add modal weighting to time-domain nfchoa

[fietew]

in order to study the effect of different modal windows on the reproduced sound field

[hagenw]

I have corrected a small error and changed a little bit the behavior of the default parameters to be in line with the rest of the Toolbox, hopes this is ok with you.

At the moment the modal_weighting() function has only Input parameters and Output parameters in the comments at the beginning. Have you a good idea what we could add for a more detailed explanation, that maybe could roughly explain what it is doing to someone not familiar with the concept of modal weighting?

For adding a reference I guess we have to wait until you publish some results?

[fietew]

Maybe we should use the varargin functionality of MATLAB instead of assigning ndft to conf. I think it's quite hard to understand, what happens, if only two arguments are given.

I could add some reference for the kaiser window function.

[hagenw]

I have updated the parameter handling a little bit, but did not change the behavior to varargin as the current behavior is the standard procedure done in SFS, see this example.

There is a TODO left, could you fix this and add the reference you mentioned, this would be nice.

[hagenw]

I added a small description, maybe you can check if it is correct.

Then, I was wondering about the usage of the conf.nfchoa.parameter entry. If I see it correctly, a setting of conf.nfchoa.wtype = 'kaiser' and conf.nfchoa.parameter = 0.0 will be the same as using conf.nfchoa.wtpye = 'rect'. Are you planing to integrate more windows? Otherwise I would propose to use only the Kaiser-Bessel window and provide just one conf parameter, something like conf.nfchoa.modal_window which can go from 0 to 1.

[fietew]

Yes, I would like to add some more windows in the future. conf.nfchoa.parameter is not that intuitive, but many window types have one parameter, but the meaning of the parameter is quite different. Furthermore, the parameter of the Kaiser window can go from 0.0 to infinity. I will add that to the comments of the SFS_config

[hagenw]

OK, then I agree to have the setting in conf. I'm still not completely happy with the naming, what do you think of the following proposal? conf.nfchoa.wtype => conf.nfchoa.modal_window conf.nfchoa.wparameter => conf.modal_window_strengths or conf.modal_window_parameter

[fietew]

I think the window parameter should still have conf.nfchoa. in front. conf.modal_window_strengths is misleading. If I had to chose, I would use conf.nfchoa.modal_window and conf.nfchoa.modal_window_parameter

[hagenw]

Sorry, my fault, I meant conf.nfchoa.modal_window_parameter. I have changed the code accordingly, could you test if it is still working.

Currently the modal_weighting is only applied in the temporal NFC-HOA driving function. Is it also possible to apply it to the mono-frequent driving function or do we have to change the NFC-HOA structure (as you proposed) before?

[fietew]

Yes, I had a look at the current monochromatic driving functions and it would be quite complicated to add the weightings there. With a new structure, i.e.

1. compute the circular/spherical basic expansion of the sound field,
2. compute the circular/spherical harmonics coefficients of the driving functions,
3. convert it to the loudspeakers driving function,

this would be easier. The modal weights could be applied directly in the second step.

[fietew]

I tested it with a few parameter combination and everything looked fine for me.