Below is what I have done. I offer no indication it is "right":
>Here are some questions that I have regarding the testing >protocol. I am using FoCal Pro version 126.96.36.199. >1. Should Target Optimization be turned on? I think so.
I did so, except for long distance or very wide angle when it couldn't "see" the target.
>2. Does Color Temp really matter? I am testing with two >Elinchrom studio strobes using the modeling lights at max >setting. I am now setting the color temperature in FoCal at >3000K vs 5200 originally.
There was a posting around here somewhere indicating it made a big difference. I've done some of mine with daylight, some with incandescent, I do not know. I was surprised by the posting saying it did.
>3. Should I be using mirror lockup? I haven't been.
No. I don't even think it woudl work.
>4. I left the battery grip on my D800, is that OK?
I can't imagine why it would matter.
>5. I have the illumination consistently between 8-9EV using >my studio lights. I check it with a light meter. Is the >variability between 8 and 9 significant?
I don't know. I did some testing outside, and if a cloud rolled by it would sometimes abort the test, but when a light cloud rolled by that was not enough variation to abort the test, the test did not work well. I think in any given series (where it is comparing image A to B to tell sharpness) you need really consistent lighting. Blue sky and in the shade seemed OK for me, and often better than incandescent as it was brighter (I do not have studio lights, I have been using two 100w bulbs with reflectors and it is barely bright enough).
>6. I printed the targets on 8X10 Luster paper. I know they >recommended matte. I don't see any reflections or luster with >my setup. What are your experiences using different paper?
I have had trouble on regular laser printer paper with reflections from the lights, and have to fiddle around to make sure that the area around (not just on) the target does not have any glare. And I think it is very important that it not change at all during any one run (or any runs you desire to compare to another), as the contrast difference makes a large difference in the focus quality number. In fact in general I think comparing two separate runs in terms of absolute numbers that were not at the same time/place without moving anything is almost pointless.
>7. I have been using the default AF Consistency Constraint of >6%. Is that what you are using or should it be tighter?
DIdn't do anything there.
>8. I have done one set of tests allowing non-centered >alignment to test the left, center and right focal points. Has >anyone else done this? My first test results on my D800 with >24-70mm showed -11, 1, -14. Yikes! I want to finalize the >protocol before retesting.
That's not terribly diffreent from some of mine, I don't recall off hand whether it was that or the 14-24. The D800 even post repair still has quite a difference, and yours is similar to mine -- right and left match each other, and are off from center.
>9. What distances are you testing various lenses at? FoCal >suggests the 25-50X the focal length.
I have taken to doing each lens at portrait distance, at kind of a normal "interior" distance where I might be in a building, and at about 20-30m distance (whereever I can find a usable target outside to simulate sports-distances). And see how they vary. Some (like the 200-400) I also did at very close up distance.
That's a lot of testing for a zoom, as it's at least 3 runs at each distance times 2 for min/max zoom.
>10. For 200-400mm lenses what target size are you using?
I have been printing the target zoomed slightly and using it for all the tests. For the 200-400 up very close I used the macro target. For longer telephotos I've been trying several different non-target targets - insulators on a light pole against blue sky, mortor joints in a building across the way, textured tree trunk bodies. None of these are great (and you have to make sure they aren't MOVING), but of these the higher contrast work best, something that at 1:1 still has strong horizontal and vertical, high contrast transitions in the focus point. I now make sure I do at least two different ad hoc targets to make sure they give similar results (sometimes they do not, I discard that as a target and keep looking for two more).
It's not that I can't print a multi-page target blown up, but I can't find places out of doors to MOUNT such a target 30m away from where I have the computer/camera.
>11. My optimization tests normally peak around 2000. What is >focal measuring? What is this number?
I think it's an arbitrary edge sharpness number related to the entire scenario and only meaningful in comparison to other images taken at the same time. Others have tried to compare but I think it's questionable at best.
Your mileage may vary, I think Focal added some science to this but it's still heavily in the art and trial and error arena.