Room impulse response synthesis in simulations with simple models

Cola MD, Martignon P. Room impulse response synthesis in simulations with simple models. In: Express Paper. Audio Engineering Society; 2025. Paper 399. Available from: https://aes.org/publications/elibrary-page/?id=23073

@inproceedings{Cola2025_23073,
  author = {Cola, Mario Di and Martignon, Paolo},
  title = {{Room impulse response synthesis in simulations with simple models}},
  note = {Paper 399},
  year = {2025},
  month = oct,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=23073}
}

TY  - CPAPER
TI  - Room impulse response synthesis in simulations with simple models
AU  - Cola, Mario Di
AU  - Martignon, Paolo
M1  - Paper 399
PY  - 2025
DA  - 2025/10/14
UR  - https://aes.org/publications/elibrary-page/?id=23073
PB  - Audio Engineering Society
LA  - en
AB  - In recent decades, the acoustic simulation of indoor environments has been a significant challenge in computational acoustics. Advances in computational power have enabled the implementation of highly accurate acoustic modelsaccounting for phenomena such as reflections, scattering and diffractionand, as a result, the realistic simulation of room impulse responses (RIRs). Since teh simulated RIRs include multiple directional components, they allow for the auralization of the simulated acoustics, enabling users to experience an immersive representation of the acoustic field. This is achieved by convolving the auralized RIR with anechoic signals and reproducing the result via headphones or multichannel immersive audio systems. Accurate RIR simulation typically requires the creation of detailed geometric models, a process that is time-consuming and demands precise geometrical data. However, in professional audio applicationssuch as for the virtual installation of sound systemsusers often work with simplified geometric models of large spaces. Some software tools designed for these applications support the creation of basic closed-space models and allow the assignment of frequency-dependent absorption coefficients to surfaces for reverberation time estimation. These surfaces are generally large and lack detailed geometric features or scattering coefficients. Attempting to produce convincing auralizations directly from such low-resolution models using simple ray-tracing algorithms can lead to unsatisfactory results. This paper presents a technique to enrich the RIR by compensating for the missing spatial and acoustic details in low-resolution models, thereby enabling more plausible auralizations even in the absence of high-fidelity input data.
ER  -