Source directivity

[Picaut]

It would be great to consider specific source directivity (in addition of current reflection law) in I-Simpa.

[nicolas-f]

Directivity - Energetic

Apply the linear interpolation of two attenuation value for each angle.

Convert attenuation to linear scale using lAtt = 10**(spl/10)

particle.energy *= lAtt

[wbinek]

There is a clf format in which directivity for mamy speaker sets can be found. http://www.clfgroup.org

I think it would be good to be able to import directivity from these files.

[nicolas-f]

Thanks for sharing. We already discuss with the CLF Group, the specification of the binary format is not intended for open source software. Reading CLF files must be done using provided dll file from CLF group.

As all components of I-SIMPA have to be open source (multiplatform & license issue) we can not use this dll. We will begin with tabular text format directivity as specified by Ease (export format).

[wbinek]

Thanks for the response. I understand. Using text files as in Ease should be equally good solution.

[austil]

Work in progress :smiley:

[wbinek]

The image looks as if the source directivity is obtained dy modifying the energy of the particle. Wouldn't it be better to consider directivity as probability function and send more particles in higher energy regions? This will increase the simulation accuracy in important regions. I believe such approach is common in image rendering problems.

[Picaut]

Both solutions will be developed, depending on the method used for the simulation "Random" or "Energetic". At a first step we implement the energetic one, and after the random one.

[wbinek]

I think that the ability to choose the source directivity method should be independent from method used in the rest of the calculation.

I believe that using random source approach with energetic calculation method may give better results than using energetic both for directivity and calculation.

The problem with random calculation method is that the quality of the simulation is high at the beginning where there are a lot of particles, but then it decreases as the particle number falls. To compensate it the initial number of particles should be much higher than in energetic method what makes it actually slower. Therefore in my opinion the method isn't really useful.

[Picaut]

There is now an open format for directivities. http://www.opendaff.org/ This could be included in I-Simpa instead of the current way.