Computing interaural differences through finite element modeling of idealized human heads.

Area of Science:

• Acoustics
• Bioacoustics
• Computational Auditory Neuroscience

Background:

• Acoustical interaural differences are crucial for sound localization.
• Previous models often used simplified spherical head shapes.
• Understanding anatomical influences on these differences is key to realistic auditory perception.

Purpose of the Study:

• To investigate how idealized human head and torso shapes affect acoustical interaural differences.
• To compare computational models with experimental measurements.
• To explore the relationship between anatomical features and auditory cues.

Main Methods:

• Finite element modeling of spherical and ellipsoidal head shapes, with and without torso representations (plate or box).
• Calculations of interaural time differences (ITDs) and interaural level differences (ILDs) across frequencies (100-2500 Hz) and source azimuths (10-90 degrees).
• Comparison of model computations with free-field measurements using a manikin.

Main Results:

• The ellipsoid model showed greater frequency variation in ITDs and ILDs compared to the sphere, improving agreement with manikin measurements.
• Incorporating a torso (plate or box) further enhanced agreement by introducing finer frequency variations.
• The study identified correlations between anatomical features (head shape, neck length) and specific interaural difference patterns.

Conclusions:

• Ellipsoidal head models with torso representations provide more accurate predictions of acoustical interaural differences than simple spheres.
• Anatomical details significantly influence the spectral and temporal cues used for sound localization.
• These findings aid in understanding the physical basis of human auditory perception and developing more realistic auditory models.