Exploring the limits of my gear - Nikon strobes and the iTTL system

Couple of months ago during the portrait shooting session I have noticed quite annoying fact that I couldn't control the light strength of one of the speedlight flashes used for the lighting setup. In this particular case it was the rim light that was causing problems. I have set the strength of this flash to be +1.7 stops above the default strength determined by the iTTL system. Since this light was too strong, I have reduced it by 1 and then 2 stops, but without any result. The exposure corrections were neglected by the system. Due to the time pressure I didn't investigate it, just moved the light stand a bit further away and obtained the required exposure in this way. But the questions remained:

- why did the system bahave like that?
- how can such behaviour be circumvented?
- how do I know such limits in advance next time?

Due to the other activities I couldn't answer all those questions till this weekend.

I had suspected that the problem with this flash was related to the distance between the flash and the subject. I thought that when the distance is too big, iTTL fires as much light as it can and gives up the control possibilities. To prove it I have tested the system with a simple setup, sketched below:

When the subject is properly exposed by the flash it looks like shown below:

The test was straightforward. I selected a number of distances between the subject and the flash. For each distance I took the series of shots, with the relative exposure of the flash set to -3, -2, -1, 0, +1, +2, +3 stops. For each shot I observed if the flash exposure correction had effect. I did the observation based on the shot but also taking the histogram into account. To illustrate what I mean look at the two tables below, showing the situations where I had full control over the exposure and where I had the control over the exposure from -3 to +1 stop.

First the situation where I had full control over whole range of exposures (distance from the subject to flash was 0.9m):

Exposure	Histogram

It is clearly visible that the flash can produce enough light for the whole range of the exposure corrections.

Then the situation when the control is no more possible above +1 stop  (distance from the subject to flash was 2.9m):.

Exposure	Histogram

So indeed, my assumption was right: if the distance between the subject and the flash grows, the iTTL narrows the range of the relative exposures that can be used. For my measurements I have got the following results (for SB910, set at ISO 100, zoom 24mm):

Distance	0.9m	1.9m	2.9m	3.9m	4.9m	5.9m	6.9m	7.9m
Control	-3..+3	-3..+2	-3..+1	-3..0	-3..0	-3..-1	-3..-3	-3..-3

So the lesson learned is that one can manipulate the distance between flash and the subject as long as loosing of the relative exposure control is acceptable. Which answers my second question: to have the control back over whole range of exposures, one has to move flash closer to the subject.

How far? I found that iTTL can give quite accurate hint about the distance (provided you have SB-800, SB-900 or SB-910 flashes). Here is the procedure:

1. Connect the flash to your camera (either directly or by using the flash release cord)
2. Setup your flash to work in TTL or TTL BL mode, depending on your needs (or planned exposure)
3. Setup the zoom of the flash for the value you intend to use (either by the flash settings or by using the lens zoom, if applicable), required ISO to be used (this one needs to be set in the camera) and the aperture (again, either in the flash or in the camera).
4. Note that the distance scale of the flash shows the range of the distances where the flash will expose the subject properly.
5. Now change the relative exposure of the flash (example for SB-910):

6. Observe that the maximal distance changes with the relative exposure (here an example for the reading for relative exposure +1.3 stops:

The maximal distance together with the exposure correction can be interpreted as a maximal distance at which the exposure can be corrected from -3 up to the value set. In the presented picture, if the subject is at 2.4m from the flash, the exposure can be corrected between -3 and 1.3 stops. Setting it above 1.3 stop will not have effect on the exposure (the exposure will be the same as for +1.3 stop).

Note that it is not exact math, rather indication. I have noticed, for instance, that while iTTL indicates that I can control exposure up to -2 stops at the distance 7.7 meters, I was able in fact to do so at the distance less than 6.9 meters. But for the values -1 to +3 it matches my experimental results quite accurately.

So for the next time I have to remember:

- Check the distance between the subject ant the flash,

- For this distance validate the ranges of exposure correction that will make sense. 

- Don't waste time on trying to go above maximal exposure correction. It will not work.

D800E - autofocus issue: confirmed after the tests

As mentioned in previous post, I suspected that my D800E might have an autofocus issue when using autofocus points located at the left side of the viewfinder. I have decided then to perform more thorough test to see if the issue is real. I can say that there is a problem with my camera (left focus points don't work well, but on the right side the results are not so great as well). If one is interested in the technicalities of the test, read on.
The test I have conducted is based on the setup as described by Nasim Mansurov on his blog here and here.
It is pointless to repeat the test procedure in my post. So let me point out just a few aspects I have gone through when testing my camera:

Test approaching the autofocus point from both sides (low and high distances)

It means that for a given focus point I did two tests. I made the exposure initially unsharp by setting the lowest available distance for the given lens. So I turned the lens focus ring manually to the lowest possible distance and did the test. Then I turned the focus ring to infinity and redo the test again.

Why is it important? It turned that for both lenses I have tested, my D800E performs better when setting autofocus starting from low distances. When the exposure distance of a lens was set up to infinity, the autofocus performance was much worse (i.e. the autofocus didn't work). 

So by taking only one side one might not notice the problem.

Speed up the test by using tethered mode

I have connected my camera to the laptop and to 40'' TV set to see the results in a big format. Tethered mode for D800 is available in Adobe Lightroom 4.2. It is not officially released yet, but the release candidate version works pretty nice.  Having the results immediately on the big screen speeds up the procedure

So, what's next?

I have contacted Nikon Netherlands and they are aware of the issue (as a matter of fact they expect quite some more calls with respect to the issue). They claim to be able to fix it in one week. Let's see.