JML 21, Canon 35, Oly 38 with Sea Urchin Spine at 5x

[PaulFurman]

Some tests of 3 lenses with pretty clear results that surprised me. I think my step was too much but still plenty to see here. I thought I had all the lenses stopped down to f/3.5 but the 38 needed to be darkened one stop. In an ideal world I'd have tried different apertures and step distances but these things are endless.

Sea urchin spine, near the base where it attaches (seems like a perfect benchmark for testing lenses.
5x magnification full frame 12MP stacks of 42 at 25 micron increments, 6 second exposures. All sharpened rather heavily at minimum 0.5 pixel radius.

Corner & center crops full pixels:


Full frames link to original 12MP versions.

21mm f/3.5 JML


35mm f/2.8 Canon MP-B


38mm f/2.8 Olympus


Conclusions:

The Canon 35 isn't as good as I thought. Perhaps because the Oly is faster than I thought? That would explain why I didn't like the Oly; things are so darn blurry and smaller steps are needed.

All of them hold up quite well in the corners on full frame.

The JML is very nearly as good as the Oly, hardly worth mentioning the difference. Interestingly, jpeg file size can be a good way of objectively measuring and here the numbers are:
straight out of zerene:
6,265 - 6,307 - 6,452
web versions from lightroom, sharpened & compression 80:
3,195 - 2,740 - 3,272

[Pau]

Paul, very good comparation.
I need to test the JML against my old Oly 20mm f3.5 (single coated) but I think I already know the answer: the JML may be the winner.

[PaulFurman]

Pau, my previous (first) test at 3x put the 20mm f/2 and 38mm Oly against the JML and the 20mm Oly did poorly, perhaps because it's out of it's range of 5.3x or a number of other issues. The 3.5 version is rated for 4.3x so that's certainly not a foregone conclusion.