Implement filter tests for complex signals #460
Conversation
f-brinkmann
changed the title
|
We might get away with fewer tests if we test only if the three filter classes FilterFIR, FilterIIR, and FilterSOS can handle complex input. If this is the case, all pyfar filter functions will be good, as far as I understand. |
tluebeck
changed the base branch from
complex_time_signals
to
develop_complex_signals
| def test_filter_sos_process_complex(impulse_complex): | ||
| sos = np.array([[1, 1/2, 0, 1, 0, 0]]) | ||
| filt = fo.FilterSOS(sos, impulse_complex.sampling_rate) | ||
| coeff = np.array([[1, 1/2, 0], [1, 0, 0]]) |
There was a problem hiding this comment.
Why are you initializing twice? According to the docs the shape of the coefficients should be (1, 1, 6) did that not work?
There was a problem hiding this comment.
It was a copy and paste from the test for the non complex data, to be honest :) ... I removed. Also in the non complex test. Should I also test it with coefficients (1, 1, 6)?
There was a problem hiding this comment.
I tried it and filt.coefficients has shape (1,1,6) in both cases. but maybe inputting the coefficients according to the docs is less confusing, i.e., using sos = np.array([[[1, 1/2, 0, 1, 0, 0]]])
Co-authored-by: Anne Heimes <64446926+ahms5@users.noreply.github.com>
unit testing for filtering complex signals