Skip to main navigation Skip to search Skip to main content

Assessing attenuation curve compliance in off-the-shelf sound spatialisation methods for sound localisation testing

    Research output: Contribution to conference › Poster › peer-review

    Abstract

    Powerful sound spatialisation engines offer immersive gaming experiences [1,2] but it is unclear if they are suited for clinical sound localisation testing [3]. We assessed if the attenuation of sound from off-the-shelf sound spatialisation engines was consistent. Unreal Engine (UE5) was used to virtually record 250Hz, 4KHz and 8KHz pure tones played through a virtual speaker. Recordings were taken at 0.001m, 0.05m, 1m, 2m, 3m, 4m, 5m, 10m and 15m. The audio was spatialized using: built-in (UE5 Panning), plugins (Steam Audio, Google Resonance and Project Acoustics) and middleware (Wwise and FMOD Studio). We applied distance-based attenuation only, in a 0.1cm – 25m range. Inverse square law attenuation was applied for all implementations, except for Google Resonance, where logarithmic attenuation was used. We captured audio using Audacity and loopback on a Behringer U-PHORIA UMC404HD audio interface at 96kHz. The average power of the signals was calculated at each distance and attenuation functions were plotted. The attenuation of the three pure tones was then compared for each implementation. UE5 Panning and Wwise showed no difference in attenuation. FMOD attenuated the 8kHz signal slightly faster compared to the others. The three plugin implementations attenuated the three signals at different rates, suggesting built-in frequency dependence that may affect sound localisation assessment in unexpected ways, especially when using complex signals (Speech). We expected no attenuation differences between the three signals. The presence of this difference is undesirable and is evidence of an uncontrollable simulation parameter. The effects of this phenomenon on human perception must be investigated.

    [1]K. Rogers et al., 2018, In proceedings of 2018 CHI Conference on Human Factors in Computing Systems, pp.1-13.
    [2] R.W. Lindeman et al., 2012, In proceedings of 2012 ISMAR - ARTS, MEDIA, AND HUMANITIES. pp. 93-94
    [3]Kitterick PT et al., 2011, Cochlear Implants Int., vol.12, no.3, pp. 164-169.
    Original languageEnglish
    Publication statusPublished - 4 Sept 2025
    EventBioMedEng25 - Technology and Innovation Centre, University of Strathclyde, Glasgow, United Kingdom
    Duration: 4 Sept 2025 → 5 Sept 2025
    https://biomedeng.org/biomedeng25/

    Conference

    ConferenceBioMedEng25
    Abbreviated titleBioMedEng25
    Country/TerritoryUnited Kingdom
    CityGlasgow
    Period4/09/25 → 5/09/25
    Other2025 conference of the Association of Biomedical Engineers, Medical Engineers and Bioengineers
    Internet address

    ASJC Scopus subject areas

    • Acoustics and Ultrasonics
    • Modelling and Simulation

    Fingerprint

    Dive into the research topics of 'Assessing attenuation curve compliance in off-the-shelf sound spatialisation methods for sound localisation testing'. Together they form a unique fingerprint.

    Cite this