Finding speech formants and modulations via energy separation: with application to a vocoder

An experimental system that uses an energy-tracking operator and a related energy separation algorithm to find speech formants and amplitude/frequency modulations automatically in voiced speech segments is presented. Initial estimates of formant center frequencies are provided by either LPC (linear predictive coding) or morphological spectral peak picking. These estimates are improved by a combination of bandpass filtering and iterative application of energy separation. The system is shown to be effective. Its application to an AM-FM vocoder is also discussed.<>