Identification of perceptually relevant methods of inter-aural time difference estimation.
Level 4 - case-series / case-control
Laboratory psychophysical study evaluating algorithm performance against human perceptual judgments without randomized clinical comparator
PubMed 28863557 · doi:10.1121/1.4996457
What was done
A subjective lateralization psychoacoustic study was conducted to evaluate which objective inter-aural time difference (ITD) estimation algorithms best match human perception when applied to head-related impulse response (HRIR) data. Perceptually evaluated inter-aural pure delay offsets were compared to various objective estimation techniques, including onset thresholding, inter-aural cross-correlation methods, and phase-based approaches.
What was found
The abstract reports no exact numerical values, error metrics, or statistical intervals. Qualitatively, a first-onset threshold detection method using a -30 dB relative threshold applied to 3 kHz low-pass filtered HRIRs was identified as consistently the most perceptually relevant across metrics. Inter-aural cross-correlation methods (maximum or centroid) using filtered HRIRs or envelopes provided reasonable performance, whereas phase-based methods (integrated relative group delay or auditory models) performed poorly.
Why it matters
Standardizing ITD estimation to the most perceptually aligned algorithm helps minimize localization artifacts when rendering personalized binaural audio in virtual acoustic environments.
Limits
The abstract omits the sample size, listener characteristics, number of HRIR datasets tested, and all quantitative metrics. Generalizability to non-filtered conditions, diverse acoustic environments, or listeners with hearing impairment was not reported.