Audio example from fast ISM simulations, Eric A. Lehmann

Fast ISM simulation: audio data example

This page contains an example of the results obtained with IS-DRM, a method based on a diffuse reverberation model (DRM) for the simulation of room impulse responses (RIRs) according to the image-source (IS) method. See [1] for implementation details. The following figure provides an overview of the simulation setup.

The environment is a 4 x 5 x 2.7 m³ room with reverberation time T₆₀ = 0.55 s. The sound source moves along a 2.64 m circular trajectory at a constant height of 1.55 m, and the acoustic receivers are located at a height of 1.35 m.

The following data sample was used as source signal: anechoic_source_signal.wav [280kB]

The following audio samples are the result from ISM simulations in the considered environment:

standard ISM simulation [2], coarsely-sampled trajectory (212 RIRs, 2.51 cm increment): standardISM_RT550ms_CoarseTrajectory.wav [580 kB]
IS-DRM simulation [1], coarsely-sampled trajectory (212 RIRs, 2.51 cm increment): fastISM_RT550ms_CoarseTrajectory.wav [580 kB]
IS-DRM simulation [1], finely-sampled trajectory (1056 RIRs, 0.5 cm increment): fastISM_RT550ms_FineTrajectory.wav [580 kB]

Note how clicking noise is distinctly audible in the audio data generated for the coarsely-sampled trajectory; this directly results from the RIRs differing significantly from one trajectory point to the next. Apart from this specific artifact however, both ISM simulation methods (standard ISM and IS-DRM) lead to audio samples that are virtually identical from a subjective point of view. A fine discretisation of the source trajectory further allows smoother transitions between RIRs, thereby avoiding the above mentioned clicking noise artifacts.

The simulation time required by the standard ISM for the coarsely-sampled trajectory (212 RIRs) was approximately nine hours on a standard desktop computer (2.4GHz), whereas the IS-DRM method generated the audio data for the finely-sampled trajectory (1056 RIRs) in about 21 minutes.

References

[1]	E. Lehmann and A. Johansson, Diffuse Reverberation Model for Efficient Image-Source Simulation of Room Impulse Responses, to appear in IEEE Transactions on Audio, Speech and Language Processing, 2010 (accepted for publication).
[2]	E. Lehmann and A. Johansson, Prediction of energy decay in room impulse responses simulated with an image-source model, Journal of the Acoustical Society of America, vol. 124(1), pp. 269-277, July 2008.

Home