Flash

First of all, I want to come clean about not being a great expert at flash, studio lighting, or any of the truly bewildering array of available lighting products, stands, modifiers, radio control systems, and et cetera. I have several years of experience with Nikon’s Creative Lighting System (CLS); and in every practical case that I’m going to work with here, that’s what I have used.

Nonetheless, the basic ideas, in the context of the Zone System, are applicable to other lighting systems. In this, I want to emphasize again that, as Ansel Adams said, the Zone System is a practical expression of sensitometry, and, as I would add, sensitometry is an expression of physics. I will claim to have some notion of physics (see one of my other blogs at Face-Paging for that). Anyway, with that disclaimer out of the way, let us proceed.

The Problem

As I have mentioned before in this set of pages on the Zone System, there are many occasions in which direct or reflected sources of light can overwhelm the intensity of light from diffuse surfaces. The classic case is a landscape scene with clouds, reflections of sunlight from a river or lake and from snow; take Adams’ Tetons and Snake River as a paradigm of this.

I have offered a few ways to deal with situations like this, and they work reasonably well in a wide variety of cases. However, a situation in which the approaches that I’ve already outlined may not be “best” include portraiture in strong ambient light. These cases can be exacerbated when you, as photographer, want to use wide open aperture to achieve a shallow depth of field and when you backlight your subjects to avoid harsh, direct shadows. In such cases, strong ambient light can be so extreme, on the one hand, and the light on your subject’s face can be so minimal, that some kind of compensation between the two is called for.

In the following, I’m going to provide some specific examples of this situation; but allow me to set up the case in a bit more detail. Let’s say you have some wonderful outdoor setting, beautiful sky, clouds, greenery, sea-shore, whatever. You want to make this part of your portrait capture. However, it is mid-day, because, well because that’s what it is; and you have a client who wants the picture there. It’s a wedding. It’s some client session, and you have to get the shot. Perhaps it was your idea in the first place. Doesn’t matter. You need the shot.

You could shoot with the sun directly in the subject’s face, and take the harsh shadows and the squinted expression. You could put the subject’s back to the sun, and get some hair light for your effort, but then your subject’s face is in deep shadow, with no catch light in the eyes and 5 or 6 stops lower than the background. You could side light your subject, but that just lengthens the shadows, emphasizes any wrinkles or other skin defects, and still leaves them squinting. None of this yields the image that you want or that the client wants. So, there’s a problem to fix.

The Solution: A Recipe

The essence of the solution is to provide another lighting source for the subject’s face that is independent of the sun and under your own control. With the Nikon CLS, the steps are fairly straightforward, and you can vary the recipe that I’m about to outline.

I expect that you’ll have your camera set for either aperture priority or in manual mode. Make sure that your camera body is capable of handling remote flash in Auto FP (Focal Plane) mode. Set your camera for remote flash and that the shutter speed will handle Auto FP (typically marked with an asterisk at the fast end of the shutter speed option list). First, spot meter off your subject’s face. Next, add in enough exposure compensation that you are not burning out the highlights in the background. Note this value of exposure compensation. Go to the Flash control settings in your camera body and set the flash for TTL (Through the Lens). Then add in the same amount of positive flash compensation as you just entered negative exposure compensation in the camera. You may need a VLA (voice-activated light stand, aka an assistant) to hold the flash for you. Place the flash so that it is in a “sun-like” position; that is, so that catch-lights in the subject’s eyes are at around 2 or 10 o’clock. If your camera body has a pop-up flash, like my D700, then make sure that it’s set for either no power or at an extremely low setting for fill. The remote flash will do all, or most, of the work. If not, like a D3 for example, you’ll need one speedlight for control and one as remote. [If the sun is extremely bright and you need a significant amount of flash power, you can work with two or more remote speedlights on a single channel. Camera settings are the same in this case. The CLS will figure out the right power for each remote speedlight.]

That’s the recipe. It will be very nearly right. You may want to fine tune either the flash compensation or the exposure compensation by 1/3 stop at a time.

