ReadingClouds / Subfilter

python code to compute sub-filter quantities from MONC output.
0 stars 1 forks source link

Correct Filters and true FFT convolution. #4

Closed ReadingClouds closed 4 years ago

ReadingClouds commented 4 years ago

The Gaussian and wave_cutoff filters were written to average the filter over gridboxes. This has been removed, though previous behaviour is available via keyword. Note, 1D versions have not yet been changed, as they are not used in application code. WIll get round to it sometime.

New options to use fft2 or rfft2 in convolve (with filter fft for rfft saved for efficiency) have been implemented. Selected via options = {'FFT_type': 'FFT' } or options = { 'FFT_type': 'RFFT' }

Previous behaviour (not recommended) selectable via options = { 'FFT_type': 'FFTconvolve' }

Note changes to argument lists to include options.

See 'subfilter_file.py' for example of use. Now files test_signal.py, test_signal_real.py and test_filters are for testing filter implementations - note last in particular.