Aperiodicity cue for fricative voicing contrasts

This study examines acoustic cues to voicing contrasts in French consonant–vowel–consonant–vowel sequences containing fricatives. We introduce an aperiodicity metric derived from the YIN difference function and evaluate its ability to capture voicing properties in controlled, noise-free speech materials. Results show that the proposed measure enables reliable, objective classification of fricatives as voiced or unvoiced. When combined with a conventional acoustic feature (root mean square amplitude), the metric provides complementary information and supports accurate classification (approximately 83% overall) among unvoiced fricatives, voiced fricatives, and vowels using standard classifiers. These results demonstrate that aperiodicity-based cues offer a valuable addition to established spectral and amplitude-based features for detailed phoneme characterization.