Signal properties that reduce masking by simultaneous, random-frequency maskers

Area of Science:

• Auditory perception
• Psychoacoustics
• Signal processing

Background:

• Simultaneous masking in auditory perception can be substantial, often stemming from informational masking due to stimulus uncertainty.
• Multicomponent maskers with changing frequency content exacerbate this masking effect.

Purpose of the Study:

• To investigate if altering signal properties or presentation mode can enhance signal saliency and mitigate informational masking.
• To determine the effectiveness of different signal types, durations, and presentation modes in reducing masking.

Main Methods:

• Experiments varied the number of masker components (2-100) and used a 1000-Hz sinusoid as the reference signal.
• Signal types included amplitude-modulated (AM), quasifrequency-modulated (QFM), and narrow-band noise (NBN) signals.
• Signal durations (100 ms or 10 ms) and presentation modes (diotic, dichotic, cross-ear) were systematically manipulated.

Main Results:

• AM and NBN signals generally improved performance, while QFM signals sometimes degraded it.
• Dichotic presentation reduced masking more consistently than monaural presentation across listeners and conditions.
• Shortening signal duration (10 ms) yielded the most significant and reliable reductions in masking attributed to masker-frequency uncertainty.

Conclusions:

• Signal duration is a critical factor in reducing masking related to auditory uncertainty.
• While dichotic presentation offers benefits, its advantage may not solely be due to reduced uncertainty.
• Manipulating signal onset/offset times provides a robust method for enhancing signal detection in uncertain masker environments.