Acoustic parameters that identify listeners' perceptions of good near-total laryngectomy voice

Ten near-total laryngectomy (NTL) speakers performed a number of voicing tasks that were recorded for acoustic, temporal, and aerodynamic feature analysis. A paired comparison within groups design was used to objectify the perceptual parameters that constitute good near-total laryngectomized voice. The pairs were judged by three speech-language pathologist who chose which subject sounded "most like normal." The judgments were used to assign the NTL speakers into three groups (Most Like Normal, Middle, Least Like Normal). The analysis revealed that listeners could reliably rate NTL speakers as Most Like Normal (MLN). A correlation was found between listener ratings and the acoustic parameter shimmer (r = .684, p = .029). No other significant correlations were noted. A step-wise regression revealed that speaker rating could be predicted by shimmer, H/N ratio, jitter, weight, and flow. These parameters accounted for 93% of the variance.