Unattended speech processing: Effect of vocal-tract length

Abstract

Rivenez et al. [J. Acoust. Soc. Am. 119 (6), 4027-4040 (2006)] recently demonstrated that an unattended message is able to prime by 28 ms a simultaneously presented attended message when the two messages have a different F0 range. This study asks whether a difference in vocal-tract length between the two messages rather than a difference in F0 can also produce such priming. A priming effect of 13 ms was found when messages were in the same F0 range but had different (15%-30%) vocal-tract length, suggesting that the processing of unattended speech strongly relies on the presence of perceptual grouping cues.