Simultaneous relative cue reliance in speech-on-speech masking

Area of Science:

• Auditory Perception
• Psychoacoustics
• Speech Processing

Background:

• Listeners segregate simultaneous speech using fundamental frequency, level, and location cues.
• Few studies examine simultaneous cue reliance or relative cue weighting in naturalistic listening.
• Understanding cue reliance is crucial for interpreting speech-on-speech masking research.

Purpose of the Study:

• To quantify simultaneous reliance on fundamental frequency, level, and location for speech segregation.
• To compare listener reliance to an ideal observer model.
• To evaluate the implications for speech-on-speech masking research methodologies.

Main Methods:

• Trial-by-trial analysis of listener performance in a speech segregation task.
• Presentation of simultaneous, naturally spoken sentence recordings with varying target talker cues.
• Comparison of listener cue reliance against an ideal observer standard.

Main Results:

• Only one of 33 normal-hearing adults predominantly used voice level.
• Most listeners relied on voice fundamental frequency and/or location cues.
• Significant individual variability exists in cue weighting for speech segregation.

Conclusions:

• Listeners' speech segregation predominantly relies on fundamental frequency and location, not just level.
• Current practices in speech-on-speech masking research may overemphasize level.
• Further research should consider the complex interplay of multiple auditory cues in masking studies.