Safety shift has existed in some form as far back as the original EOS 1D. Its options have changed slightly over the years and though varying cameras but the intent is the same. The function, unfortunately, is somewhat cryptically labeled, potentially leading many to skip over enabling what can be a rather useful function.

Safety Shift Mode 1: Tv/Av

Safety shift mode one applies to all Canon cameras except the D30, D60, and Rebels (EOS ##0D and EOS 1000D). The manual describes it as follows.

1: Enable (Tv/Av)

This works in shutter-priority AE (Tv) and aperture-priority AE (Av) modes. If the subject’s brightness changes suddenly and the current shutter or aperture becomes unsuitable, the shutter or aperture is shifted automatically to obtain a suitable exposure.

From reading that, one might assume that enabling safety shift isn’t necessarily a good idea. The implication is that even in one of the priority modes the camera will be constantly overriding your shutter and aperture settings as the “exposure changes”.

The reality is that Safety Shift only kicks in when your metered exposure exceeds the limits of one of the camera’s ranges.

In Av mode, Tv/Av Safety Shift works something like this; suppose the camera is set to f/2.8 and the metered shutter speed is 1/16000^th of a second, 1 stop faster than the maximum shutter speed the camera can support. Without safety shift, the image will be 1 stop over exposed. With safety shift, the camera will automatically stop the lens down 1 stop so you’re shooting at 1/8000^th at f/4 and the image will be properly exposed. On the flip side, an exposure of 60s @ f/22 would prompt the camera to open the aperture 1 stop to get an exposure of 30s @ f/16.

In Tv mode, the Tv/Av Safety Shift behaves the same way only changing the shutter speed instead. For a required exposure of f/1.4 at 1/250^th with a lens that can only open up to f/2.8, Safety Shift would cause the camera to drop the shutter speed to 1/60^th to insure a proper exposure. Alternatively, an exposure that required an aperture smaller than the lens could produce would result in shortening the shutter speed until the exposure was correct.

What exposure is shifted?

Safety shift uses the meter to determine the exposure. However, exposure compensation settings are factored into the exposure before the safety shift is applied. Therefore, if one is shooting a high-key scene with +2 stops of exposure compensation set, the compensated exposure will be the base exposure and the Safety Shifted exposure will be equivalent to the compensated exposure.

Safety Shift Mode 2: ISO

Canon’s EOS 1 series bodies since the Mark 3s have a second Safety Shift mode to adjust the ISO instead of the aperture or shutter speed. The manual entry for mode 2 reads:

2: Enable (ISO Speed)

This works in Program AE, shutter-priority AE, and aperture priority AE modes. When the subject’s brightness changes erratically and the correct auto exposure cannot be obtained, the camera will change the ISO speed within 100-3200 automatically to obtain the correct exposure.

Like mode 1, mode 2 only alters the exposure once the shutter speed or aperture has reached the limits of what the lens or camera can do. In addition, like Tv/Av mode 2 factors in the exposure compensation that is set when correcting the exposure.

Faking Auto ISO on a EOS-1D(s) Mk. 3 or newer

Auto ISO is either handy or completely useless, depending on the implementation and photographer. Nikon’s implementation is arguably one of the best. In starts with allowing the photographer to set their preferred ISO, from there the photographer can configure a minimum shutter speed and a maximum ISO for the camera to use.

In other words, a camera could be configured so that the shutter speed won’t drop below 1/60^th and the ISO won’t be set above ISO 800.

If you aren’t providing at least that much, the implementation borders on worthless.

Fortunately, the EOS 1 series bodies don’t provide an auto ISO implementation as such at all. However, mode 2 (ISO) Safety Shift combined with the built in shutter speed limitations (Custom function I-12) can be used to mimic Nikon’s auto ISO to some degree; you can’t register a lowest shutter speed faster than 1/60^th and you can’t stop the camera for shifting outside of the registered ISO limits (Custom Function 1-3).

However, even with those limitations this “fake” implementation of auto ISO provides more control than Canon’s auto ISO as implemented on any other Canon camera.

In my experience, Safety Shift, especially the ISO shift on 3^rd generation EOS-1 bodies, provides the latitude to be able to quickly capture a fleeting moment without having to worrying about overexposing an image because the camera couldn’t reach a high enough shutter speed or narrow enough aperture. Moreover, coupled with the ability to limit the lowest shutter speed to something approaching a generally hand-holdable speed (1/60^th) you have a rather good, if slightly more complex to configure auto-ISO system.

I’ve previously written about my frustrations with calibrating my camera’s auto focus system. Frankly, I hate it. It’s finicky, time consuming and requires a great deal of precision to setup and execute properly. In addition, most targets aren’t large enough to test at actual working distances so there isn’t any guarantee that the results are going to be good for normal photography anyway. The worst part is that the whole system is largely open to interpretation.

