原创 Nikon CLS Advanced Wireless Lighting, Part IV

After going through the boring parts of the series (Part I, Part II, Part III), in which I described the “secrets” of Nikon CLS Advanced Wireless Lighting, I will discuss some real world examples this time using Nikon CLS Advanced Wireless Lighting. The examples are intended to show some of the common issues people may have seen.

To run the test, I mounted a SB-800 Speedlight on one end of the Nikon SC-28 TTL Remote Cord, then mounted them on a light stand using the 1/4″ thread of the SC-28. If I connect the flash to the camera’s accessory shoe, I can use the available flash modes offered by SB-800, such as TTL/BL, TTL, etc. I can also disconnect the cable from Camera and set the flash as a remote unit that I can control using the camera’s built-in flash as master. This setup allows me to run the test without moving either the flash or the camera.

The first image was taken using Nikon D200 in Commander mode and a SB-800 as a remote in TTL mode. The camera was at f/2.8 and with 1/30s shutter speed. Flash sync mode is set to front curtain on the camera.

The second image was also taken using the same camera and flash setting. Did you notice the big difference?

The second image shows a slightly brighter overall scene and much brighter shadow on the right. The extra shadows of the book shelf in the image indicate that there was a second light source contributing to exposure. If you have followed through the previous articles of the series, you won’t be surprised to know that the extra light is from the camera’s built-in flash that I used as the commander unit. The remote flash trigger pulse is supposed to happen at the same time when the shutter is open and the remote flash fires.

In many cases this extra light can be a problem. The trigger flash is typically quite weak but can still be very noticeable when the flash-subject distance is short or the ISO sensitivity is high or there is highly reflective surface in the field of view. In rear curtain sync mode with shutter speed slower than 1/2 second, the remote flash triggering is done with a series of pulse (we explained this in part III).  It will appear much brighter and get captured by the camera. To prevent the problem, Nikon has provide the Nikon SG-31R IR Panel for Built-in Flash you can mount on the hot shoe. It has a plastic panel that is opaque to visible light but allows infrared light to pass. Remote flash units can function perfectly fine with infrared signal.

The following image was taken in with the SB-800 off camera in TTL mode using a SC-28 connection. The image appear to be significantly brighter than the first one. To match the brightness I had to add a 2/3 stop exposure compensation to the first image. The question is, why would wireless TTL and wired TTL lead to different exposure?

Actually I am not the only one seeing this problem. The same problem (underexposure in wireless TTL mode) has been observed by many. Here are just a few examples I found at photo.net:

• D200 and sb-600 images underexposed in wireless mode
• Remote SB800 + D200 = Underexposure
• Does Nikon CLS suck or is it just me? Consistant underexposure

And there are a few more I found on Nikonians.org:

• CLS underexposure
• D200 and sb-600 as slave problem

Answers that people could get from Nikon is like the following:

“…… There will be a slight exposure difference when the flashes are used off camera wireless because the camera uses a slightly different calculation with off-camera flash than when on-camera……”

If the difference is intentional, we can safely assume that the Nikon had good technical reasons for such a difference. We may never know the exact reasoning behind this. If you are really curious, here is what I think may be the reason. A flash attached to a the camera’s accessory shoe is often the only light source or the main light source when it is set to contribute to the exposure. The flash should point to the subject. So the camera needs to make sure the flash in TTL mode provide the exposure level expected to be normal by most users.  With remote flashes in TTL mode, the camera has no way of knowing where the remote flashes are pointed at or what the intention of the photographer is. It could be providing the main light but it could also be providing the background light, or fill light etc. Without knowing what the photographer wants it to be, it will be foolish for the remote flash to try to bring the exposure of the main subject to the standard level. If the remote flash is not pointed at the subject, it may mean a blown-out region somewhere. The only practical and safe decision it can make is to make sure no regions illuminated by the remote flash are overexposed. Due to the typical auxiliary nature of the remote TTL flashes (Check out the Nikon SB-800 flash shooting techniques guide by Nikon) it probably makes sense to intentionally underexpose with the remote TTL flashes.

