Saving and loading (fixpoint) filters
chipmuenk opened this issue · 2 comments
Several problems exist when saving / loading filters:
-
[ ]: When saving filters, it seems as if coefficients are only saved with an accuracy of 8 digits. This would be due to
pyfda_io_lib.save_data_np()
and its default optionfmt=%f
. -
[ ]: When reading back a fixpoint filter, selecting the b,a tab can crash the software with
File "d:\daten\design\python\git\pyfda\pyfda\input_widgets\input_coeffs.py", line 548, in _refresh_table_item if self.ba_q[col + 2][row] > 0: TypeError: 'int' object is not subscriptable
This can be avoided by running a fixpoint simulation, but it seems the ba_q[][] array is not loaded / updated / created. Coefficient formats between the fixpoint tab and b,a can also be out of sync.
-
[ ]: The key
fb.fil[0]['fxqc']['QCB']['fx_base']
is not present in older filter designs, causing KeyErrors and an empty coefficient table. The import routine is not robust enough, there should be a default fallback.
Version 0.8.3. The program doesn't crash when I click "Save" with fixed point coefficients. But the coefficients are not saved. Nothing happens