Deriving content-specific measures of room acoustic perception using a binaural, nonlinear auditory model

Area of Science:

• Acoustics
• Psychoacoustics
• Signal Processing

Background:

• Traditional acoustic measures (e.g., reverberation time, clarity index) often fail to accurately predict perceptual qualities of rooms.
• Existing physical measures exhibit limitations, including dependence on room occupancy, spatial variability, and poor discrimination of acoustic stimuli.

Purpose of the Study:

• To develop novel signal-based acoustic measures for predicting room acoustic perception.
• To utilize content-specific signal representations from a binaural, nonlinear auditory model.
• To evaluate the efficacy of these new measures for speech and music perception.

Main Methods:

• Derivation of signal-based measures using a binaural, nonlinear model of the human auditory system.
• Generation of content-specific signal representations.
• Conducting listening tests with speech and music stimuli to assess perceptual attributes.

Main Results:

• The proposed signal-based measures demonstrate improved correlation with perceptual attributes compared to traditional measures.
• Listening test results indicate that the new parameters effectively predict aspects of room acoustic perception for both speech and music.
• The auditory model-based approach shows potential for more accurate acoustic quality assessment.

Conclusions:

• The developed signal-based measures offer a more robust method for predicting room acoustic perception.
• The findings suggest that incorporating human auditory system models enhances the prediction of acoustic perception.
• Further research is warranted to refine and validate these promising new acoustic parameters.