A sinewave (SWS) replica of a natural utterance discards the fine-grain acoustic properties of speech, retaining only the coarse-grain changes in the spectra over time. This pattern of spectral changes generally is estimated by linear prediction (LPC) to pick the frequencies and amplitudes of the formants.

SPECTROGRAMS OF THE NATURAL WORD  ("BORDE") AND THE SINEWAVE REPLICA

Here three pure tones (sinewaves) rise and fall in imitation of the frequency and amplitude variations of three vocal resonances (formants) when the speaker says the Spanish word "borde" [edge].

Spanish Sentences

Back