However, there has been a solution all along. It just seems that nobody had really thought of it until recently. The solution comes from the ability to remotely control a tethered camera’s focus while looking at the live view image on the computer. Even better, there is now software that automates the process and that software is the AF Calibrator feature in Helicon Remote, part of Helicon Soft’s Helicaon Focus.

What Helicon Remote’s AF Calibrator does is use the camera’s built and its own contrast detection algorithms to determine what the sharpest focus setting is without having the user guess. Even better, it does this automatically and the result isn’t a hard to interpret picture but an actual setting for your camera. Further, because this approach doesn’t require reading a ruler off towards the edge of the frame, precise alignment and target design is much less important and setup is a snap.

A Quick Review of Traditional Method and its Problems

Let’s, quickly look at what you need to perform AF calibrations the old way.

1. Place your camera on a tripod and the test target on a flat surface.
2. Carefully align the camera and target.
3. Focus the lens to infinity and then let the camera autofocus on the target and take an exposure
4. Change the AF Microadjust setting and repeat step 3.
5. After all, of the focus step images have been made load them into an image-processing program and evaluate which one places the focus in the right place.

It doesn’t take doing this more than once to realize that it’s a pain and quite problematic if not done extremely carefully. For starters, the target, scale and film plane must be aligned with a high level of precision. This is necessary since the scale is shifted away from the target.

Further, the target can’t be anything that’s handy; it must be specially designed strictly for focus testing. It also has to have specific design features, like a high-contrast focus point sounded by no other detail.

On top of that, very few AF calibration targets are suitable for use at normal working distances. Canon, for example, recommends that all AF tests be done at 50x the lens’s focal length. If you’re testing a 70mm lens, that works out to 12 feet. A 200mm lens requires more than 32 feet between the camera and target.

Finally, you must manually step though each setting (for most cameras that’s 40 images from -20 to +20) manually refocusing the camera to infinity between each shot and letting the AF system refocus.

At 15' with a 70mm lens (even on a 1.3 crop body) this test chart is almost completely useless. Not only isn't there enough room insure that the camera will lock on the target bar, but the scale is smaller than an f/2.8 lens's depth of field.

All told setting up and shooting an AF test, using a paper target can take a couple of hours or more for a single lens. Even commercial solutions designed for focus testing only improves upon that slightly. Moreover, the results are dependent on the photographer’s ability to interpret what he’s looking at accurately.

The Live-view Method

Almost all of the problems with the traditional technique can be solved quite easily if we can insure that the focus point and measurement point are the same. Unfortunately, doing this precludes easy interpretation of the target by the photographer. Fortunately, this type of measurement is something that can be done readily by a computer. In fact, it’s how contrast detection auto focus systems, like those in point and shoot cameras, work. Additionally, since the focus settings on many cameras can be controlled remotely when they are tethered to a computer, the whole process can be automated in software.

In the broad strokes, for “live view” focus testing you:

1. Place the camera on a tripod and align the camera and target. The two planes should be close to parallel but ultimate precision isn’t necessary.
2. Using the center AF point, focus on the target by half pressing the shutter release.
3. Connect the camera to your computer and fire up your remote control software.
4. Using the smallest AF shift amount move the focus forward and backward while keeping track of how many times you click the button in either direction.
5. Observe the sharpness of the image each time you change the focus. The number of button clicks that made the sharpest image will be the value you enter into the camera’s AF adjustment setting.

This works well for a number of reasons. First, since this method focuses and measures at the same point, any errors induced by misalignment are negligible at best. Even better, there is no guesswork involved in coming up what adjustment to use. This is because the smallest step that the camera can shift focus is the same size as an AF micro adjustment step. Finally, since there is no need for a specialized target, the procedure can be done at normal working distances where it will better reflect the real world.

The only hurdle is judging the sharpest image, and this is where the computer takes over. Helicon Software’s Helicon Remote presents us with, as far as I can tell the first solution to computer aided AF calibration.

Live View Focus Testing in Helicon Remote

As I’ve said, Helicon Remote simplifies the process of Focus Calibration significantly since it eliminates the need to determine which image is the sharpest.

Setup for focus testing is as follows.

