Formant

(redirected from Formant synthesis)
Also found in: Dictionary, Medical.

formant

[′fȯr·mənt]
(acoustics)
A set of resonances of a musical instrument or voice mechanism that form partials of sounds produced by the instrument, independent of the fundamental frequency, and give these sounds their quality.

Formant

 

a term used in phonetics to denote an acoustic characteristic of speech sounds, especially vowels; formants are associated with the frequency of the voice and are responsible for the timbre of the sound. A formant is part of the frequency spectrum of a sound, which can be obtained in the laboratory by means of a sound spectrograph. The formant represents a range of frequencies defined with respect to an average frequency and designated by the letter F.

In a sound spectrum, several formants may be isolated, for example F1 = 500 hertz (Hz), F2 = 1,500 Hz, and so on. For men, the average interval between formants is 1,000 Hz; for women and children it is somewhat greater. In most cases, however, the first two formants are sufficient to distinguish vowels. F1 (150–850 Hz) corresponds to the articulatory feature of height, or the difference between close and open vowels, and is lower for close vowels. F2 (500–2,500 Hz) corresponds to the degree to which a vowel is front or back, being higher for front vowels and lower for back vowels. The sum of the frequencies F1 + F2 corresponds to the feature of roundedness; labialization causes a reduction in the frequencies corresponding to F1 and F2.

REFERENCES

Zinder, L. R. Obshchaiafonetika. Leningrad, 1960.
Fant, G. Akusticheskaia teoriia recheobrazovaniia. Moscow, 1964. (Translated from English.)

V. A. VINOGRADOV

References in periodicals archive ?
While neural network-based synthesis has been demonstrated, these methods so far have failed to approach the quality of formant synthesis or concatenation.
Centigram's TruVoice TTS converter uses a patented formant synthesis, rules-based form of TTS to produce clear speech for end users.