`offset` argument to Detector class doesn't actually work

[mperrin]

Looks like I left a few minor bits out of this implementation. Will fix.

Somewhat more complicatedly, _propagateMFTinverse doesn't support offset image planes.

I actually no longer remember the details of what motivated this feature - but I just came across a use case in making one of the MFT to FFT comparison tests. (Detector always uses ADJUSTABLE style MFTs, so there's no way to get exact equivalency with FFT outputs -- unless one uses this offset argument!)

[mperrin]

not likely to actually have time to look at this before 0.4, alas.

[mperrin]

nobody's asked about this, not a high priority.