VOSIM synthesizes vowel formants using decaying squared-sine pulse trains

VOSIM (Voice Simulation) synthesis generates formant spectra through trains of decaying sin-squared pulses. Each toneburst in the train produces a strong formant component. The shape of each pulse (a sin-squared function - a half-cycle of a sine wave squared) creates a spectrum with a specific formant peak. Multiple VOSIM generators, each tuned to a formant frequency, can simulate vowel sounds. The decay rate of successive pulses controls the bandwidth of each formant. VOSIM was developed at the Institute of Sonology, Utrecht, and offers computational efficiency combined with natural-sounding vowel synthesis.

Examples

VOSIM for vowel /e/: pulse width = 0.8ms (formant at ~1250 Hz), 4 decaying pulses per pitch period, fundamental = 150 Hz.

Assessment

What type of waveform does VOSIM use for each pulse, and how does this differ from a simple sine wave grain? What aspects of vocal sound does VOSIM model well?