The decisions for the remote flash output amounts are made entirely by the camera. Nikon may choose to change the behavior if they want to and there are good reasons to do so. Actually I haven’t seen much complaints like this lately. Perhaps newer cameras are different now.

Another argument for the wireless TTL underexposure goes like this: the camera determines the output for each remote group based on the reflected light of the pre-flash of that particular group independently of other groups (for example, the test here). To avoid overexposures in the overlapping flash illumination areas, the output for remote groups are intentionally underexposed.

This is obviously wrong and it is easy to prove. I took the following shots with the camera in Aperture Priority exposure mode and Front Curtain flash sync mode. I used two remote flashes that pointed to the same area. In the first image, I triggered both remote flashes (one in Group A and one in Group B) in TTL mode. In the second image, Group A flash was left on in TTL mode. Group B flash was turned off.  In the third image, only Group B flash was left on in TTL mode. Group A flash was turned off. There is some degree of variation in exposure but the overall exposures are very similar. If the camera indeed determined the flash output amount for a remote group based on its pre-flash only, I’d expect the image taken with both groups in TTL mode (the first image) would be much brighter (+1 stop). It was not the case.

The following image is the 2nd image above with +1EV exposure compensation done in Capture NX2. It is much brighter than the 1st image (Both Group A and B flash in TTL).

Recap

• The trigger flash from commander unit does contribute to exposure, which is a problem in many cases. You can avoid the problem by using Nikon SG-3IR IR panel or use a dedicated wireless commander SU-800.
• Wireless TTL flash tends to underexpose compared to wired TTL flash. The design may be logical.
• Camera analyzes the reflected light of pre-flash from all remote groups then determine the flash output amount for each remote group.

Update [November 15, 2009]

In a Nikonians.org discussion thread, Russ McDonald raised some concerns with the specular reflections in my tests photos published in this post. I also realized that the close up of books may not be a good example. So I redid the tests. In the new tests, I put a toy bear about 2 feet in front of a black background. I then placed two flashes on left and right of the camera, which is about 5 feet away from the bear. Both flashes heads were aimed at the bear. The camera was set to front curtain sync mode, f/5.6, 1/60s, and ISO100.

The first image below was shot with both flashes on and in TTL mode.

The second image below was shot with the flash on the left (group A) in TTL mode. The flash on the right (group B) was turned off.

The third image below was shot with the flash on the right (group B) in TTL mode. The flash on the left (group A) was turned off.

The results are very similar to the book shelf example I tried earlier. This again clearly demonstrated that the camera makes decision on the final flash exposure after considering all remote TTL groups. If this was not the case, we’d expect to see the first image have more exposure than it had now.

I included the following images for comparison. The first one below was the first bear photo above with an 1EV exposure compensation done in Capture NX2. The second one below was a photo taken with two remote TTL flashes and +1EV flash exposure compensation dialed in on the camera.

Finally, the wireless TTL and wired TTL (not TTL-BL) comparison. I used the same method described at the beginning of the post to avoid moving the camera and flash. By playing with the exposure compensation slider in the Develop Section of Capture NX2 to match the histograms, I found that the photo (top) taken with wireless TTL was 2/3 stops underexposed compared to the photo (bottom) taken with wired TTL.

Recap #2

The previous conclusions about how camera manages the exposure of multiple remote TTL groups are still valid. The final flash exposure determined by the camera is not the simple sum of the flash exposure of individual groups when they are present alone. The camera instructs the remote TTL groups to fire the pre-flash one by one then analyze the reflected light intensity data from all remote groups to come up flash output amounts for each remote TTL group so that the final result has consistent exposure no matter how many (up to 3 ) remote groups are present.

And again, wireless TTL underexposes compared to wired TTL.