Parity issue with MFT

[mperrin]

Issue by mperrin
Thursday Jan 22, 2015 at 23:31 GMT
Originally opened as mperrin/poppy#38

---

Properly speaking this was a poppy issue, not webbpsf.

See mperrin/webbpsf#35

[mperrin]

Comment by mperrin
Thursday Jan 22, 2015 at 23:34 GMT

---

• Commit 971436b adds test functions that MFT and FFT both preserve parity after forward-then-inverse transforms, which was the original problem that started this.
• Consider expanding above test function to also compare multiple image planes as well as pupil planes?
• Should we also test the equivalence of inverse MFT with inverse FFT?

Add test functions for MFT / FFT equivalence:

• one just testing the base matrixDFT interface based on what we did today,
• and one testing higher level usage via OpticalSystem and wavefront propagation.

[mperrin]

Comment by mperrin
Saturday Jan 24, 2015 at 01:05 GMT

---

This is pretty much good to go. There are a couple more tests that could be added for complete thoroughness, but we're reaching the point of diminishing returns.

[mperrin]

Comment by mperrin
Thursday Feb 12, 2015 at 22:45 GMT

---

Additional test coverage is now issue #71. This bug is fixed so this issue can be closed.