Formant

A formant is a concentration of acoustic energy around a resonance frequency of a sound-producing system. In speech, formants arise from vocal-tract resonances and help distinguish vowel sounds.

Vocal tract: The airway above the larynx, shaped by the tongue, lips, jaw, and pharynx during speech. Its changing shape sets the resonances that produce speech formants.
Vowel: A speech sound produced with an open vocal tract and no major constriction that blocks airflow. Vowel identity is strongly associated with characteristic first- and second-formant frequencies.
Harmonic: A sinusoidal component whose frequency is an integer multiple of a sound’s fundamental frequency. Harmonics are source components; formants are resonances that shape their relative amplitudes.
Nasal consonant: A consonant produced with airflow passing through the nose while the oral passage is obstructed. Nasal coupling adds resonances and antiresonances to the vocal-tract spectrum.
Source–filter model: A model of speech production in which a sound source is shaped by a resonant filter. It separates vocal excitation from the tract resonances that create formants.
Vowel space: A plot of vowel sounds by their first and second formant frequencies. It maps how formant patterns distinguish vowel categories.
Fundamental frequency: The lowest frequency of a periodic signal, corresponding to its repetition rate. Pitch-related fundamental frequency can change without moving the vocal tract’s formant resonances.
Singing formant: A spectral concentration in some trained singing voices that enhances audibility over an orchestra. It is a distinctive formant cluster associated with operatic singing.
Linear predictive coding: A signal-processing method that estimates each sample from a weighted combination of previous samples. It estimates vocal-tract resonances from speech and is widely used to locate formants.
Formant synthesis: Speech synthesis that creates sound by combining a source signal with resonant filters. It uses controllable formant frequencies and bandwidths to generate speech.

Linked from 15 pages