What makes baby cries impossible to ignore?

How can crying babies and barking dogs be harder to ignore than specially designed sirens and fire alarms?

The key may lie in their temporal structure, cleverly optimized for capturing listeners’ attention, and in a particularly irregular, unpredictably changing voice quality.

In this project I will test this hypothesis, which requires both methodological and experimental work with long sequences of human nonverbal vocalizations and animal calls.

Helped by a qualified research assistant, I will improve the existing algorithms for segmenting and analyzing the temporal structure of vocalization sequences, linking temporal irregularities at different time scales to their bottom-up salience (the ability to involuntarily attract attention).

The second task is to develop the first algorithm for automatically detecting different irregularities in voice production known as nonlinear vocal phenomena.

Coupled with experimental manipulation of nonlinearities, this will make it possible to investigate their role in making harsh-sounding vocalizations, such as intense baby cries, so salient and distressing to listeners.

The project will deliver important methodological innovations and theoretical insights into human and animal vocal communication.

It also has immediate practical applications: a better understanding of why some sounds are so distracting and annoying is crucial to managing our acoustic environments.