Perceived Azimuth of Isolated ITD and ILD Cues at LowFrequencies

Gao M, Wang H, Xu Y. Perceived Azimuth of Isolated ITD and ILD Cues at LowFrequencies. In: Express Paper. Audio Engineering Society; 2026. Paper 416. Available from: https://aes.org/publications/elibrary-page/?id=23252

@inproceedings{Gao2026_23252,
  author = {Gao, Mingyan and Wang, Heng and Xu, Yiming},
  title = {{Perceived Azimuth of Isolated ITD and ILD Cues at LowFrequencies}},
  note = {Paper 416},
  year = {2026},
  month = may,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=23252}
}

TY  - CPAPER
TI  - Perceived Azimuth of Isolated ITD and ILD Cues at LowFrequencies
AU  - Gao, Mingyan
AU  - Wang, Heng
AU  - Xu, Yiming
M1  - Paper 416
PY  - 2026
DA  - 2026/05/28
UR  - https://aes.org/publications/elibrary-page/?id=23252
PB  - Audio Engineering Society
LA  - en
AB  - Inter-ear time difference (ITD) and inter-ear level difference (ILD) are central to parametric spatial audio coding; however, their perceptual characteristics in the low-frequency rangeparticularly the relationship between the tworemain unclear. This study aims to systematically quantify and compare how low-frequency ITD and ILD cues map to perceived azimuth. Through subjective listening experiments, we measured the perceived azimuth elicited by 9 ITD values and 12 ILD values (when the other cue was zero) across the 3001480 Hz frequency range. The experiments employed an adaptive staircase procedure, in which participants were asked to match a test tone synthesised from ITD/ILD with a reference tone of known physical location. The results indicate that the perceived azimuth of both ITD and ILD exhibits strong frequency dependence, remarkably consistently, a minimum in perceived azimuth occurs around 700 Hz (i.e., the sound image is closest to the central axis). Furthermore, by analysing ITD-ILD combinations capable of producing equivalent perceived directions, this study confirmed the existence of quantifiable interactions between the two cues at the perceptual level. These findings provide key data for optimising perceptual models of spatial audio coding and lay the foundation for achieving higher-fidelity spatial audio reproduction.
ER  -