Crosscorrelation of adjacent spectra enhances fundamental frequency tracking

All fundamental frequency estimators based on spectral analysis rely heavily on a proper harmonic selection of the voice analyzed. Since in practice other spectral peaks pertaining to sources external to the speech considered may be present in the signal, various schemes have been designed to ensure a satisfactory elimination of non pertinent harmonics. This paper introduces a new harmonic selection algorithm based on the brush method, which ensures a very good selection of voice harmonics, and gives satisfactory results even in the presence of another speech source. The algorithm has been implemented and incorporated in the speech analysis program WinPitch, and tested on various examples provided in the Speech Separation Challenge.