Zone Analysis

Let’s consider this scenario in a bit more detail. We would usually place a Caucasian subject’s face on Zone VI. Perhaps we find that when we meter off the subject’s face, and then dial in enough exposure compensation to just avoid blowing out the highlights, we’ve entered, for example, -2 1/3 stops. Under normal circumstances, the subject’s face would have been Value VI, but our meter has placed it on Zone V. The highlights are at, say, Zone VIII, after we enter -2 1/3 stops of compensation. After compensation, Value VI is at Zone II 2/3. Up at Zone VIII, we have placed Value XI 1/3. Given that we have picked an ISO and an aperture setting, the only free variable we have left to get this exposure is the shutter speed. If it’s a bright day, and we’ve got a fairly open aperture, we’re going to have a very fast shutter speed, even at a low ISO, probably something faster than 1/1000 second.

I said under normal circumstances, and that would be the case if the subject’s face was lit by the direct ambient light; but if we have the subject facing away from the sun, then it’s more likely that their face was at Value IV when we metered. The meter gave us the exposure to place Value IV on Zone V in that instance. Instead of pushing the exposure down a stop (as in the previous case) it’s been pushed up a stop. In our recipe, we’ve knocked that down by 2 1/3 stops, so Value IV is now at Zone II 2/3. That puts Value IX 1/3 at Zone VIII. This is a more likely case, since our backlit subject will almost certainly have their face in shadow.

Variables to control

To get proper lighting on the subject’s face, we’re adding light from our speedlight. We have independent control over its intensity, and our aim will be to adjust this intensity to place the subject’s face on Zone VI (assuming Caucasian skin, I’ll consider alternatives in a moment). If we had no flash compensation with the Nikon CLS and with the speedlight in TTL BL FP mode (through the lens, backlit, auto focal plane), the CLS would aim to provide enough light to get our subject’s face to Zone V. You might think that cranking up the flash compensation by 2 1/3 stops is a bit much, and that is possibly true. But there is a strong mitigating factor at work, and that is Auto FP mode.

Auto FP mode and Flash Sync

Here is a digression on Auto FP mode. In a typical DSLR, the shutter involves two “curtains”. The front curtain opens to reveal the sensor, thereby allowing light to fall on it and create an exposure. The rear curtain then sweeps across the sensor to close the sensor off to light, ending the exposure. Naturally, it takes a certain amount of time for the front curtain to open completely and for the rear curtain to close completely. This implies that there is some shutter speed such that the rear curtain starts to move, closing off the sensor to light, at the instant that the front curtain has just completely opened the sensor. For any shutter speed faster than this mechanical limit, the placement of the front shutter and the rear shutter at any given time during an exposure is really a narrow window that is sliding across the sensor, letting in light through a “crack”.

Think about this for a moment. Suppose that the fastest that the front and rear shutters can move across the sensor is 1/250 second. At this shutter setting, the front curtain has just cleared the sensor when the rear curtain begins to close it. Now, we up the speed to 1/500 second. When the front curtain is 1/2-way across the sensor, the rear curtain begins to close. It still takes the front curtain 1/250 second to cross the sensor, but no single pixel is exposed to light for longer than 1/500 second. The rear curtain covers over each row of the sensor exactly 1/500 second after the front curtain opened up that row to light.

The same would be true at, say, 1/1000 second. Now, the front curtain is only one-quarter the way across the sensor before the rear curtain starts to move. Only one-quarter of the sensor’s area is open to light at any time. Again, it still takes 1/250 second to achieve the entire exposure, even though the shutter setting is 1/1000 second. That’s not the time that the exposure took. It actually took 4 times as long. It still took 1/250s. No single pixel was exposed for longer than 1/1000s, true; but the exposure’s total time was 1/250s. This can yield some odd effects, especially with fast moving objects. You might be thinking that with a 1/8000s shutter speed, you’d really be stopping motion; however, your total shutter time is still 1/250s or so with a typical DSLR. Even at 1/8000s shutter speed, what was exposed on the bottom of the image happened 1/250s before or after what was exposed at the top (depending on how the curtains move in your camera, top to bottom or vice versa).

