Hybrid Acoustic Modelling of Historic Spaces Using Blender

Research output: Contribution to conferencePaper

Full text download(s)

Links

Author(s)

Department/unit(s)

Conference

ConferenceForum Acusticum
CountryPoland
CityKrakow
Conference date(s)7/09/1412/09/14

Publication details

DatePublished - 7 Sep 2014
Number of pages8
Original languageEnglish

Abstract

Historic spaces provide a challenge in terms of achieving accurate acoustic modelling and auralisation due to the large volumes typically involved, implying significant computational overhead, uncertainty in terms of the construction materials’ properties, and translating this into appropriate physically based boundary conditions. Hybrid acoustic modeling approaches seek to solve the computational problem through complementary assimilation of various modeling paradigms. SonicRender is such a hybrid acoustic modelling tool, based around the Blender open source 3D graphics development platform. Finite Difference Time Domain and ray tracing methods are used, with the FDTD method constrained to low-frequencies enabling the simulation of wave characteristics at manageable compu- tational cost. Efficient rendering of spectrally complete RIRs are produced by combining results from geometric and band-limited numerical simulations. In this study, SonicRender is applied to recreate the acoustic of a site of historical and architectural importance: the National Centre for Early Music, York, UK. A series of acoustic measurements have been made through utilisation of the Exponential Swept Sine method for one source location and multiple receiver locations. As such, objective acoustic data is used to test the validity of SonicRender as an acoustic simulation tool through comparison of recorded and simulated room acoustic metrics. Conclusions are drawn on the future potential for using such hybrid acoustic modeling methods in similar challenging application areas.

    Research areas

  • Auralisation, FDTD METHOD, acoustic modelling, room acoustics

Discover related content

Find related publications, people, projects, datasets and more using interactive charts.

View graph of relations