The effect of vehicle sound power on auditory time-to-collision estimation

Huisman, Thirsa; DeLucia, Patricia R.; Oberfeld, Daniel · 2026 · Crossref

DOI: 10.1051/aacus/2026019

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates how the variation of vehicle sound power influences auditory time-to-collision (TTC) estimation, specifically addressing the "intensity-arrival effect," where louder vehicles are perceived as arriving earlier than quieter ones. Previous research demonstrated this effect when sound power varied from trial to trial, but this design may have artificially amplified the effect by directing participants' attention to loudness differences. The authors sought to determine if the intensity-arrival effect persists when sound power is varied blockwise (constant within a block, changing between blocks), a condition that better mimics real-world scenarios where a pedestrian encounters vehicles of consistent noise levels over short periods. This research is motivated by pedestrian safety concerns regarding electric vehicles (EVs), which are significantly quieter than internal combustion engine vehicles (ICEVs) at low speeds. The researchers employed high-fidelity acoustic simulations of an approaching gasoline-powered car using the TASCAR software and a 32-loudspeaker array with sound field synthesis. Participants viewed a visual simulation of the street environment via a head-mounted display but received no visual information about the vehicle itself, relying solely on auditory cues. The experiment involved 22 participants with normal hearing and vision. In a within-subjects design, participants completed two sessions. In the first session, vehicle sound power was varied blockwise, with participants experiencing either a lower sound power condition (0 dB audio gain) or a higher sound power condition (+10 dB audio gain) for an entire block. In the second session, the same sound power levels were presented in a randomly interleaved, trial-to-trial fashion. The vehicle approached at speeds of 10, 30, and 50 km/h, with TTCs at occlusion ranging from 1.25 to 5.0 seconds. Participants estimated the TTC by pressing a button when they believed the vehicle would reach their position. The results confirmed a significant intensity-arrival effect in both experimental conditions, meaning participants consistently estimated that louder vehicles would arrive earlier than quieter ones. However, the magnitude of this effect was substantially weaker when sound power varied blockwise compared to when it varied from trial to trial. Statistical analysis revealed a significant interaction between vehicle sound power and the type of power variation, indicating that the context of the sound presentation modulates the perceptual bias. The effect remained significant even when controlling for vehicle speed and actual TTC, though the specific interaction patterns suggested that the attentional focus on loudness differences in the trial-to-trial condition amplified the perceptual distortion. These findings imply that the intensity-arrival effect is robust but sensitive to the experimental context, particularly the availability of within-session comparisons for loudness. The weaker effect observed in the blockwise condition suggests that the perceptual bias may be less pronounced in real-world scenarios where pedestrians do not immediately compare vehicles of differing noise levels. This has important implications for pedestrian safety, particularly regarding the design of acoustic vehicle alerting systems (AVAS) for EVs. Ensuring that EVs produce sufficient sound power may be critical to mitigate the risk of pedestrians underestimating the arrival time of quieter vehicles, although the exact magnitude of this risk may depend on the specific auditory context and the pedestrian's ability to compare sound sources.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-25; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• time to collision looming