Front Curtain Sync

So far, I haven’t talked about flash. Let’s throw flash into this mix. The default flash mode is “front curtain sync”, which means that the flash fires a burst at the point in time when the front curtain just clears the sensor. Let’s say you’ve got a 1/60s shutter speed. Typically, the flash burst takes about 1/1000s. Assume that there is some ambient light in your scene. The front curtain opens, the intense flash “stops motion” for 1/1000s yielding the dominant component of the overall exposure that the sensor accumulates in the time it’s opened up to light; but the rear curtain doesn’t move for about another 1/60s. Anything that does move will be moving after the flashed image. This can yield some odd effects if there is motion going on. For example, say you’re shooting something moving like billiard balls or a car or someone’s hand. The flashed image happens and then the balls or car or hand keeps moving. The final image shows a “ghost trail” due to ambient lighting that seems wrong, since it happens after the stopped image due to the flash.

Rear Curtain Sync

In the case of strong motion, “rear curtain sync” makes more sense. With rear curtain sync, the flash is fired just before the rear curtain begins to move. Suppose again that you’ve got a 1/60s shutter speed. The flash is fired just before the rear curtain moves; and typically, the full burst might take 1/1000s for the flash to complete. Ambient light that might yield “ghost trails” of moving objects happen before the flash fires. The final image now shows the motion of the billiard balls, the car, or the hand, moving causally into the bright image that was “stopped” by the flash burst.

Auto Focal Plane Sync

These two flash modes, rear curtain or front curtain, make sense for shutter speeds slower than the maximum sync speed. Suppose that that maximum sync speed is 1/250s. If we fire with front curtain sync and our shutter speed is 1/500s, the last half of our image will be dark. If we fire with rear curtain sync, then the other half of our image is dark. This is where Auto Focal Plane sync comes in, for those camera bodies and speedlights that support it. With Auto FP enabled, and a 1/500s shutter speed, the flash is fired twice at half power: once to illuminate the scene for the first half of the sensor just before the rear curtain moves (rear curtain sync) and once again just as the front curtain clears the far side of the sensor (front curtain sync).

Other combinations are possible. Suppose we go to a shutter speed of 1/1000s. Now with Auto FP, the flash will be fired four times during the exposure. And so on. Remember I said that a typical flash burst might take about 1/1000s. What happens when we’re at a shutter speed of 1/8000s in Auto FP? Unfortunately, we can no longer squeeze all of the power out of a speedlight at these very high shutter speeds. The exact “break points” depend on the combination of the camera body and the speedlights being used. The camera body will define the maximum flash sync speed and the shortest possible shutter speed. The speedlight model will define flash power and the amount available in Auto FP mode at any given rapid shutter speed setting. The various options can become complex quickly. Using third party radio controls with Nikon bodies and Canon speedlights can, apparently, yield better results than with the Nikon CLS only (see Rob Galbraith on this. Nikon builds some power limits into their speedlights for lamp protection that Canon does not use.)

Reduced Power in Auto FP at high shutter speed

The bottom line is simple though: in Auto FP mode, available flash power falls off quickly at shutter speeds above 1/1000s. We can compensate to some extent by adding more flash compensation and by adding more speedlights to a channel. The bottom line is that mileage will vary. You will just have to experiment with your flash power, holding other exposure elements fixed (ISO, aperture, shutter speed) to get the right settings. The recipe above is a rough-and-ready method to get you into the ballpark without much tweaking. [Note that working at a higher ISO setting would usually make life easier on your flash, but with Auto FP mode, this may yield worse issues with high shutter speed and lost flash power.]

