Instrumental Quality Predictions and Analysis of Auditory Cues for Algorithms in Modern Headphone Technology

Area of Science:

• Acoustics and Signal Processing
• Psychoacoustics and Auditory Perception

Background:

• Smart headphones (hearables) employ algorithms like noise cancellation that can inadvertently degrade audio quality by introducing monaural or binaural distortions.
• Evaluating these distortions via subjective listening tests is time-consuming and expensive, necessitating objective, instrumental measures.

Purpose of the Study:

• To evaluate the effectiveness of existing monaural and binaural audio quality models in predicting distortions common in hearables.
• To investigate how hearable signal processing affects spectral, temporal, and binaural auditory cues.

Main Methods:

• Applied various auditory-inspired instrumental audio quality models to assess distortions introduced by hearable signal processing algorithms.
• Examined the impact of these distortions on monaural (spectral, temporal) and binaural auditory cues.

Main Results:

• Signal processing in hearables primarily impacts monaural spectral cues, leading to perceived audio quality degradation.
• Monaural audio quality models focusing on spectral distortions demonstrated superior prediction accuracy.
• A revised audio quality model incorporating binaural aspects achieved a high correlation (0.89) between predicted and subjective ratings.

Conclusions:

• Monaural audio quality models are crucial for accurately assessing distortions in hearables, particularly those affecting spectral information.
• Objective, auditory-inspired models offer an efficient alternative to subjective testing for evaluating hearable audio quality.
• Further refinement of audio quality models, including binaural aspects, enhances their predictive power for hearable applications.