrimination of three-formant, stop-consonant approximations were measured for several types of stimuli. Stimuli were synthesized using: 1) pure tones, 2) formant filters swept across a harmonic voicing-source spectrum, and 3) formant filters swept across a frication (noise) source to approximate whispered speech. Despite the fact that the pure tones represented the formant frequencies with the greatest spectral precision, discrimination and identification of the stimuli was more accurate for the broader spectrum “source and formant” stimuli.