[joachim]

Well, these auto corrections are a borderline thing from a testing perspective. The MTFs are better without. 

The Oly 17mm f/1.8 has a raw distortion of 4.5%. As a consequence there is simply quite a bit of stretching and interpolation going on in the border region. 

Since this is enforced in MFT land, it is hard to get around this.

 

Interestingly CaptureOne 7 provides the uncorrected image from RAW now. 

It is probably hard to tune this but I could try to provided the RAW MTFs besides the 'corrected' MTFs.

Not sure in how far this is helpful though.

 

The results aren't bad for the 17/1.8 but they stay way short of the 75/1.8, of course (which doesn't need any auto-correction).

 

I think at some point you have to penalise them.  Either you report top MTF and give them poor marks in the distortion department or you fry them for poor boarder performance but give them a good rating for low distortion.  I think, since most people use the distortion correction and suffer the poor boarder performance the later is more relevant to how people use these lenses.  Your comments seem to imply that this (measuring MTF and distortion with distortion correction "on") is what you are doing.

 

On your comparison of the 17 to the 75, I would never expect a medium WA for digital to come close to a medium length telephoto.  Just to much physics against you.  Compared to the 75 the 17 is a cheapo.

 

I am looking forward to your review of the 17.