A Detection-Theoretic Analysis of Auditory Streaming and Its Relation to Auditory Masking

Area of Science:

• Auditory perception
• Psychoacoustics
• Information theory

Background:

• The cocktail party problem highlights the challenge of isolating individual sounds in complex auditory environments.
• Auditory streaming and masking are two primary research approaches to understanding auditory scene analysis.
• The analytical relationship between streaming and masking has not been previously established.

Purpose of the Study:

• To provide a detection-theoretic analysis of the relationship between multitone auditory streaming and masking.
• To predict the similarities and differences between streaming and masking phenomena.
• To evaluate the outcome of experimental studies on these auditory processes.

Main Methods:

• Utilized a detection-theoretic framework to analyze auditory streaming and masking.
• Employed sequences of tones as stimuli in experimental paradigms.
• Investigated the function linking listener performance to the information divergence (DKL) of tone sequences.

Main Results:

• Established a common functional relationship between auditory streaming and masking based on DKL.
• Demonstrated that performance in both streaming and masking tasks is predictable using DKL.
• Experimental results support the prediction of a shared dependency on DKL under symmetric sequence properties.

Conclusions:

• Auditory streaming and masking are analytically linked through the information divergence (DKL) of tone sequences.
• The findings provide a unified framework for understanding these fundamental auditory phenomena.
• Future research can leverage this DKL-based relationship to further explore auditory scene analysis.