ABSTRACT

Quantification of the relative amount of additive noise in the voice signal is an essential aspect of the diagnostic and therapeutic protocol with voice disordered patients. Traditional methods (jitter, shimmer, and harmonics-to-noise ratio; HNR) of quantifying aperiodicity in the voice may be error prone due to the necessity of accurately identifying cycle boundaries (i.e. ‘peaks’ or ‘zero-crossings’). As an alternative to the aforementioned traditional voice analysis methods, this study attempted to quantify the degree of additive noise using cepstral analysis.