There is a direct correspondence between a set of formant frequencies and a vowel.

The synthesized vowel gives a perception of the expected vowel, except for the loss of naturalness.

The loss of naturalness is mainly due to the approximation of the speech production mechanism by a source-system model.

The system is approximated as an all-pole system which is reasonable for vowel sounds which do not have significant anti-resonances.

The excitation source represented by a train of impulses is not a good approximation and better models of excitation are required to gain naturalness.