Originally posted by dragra I haven't noticed any missed focus issues on my new K-r so far, but this WB-FF discovery seems interesting though. Both test subjects are pretty much close focus range, how about something in more distant range? I will try to do my own testing on this, just need to find a place lit by tungsten light.
I focused near minimum distance because the 18-55 lens gives the smallest depth of field at 18mm f3.5. At greater distances the focus chart becomes too tiny to read. If I increase the zoom level to compensate, the aperture also changes and gives a greater depth of field that makes it tougher to spot focus issues.
This is admittedly going to a level of pixel peeping I'd never use for real photos, but I want to exaggerate any focus limitations. If I get things working well for a worst-case scenario then all of my real photos will be great (or, to be more accurate, any bad photos will be my fault, not the camera's).
It's very interesting that the WB affects focus at all. I originally expected to get the same exact DNG file regardless of how I set my WB. I used cheaper cameras and never bothered with DNG; I clearly have much to learn about the technology behind my new KR.
I expected: Autofocus -> expose photo -> save DNG -> apply WB -> save JPG.
However, WB seems to come into play much sooner: Apply WB -> autofocus -> expose photo -> save DNG -> save JPG. It may be splitting hairs but the tungsten front focus may be "a type of light detection problem" rather than an "autofocus problem". Manually telling my camera "hey, it's tungsten" results in pretty good autofocus.
Later this week I will try more tests, playing with distance, focal length, light type, light intensity, WB, and different lenses. If I was really smart I would just replace my last old tungsten light fixture at home (it uses custom bulbs and can't fit CFLs) and mostly forget about this whole thing, but it's too late because this has become a fun science project.