TUB Acoular

[esarradj]

Adds a notebook for Acoular and the data

Work in progress as benchmark_results notebook needs to be updated.

[esarradj]

Comparison is now done. The following issues remain:

• factor in the CSM
• PSF should not be given in decibels (as it is not an SPL)
• map resolution: it would make sense to have the maps with an homogeneous resolution (same in x and y); at least in Acoular it would be presently a bit more complicated to do it otherwise (and why should we?), if aeroacoustics.jl has a problem with this we could settle for a square map (say 3x3 meters)

[1oly]

Overall, I think it looks great. Regarding the csm, I suspect is is a scaling option: https://github.com/1oly/AeroAcoustics.jl/blob/309f34e766888dec602879cd8858df9b958d0643/src/csm.jl#L62-L66

I set it to 'spectrum', but it looks like acoular use the 'density' : https://github.com/acoular/acoular/blob/d41bc2c41d743c17a48faf50a40bd9835046befd/acoular/spectra.py#L403

Regarding PSF, you are right, I will export the correct numbers.

Map resolution: I think we should use nx = 81, and ny = 41, looks better in your version.

[esarradj]

No, I am afraid what we use in Acoular is also 'spectrum' scaling. Otherwise the factor would be much larger I guess.

The factor in https://github.com/acoular/acoular/blob/d41bc2c41d743c17a48faf50a40bd9835046befd/acoular/spectra.py#L403 is just the window, i.e. self.block_size / weight will give 1.0 in case of a rectangular window.

The factor is 1.5003663 for all csm elements and also regardless of the frequency.

[esarradj]

Can be merged imho