Closed fietew closed 7 years ago
@hagenw: Could you please try again? I'm getting a weird error in octave:
error: tf: /usr/lib/x86_64-linux-gnu/octave/packages/control-3.0.0/x86_64-pc-linux-gnu-api-v50+/__control_helper_functions__.oct: failed to load: /usx86_64-linux-gnu/octave/packages/control-3.0.0/x86_64-pc-linux-gnu-api-v50+/__control_helper_functions__.oct: undefined symbol: _ZNK5ArrayINSt7__cxx1ic_stringIcSt11char_traitsIcESaIcEEEE17resize_fill_valueEv
I think, this has something to do with my octave installation.
For me, everything is now working under Octave.
And I can also see the differences in the plots of test_imp_25d(true)
if I enable or disable the gain corrections.
I'm going to merge this now.
According to the original publication ( see https://doi.org/10.1109/ASPAA.2011.6082325 ), Equation (10) and (11), some weights are missing in our NFCHOA time-domain implementation for the plane waves and point source. This pull request fixes this.
Execute
test_imp_25d(true)
to compare the spectra of 2.5D SFS techniques near the reference point with the ground truth (plane wave or point source).There was also a bug in
SFS_time_domain/driving_functions_imp/driving_function_imp_wfs_fs.m
related to a wrong usage of the/
-operator.