Perceptual and acoustic characteristics of pop-out voicea)

The present study investigated characteristics of pop-out voice, i.e., a type of voice that stands out in the presence of other competing sounds. First, Japanese listeners heard a Japanese sentence produced by various speakers mixed with babble noise and rated the subjective degree of pop-out. Results showed wide variation in pop-out scores across the sentences, despite normalized intensity. Second, detection thresholds were measured by presenting the sentences to listeners at various speech-to-noise ratios. Results demonstrated that sentences with higher pop-out rank were more easily detectable than those with lower pop-out rank, both for Japanese listeners and for English listeners with no proficiency in Japanese. Furthermore, detection thresholds were lower for Japanese than English listeners. Third, intelligibility was measured with Japanese listeners using the same sentences. Results confirmed that sentences with higher pop-out ranks were more intelligible. Furthermore, correlations between intelligibility and detection thresholds were moderate. Finally, acoustic analysis revealed that pop-out voice was associated with higher relative power in the 0.85-4.0 kHz frequency band. These results suggest that pop-out voice is driven by acoustic information, which enhances perceptual salience, and by linguistic information, which further enhances detectability for native listeners.