Formant Synthesis Part 1

Formants.

What are Formants?
In the study of acoustics, speech and phonetics a formant is the part of the audio spectrum, which has a large peak in volume that results from the acoustic resonances formed by the human vocal tract, or by a room or hall.
By their nature when sound waves are created in a room some frequencies will be attenuated, and some will be boosted due to the shape and size of the room. In a room there are not only the direct sound waves that we hear, but also the reflections from the walls, and from furniture. If you want to find out more about room resonances there is an article on Wikipedia.
This also applies to human speech as well due to the mouth and vocal tract forming resonant cavities, only in this case when we create speech we are changing the dimensions of those cavities.

Formants in Acoustics.

In acoustics, a formant is usually defined as a broad peak, or local maximum, in level of the audible spectrum.
For harmonic sounds, the formant frequency is sometimes taken as the harmonic that is most strongly boosted by a natural resonance. A room can be said to have formants which characteristic of that particular room, due to its resonances, which are due to the size and shape of the room, this resonance will also be affected by the contents of the room, and any sound damping materials in the room. Room formants of this nature reinforce themselves by emphasizing specific frequencies and cancelling others. For the purposes of digital signal processing such as reverb, the way a collection of formants generated by a room or hall affects a signal can be represented by an impulse response.

Speech Formants.

In speech the formants are characteristic of the resonances of the the vocal tract which is formed by the following; The lungs, the Larynx, the vocal cords, the throat, the mouth, the palate, the shape and position of the tongue, the lips and teeth all affect the formants.
These formants make up all the sounds we recognize as vowels, and to a lesser degree also make up the sound of the consonants too.
The formant with the lowest frequency is called F1, the second F2, the third F3, and so forth. The fundamental frequency (or pitch)of the voice is sometimes referred to as F0, but it is not a formant. Normally the two first formants, F1 and F2, are sufficient for us to recognize the vowel.
In normal voiced speech, the vibration produced by the vocal cords resembles a sawtooth, rich in harmonic overtones.
If the fundamental frequency (or more often) one of the overtones is higher than one of the resonance frequencies of the system, then the resonances will be weak, and the formant usually produced by that resonance weak or almost completely inaudible. This is usually noticed in the singing of operatic sopranos, who sing at pitches high enough for their vowel sounds to become very hard to distinguish.

Note: Consonants are, to a large degree, noise bursts shaped by the tongue and lips, and we can model these using amplitude contours rather than spectral shapes, so these won’t concern us here.

Approximate frequencies found in vowel sounds. (Not a comprehensive list)

This means that we can use these frequencies in a synthesizer to create roughly human sounding vowels, and “singing”.