How do you handle the inconsistent D800/D800E PDAF with different color temperature?

In this thread I would not like to discuss all the AF trouble I already had, but the problem which bothers the most. I never could imagine, that it would be so difficult for a brand like Nikon, to build a reliable PDAF system!

With the help of the home-made test equipment I have finetuned the PDAF for daylight.

home-made test equipment  for AF finetune

In daylight conditions the PDAF works quite fine now. It’s still difficult for me, to use single point AF-C with moving objects and a 2.8 or less aperture. But I can live with 20-30% waste.

The real problem starts for me, when I take indoor shoots with different lighting conditions! What happens, I’m trying to show with the following pictures. I already tried different D800/D800E bodies, a D4 and a 5DM3.

D800, 85/1.8G, 6000K

D800, 85/1.8G, 2800K

If you look to the D800 picture where the focus is right, the color temperature is about 6000K. The camera has an frontfocus of about 8cm(approx. 3 inch.), if the color temperature is about 2800K.

Because I wanted to know, whether a D4 with the same PDAF module has the same behavior, I borrowed a D4. I was surprised that the camera did not showed the mistake in the lighting condition with 2800K color temperature! (I started with that test.) But finally, I did the test with the lighting with about 6000K. I got a backfocus! My conclusion: I didn’t finetuned the D4 AF system in daylight conditions and because it’s quite OK in the 2800K light, it must have backfocus in 6000K light. It’s the same system like in the D800!

D4, 85/1.8G, 2800K

D4, 85/1.8G, 6000K

Because I didn’t know, whether all camera manufactures have this problem I borrowed in addition a Canon 5DM3. Just have a look to the pictures!

5DM3,100/2.0, 2800K

5DM3, 100/2.0, 6000K

In my opinion, the results for the Canon 5DM3 are acceptable. It seem’s to me, that Canon fixed that problem in 2008!

http://downloads.canon.com/CDLC/EOS50D_5D-MarkII-WP2.pdf

On page 32 I red:

But under fluorescent lighting and some other artificial lighting sources, a rapid, imperceptible
flickering of the light occurs along with sudden color temperature shifts. These tend to
throw off the density measurements and therefore the accuracy of the AF calculations. In both
the EOS 50D and EOS 5D Mark II, the type of lighting is taken into effect, and readings
are averaged and processed with the help of the super-fast DIGIC 4 Image Processor. ...

Back to Nikon!

Why can’t the PDAF in Nikon cameras work that good as in Canon cameras? How do you handle this focus-shifts? If I look further I know, using the SB910 with the AF assist light the focus shift’s invers. That mean, I do not get front- but backfocus. As a photo- enthusiast I’m a little bit lost in space. What is to do? I already red in other treads from Marianne Oelund, she uses the AF finetune like the exposure correction. I can’t really imagine that this will work for me on a wedding, where I change rooms with different lighting conditions all the time. Just to use apertures higher then 4 isn’t a solution neither!

It seems to me right now, that’s worthless to buy expensive lenses with apertures less than 2.8 for the D800/D800E … because Nikon can’t build an PDAF, which handles the focus. For somebody who just can live with the live view, it’s maybe OK. But my experience with the live view and the 50/1.4G isn’t good too, especially if you move the focus point out of the lens center. The micro-contrast for that special lens I tested is really quite bad, especially in the corners.