Energetic and informational masking in a simulated restaurant environment

Area of Science:

• Auditory Perception
• Acoustic Psychology
• Speech Processing

Background:

• Investigating speech intelligibility in complex auditory environments is crucial for understanding human communication.
• Virtual reality offers a controlled method for simulating realistic acoustic scenarios like restaurants.

Purpose of the Study:

• To measure speech reception thresholds (SRTs) in a simulated restaurant with varying numbers and types of interfering sound sources.
• To evaluate the predictive accuracy of existing auditory models in complex social listening conditions.

Main Methods:

• Participants listened to a target voice in a virtual restaurant over headphones.
• Interfering sounds included speech-shaped noise or competing speech at different tables.
• The virtual environment was either anechoic or reverberant.

Main Results:

• A spatial release from masking model accurately predicted SRTs for speech-shaped noise.
• Competing speech caused more masking than speech-shaped noise, particularly with multiple or similar voices.
• Differences in fundamental frequency and dip listening explained some effects of single competing voices.

Conclusions:

• Current auditory models need refinement to accurately predict intelligibility in complex, noisy social environments.
• Informational masking plays a significant role when multiple competing speech sources are present.
• The characteristics of interfering speech (e.g., voice similarity) significantly impact speech reception thresholds.