Prism-RA – 3D Acoustic Model
PRISM-RA is a hybrid numerical and geometrical DSP, 3D acoustic model used to visualize and predict parameters used in room acoustics. There are two common methods used to predict parameters in room acoustics; empirical and geometrical.
Empirical methods (such as the Sabine formula) are employed in most spaces by acoustic consultants, architects and other professions to provide quick estimates for the reverberation time of a space. These methods are totally fine for simple spaces, which are diffuse and have a fully even distribution of absorption. The problem is that most spaces are not totally absorbent as they may have plasterboard walls, a thick carpet, wall panels etc.
Geometrical methods such as ray-tracing and image source models can account for some of the uneven distribution of absorption and therefore sound in a space. They do this by creating sound ‘rays’ or ‘particles’, which are subject to similar laws to that of light. The limitation with this method is that, it is only useful when the wavelength is much smaller than the surface area it comes into contact with. This requires an in depth knowledge of physical acoustics and waves, and cannot be predicted in the software, often meaning that the wrong results are calculated, which may lead to re-testing and mitigation measures being required afterwards! You will be glad to hear that there is a third calculation method…
PRISM and the numerical method: For years numerical calculations methods including FDTD, FEM, BEM etc have been used in physics to solve problems involving thermodynamics, electromagnetics, fluid dynamics and acoustics. These methods can solve partial differential equations directly, meaning that as long as the input data is correct and no errors are present, they can prove to be extremely accurate.
PRISM uses the 3D FDTD method for its calculations, meaning that it is one of the most accurate pieces of room acoustic software available. The software is unique when compared to existing room acoustic software, as it is coded from the ground up to process in parallel, working with the hundreds of GPU cores available to most graphics cards. This means that fully numerical modelling can be used in our calculations in a reasonable amount of time, without the need for a super-computer! For laptop computer users or those with an incompatible GPU, a hybrid model allows for the higher frequencies to be run as a geometrical model, and the lower frequencies to be run as a numerical model. An algorithm is used to opimise accuracy and calculation time depending on the hardware available, and the size of the model.
The software is not yet available for use as it is being re-written from scratch to optimise the speed of calculations and reduce the existing errors. However a first beta (trial) version of the software should be available later in 2017, with more details to come soon.
For more information regarding the latest developments in 3D audio and virtual reality, see our post here.
2017 Prism Acoustics Ltd Registered in England No.10665452.