The model and Complexity

I’ve offered you this recipe based on a portrait of a white (Caucasian) model. Such a person may not be your model; in fact, a person might not be your model at all. Perhaps you’re shooting a flower. Perhaps you’re shooting a person and a motorcycle. The basic principles remain the same. First, understand what Value your subject has relative to Value V in the ambient light. Next, come to terms with how much flash power you might require in order to “beat back the sun.” Shooting a small subject, like a flower or a face, might turn out to be fairly easy. Shooting a large and complex subject; e.g., person backlit, inside car with specular surfaces, etc., may well require lots of speedlights in well-chosen locations, modifiers, radio controllers, and so on. In spite of the apparent complexity of such a subject, it can be broken down into component parts. Break your scene down into a set of independent surfaces. Think of each independent set of surfaces of your lighting problem in terms of Values and Zones. In what I’ve described so far, we have two independent components: a landscape and a portrait subject. Suppose we add a third element: we have a landscape, a car, and a subject inside the car. We expose for the landscape, and then add independent lighting elements inside the car for the portrait and outside the car for it. And so on. The aim of the game remains the same: get Value V right in each independent aspect of the overall composite image.

Complexity and balance

Of course, when I say, get Value V right, what I mean is that you want to place some specific Value of your choosing onto some specific Zone of your choosing. The simplest choice is Value V on Zone V, but as you should be aware by now, that’s not the only choice by any means. In the simplest and least complex situation, you can place Value V on Zone V and everything falls into place. In a slightly more complex situation, we place Value X, say, on Zone VIII, and then use post processing techniques to pull our mid-tones back up. A similar kind of situation might involve a “short-scale” subject, as discussed previously, in which we use levels or curves to expand the range of light in the original scene in post processing.

In these cases, we really have only one element to control at the time of the exposure; namely, what Value to place on what Zone. Post processing provides some additional freedom of choice after the exposure is made, but this does have its limits.

By adding independent sources of light, we give ourselves more degrees of freedom in the creation of the exposure. We cannot control the sun, for example, so our sole degree of freedom there is to pick an exposure setting that will, say, control for highlights; for example, place Value X on Zone VIII. If this means that a portrait subject’s face now falls on Zone III, post processing may not do the trick. So, we add flash (or reflector), and we now have a new degree of freedom: we can independently light the subject’s face and control that light’s power and, if necessary, modify it with umbrellas, scrims, reflectors, barn doors, and all of that paraphernalia. However, the aim of the exercise is simple: use this independent light to place a Value on a Zone. In the situation that I just outlined, the exposure Zones have been defined by the ambient light. This implies that the flash power must be controlled so that the Value from the subject’s face is consistent with, say, Zone VI, assuming that we desire a “normal” exposure of the face.

If we had another set of surfaces to our lighting/exposure problem, then the analysis and solution remain the same: the ambient light defines a mapping of Values to Zones, so we use lighting to set Values on surfaces. To give one more example, and then move on to practical examples, consider a portrait subject standing in the doorway of a small shack lit by noon-day sun. The exposure will be defined by the strength of sunlight on the shack and the rest of the overall scene. Say we have clouds in the sky, and find that we need to add 2 1/3 stops of exposure compensation in order to avoid blowing out highlights in the clouds. The recipe that I described above tells us that we’ll need to add back 2 1/3 stops of positive flash compensation as a starting point to illuminate the subject’s face properly. The goal is Zone VI for the subject’s face as a starting point. But let’s say that we also want to provide some flash inside the shack to just make its contents visible through the door and windows. We now have to add an independent channel of flash within the shack and set the power, possibly manually, possibly with some other negative value of flash compensation, so that the interior surfaces fall on, say, Zone III or IV.

The problem becomes only slightly more complex if we decide that we want to light the interior of the shack behind the doorway and the window(s) independently. This means two independent control channels for flash with independent power settings. And so on. The essence of the process is the same: ambient lighting defines the mapping of Values to Zones, then we use flash power settings to place the Values on independent surfaces in the scene to exposure Zones. More independent sets of surfaces simply require more independent sources of controllable flash power (complete with lighting modifiers, control channels, and et cetera) to solve the problem. If the shutter settings defined by the ambient light situation force the use of Auto FP Sync for flash, then the remote control system and speedlights (or other lights) must be compatible.

Previous Page … ••• Next Page …