Listener descrimination between age differences in two speakers

Often when listeners hear a speaker they have an impression of the age of the speaker. The present research seeks to determine how well listeners are able to identify a speaker’s age when provided with two different recordings from one speaker. Previous research has asked participants to place a speaker’s age on a continuum or estimate the age of the speaker. In the present study, we use an AX task which asks listeners to determine whether the X recording is younger or older, from speech recorded from two speakers spanning approximately 20 years. We also asked a few questions with regard to the listeners’ language background, age, and experience with populations over the age of 40. We analyze the descrimination responses of our listeners, and find that listeners are not much better than chance in determining age differences in the recordings. We do find that as the number of years between recordings increases, participants are more accurate at determining the age difference between recordings. Listeners have fairly specific ideas about what acoustic features they are using to perform their descrimination, however these do not necessarily lead to more accurate descrimination.