Acoustic Fidelity and Perceptual Learning: The Impact of Early Reflections on Virtual Reality Auditory Training
일시11/27 금 · 10:00~10:40
발표장컨퍼런스홀
- 저자
- Viktoriia Gutor, Koh Pooseung, Kim Sungyoung(KAIST)
- 소속
- KAIST
- 좌장
- 김성현(현대차), 박춘수(표준과학연), 박현우(동아대), 손기성(세안기술)
키워드
확장요약문
1. Introduction
Virtual reality (VR) offers an engaging alternative to repetitive auditory training and has supported learning in many domains [1],[2], including spatial-hearing tasks [3], where performance improves gradually with repeated practice [4]. The effect of acoustic rendering on this learning is less understood. This work compares two head-tracked binaural renderings in a gamified localization task: Early Reflections Accounted (ERA), which adds early room reflections to the direct sound, and Direct Sound Only (DS). Ten adults (23–33 years), all second-language speakers, who can be disadvantaged even with VR-reproduced sound [5], were grouped by a speech-in-noise baseline and completed three sessions of 50 trials locating two concurrent sources with visual feedback. Contrary to expectation, the DS group showed faster, larger accuracy gains (binding success: Cliff’s δ = −1.00, p = .012), whereas response time improved equally in both groups. 2. Method
In this between-subjects design, ERA rendered early reflections up to third order with frequency-dependent absorption, without a late reverberant tail; DS rendered only the direct path. Learning-rate slopes from per-participant linear regression were compared between groups with Mann-Whitney U tests and Cliff’s δ. 3. Discussion and Conclusion
This result contrasts with reports that a simulated room response accelerates HRTF adaptation [6]. Adaptation to altered spatial cues develops gradually [4], so three sessions may have been insufficient; reflection coloration may also have hindered localization. Real-world transfer was not assessed; a follow-up study (n = 36) will test it in a real room matched to the VR scene with dummy-head BRIRs [7].
References
[1] D. Checa and A. Bustillo, Multimed. Tools Appl. 79, 5501 (2020).
[2] J. Radianti et al., Comput. Educ. 147, 103778 (2020).
[3] B. J. Parmar et al., Front. Neurosci. 18, 1491954 (2024).
[4] C. Valzolgher et al., Sci. Rep. 14, 2469 (2024).
[5] C. T. J. Hui et al., Appl. Acoust. 184, 108368 (2021).
[6] D. Poirier–Quinot and B. F. G. Katz, Acta Acust. 5, 25 (2021).
[7] W. Mickiewicz and K. Kosmenda, Int. J. Electron. Telecommun. 71, 1–6 (2025).