Acoustic parameters of speech produced as a function of background and task. Values shown are means over talkers and error bars, here and elsewhere, indicate 95% confidence intervals. Background conditions are quiet (Q), speech-modulated noise (SMN), competing speech (CS) and stationary speech-shaped noise (SSN).
Mean individual talker F0 differences between communicative and non-communicative tasks for each background.
Foreground-background overlap percentage as a function of task and background. Values shown are averages across the 8 talkers. The leftmost bar shows the degree of overlap for simulated talkers described in Sec. ???.
(a) Speech rate and (b) mean pause duration for speech produced as a function of task and background.
Example pause length (left) and contribution length (right) densities for a single talker in quiet (top), speech modulated noise (middle) and a competing talker background (bottom). The horizontal axis shows duration in seconds. Gamma density fits are also plotted with shape and rate values shown.
(a) Between-category dispersion and (b) within-category dispersion for a subset of vowels produced in the 4 types of background (quiet and 3 noisy backgrounds) with or without a communication factor. Values shown are averages across the 8 talkers.
Noise and task conditions in each recording session.
Article metrics loading...
Full text loading...