1. Attach your target on a vertical surface like a wall or door. In this case the target only needs to be a high-contrast dot or line, in fact the less there is for the computer to process the better.
2. Make sure the target is well lit with continuous lights and the more light the better. Since this system uses live view and doesn’t actually take photographs, a flash won’t work. Also, the more light available the less noise there will be and this in turn will makes test more accurate.
3. Place the camera on a tripod approximately 50 times the focal length away from the target. I use the focal length (in mm) divided by 5 or 6—whichever is easier to do in my head—in feet since it’s easy and close enough.
4. Align the camera and target. I do this by leveling the camera with a hot-shoe level then adjusting the height of the tripod so that the lens and target are vertically aligned. Finally insure that the camera is roughly straight out from the target (having wood or tile floors with the joint lines running perpendicular to the wall is handy for this, otherwise eyeballing it will get you close enough).
5. Hook the camera’s USB connection up to a USB cable (but don’t plug the other end into your computer). What you want to avoid is plugging in a cable at the camera during the test.
6. Focus the lens to infinity and half-press the shutter release to allow the AF system to lock on to the target, make sure the AF lock indicator has come on.
7. Start Helicon Remote and plug the USB cable into your computer.
8. In Helicon Remote, click the AF Calibrator button on the toolbar along the top.
9. Follow the steps in the AF Calibrator dialog.

Conclusions

I’m much happier with this method for AF adjustments than I have been with any other method I’ve tried so far. However, I haven’t yet used it enough to really gain an unshakable confidence with it. In the tests I’ve done so far, the results from Helicon Remote coincide fairly well with what I’ve found from other methods.

The biggest benefit here is the speed and ease of interpretation. With the old way, it would take me up to 30 minutes or more to insure the test was. Even with my best testing procedure, it still took quite a bit of time to insure everything was aligned properly. With this method, I can have the target and camera setup, aligned and ready to test in about 5 minutes. After that, the tests take about 30 seconds per run. In the time it took to get good results out of the old method, I can now triple check a lens at all major focal lengths.

Additionally, the Helicon Remote AF test goes far beyond what can be set in camera, +/- 60 points versus +/- 20 points for almost all cameras, so it can be used to determine if a new lens is out of spec.

From start to finish it took about 2 hours to get everything setup, designed, groomed, ready to go and shot. Fortunately, I already had my lighting designed so it was a matter of quickly dialing in the power settings. The real chore was herding popcorn where it needed to be.

The setup overview image below, shows the general overview of the setup I came up with. The background was a readily available blanket that I tend to favor when I need a dark background to paint with light.  It has raised ribs in it so the background takes on some texture but doesn’t overwhelm the image.

Setup Overview

The popcorn scene was setup (well it had been taken down when I shot the setup shot) on a piece of black corrugated plastic (so as not to get a sheet of foam core all buttery). That was placed on top of a tub (seen on the couch) to raise it a bit higher and give me some room to work and hide one of my two flashes under and behind the subject. Having the second flash on the table turned out to be helpful, as I was one lightstand short of what I needed.

Now to dial in the light; this turned out to be my biggest limiting factor in many ways. I started with the main light at 1/16th power and the background light at 1/8^th power. Since the 580Ex II is about a stop more powerful than the 430Ex; at equal powers the flashes will have a 1:1 ratio between them even though they are set differently.

I’m often asked how I figure out my exposure settings–aperature and shutter speed–for any given shot…

See how after the jump…

Lighting Tests

How I Figured Out What Settings to Use

The first thing you have to consider is what your most important exposure setting is. Are you shooting a static subject from a tripod? Or are you shooting moving things hand-held?

I’m often asked by people how I figure out what aperture or other exposure setting I should use for a given shot. So I’ll walk thought that here as this is an interesting case due to complexity and power limits.

The largest factor in producing this image is depth of field. I need certain things to be in focus, like the popcorn in the upright container and the two labels that say popcorn. Since I don’t have a tilt-shift lens that would allow me to control the placement of the area of sharp focus with out stopping down I’m forced to stop down to a narrower aperture.

I don’t have to worry about shutter speed for a number of reasons. They are in no particular order, I’m shooting from a tripod, the scene is static, and the scene is completely lit by strobes. Since the strobes contribute all of their light in a very short period, I can use the highest shutter speed that allows everything to sync. In this case because I’m triggering a master flash with old style non-TTL Pocket Wizards and workout outside of the design of Canon’s wireless flash system (see note) the highest flash speed I can use is about 1/100^th of a second. Therefore, that’s my shutter speed.

For this shot I started by picking an aperture, I think it was f/16 and taking a test image. It was dark. Time to start increasing the flash powers.

I started with the dimmest and least powerful flash, the 430ex on the background. When that got to full power and I still wasn’t getting any appreciable glow on the background, it was time to start opening up the aperture. Two stops down to f/8 and we were getting there.

At this point, I was starting to open my aperture enough that I was concerned about depth of field being a problem. The only thing to do now is increase the ISO. In the case of this shot, I ended up at with ISO500 giving me the background illumination that I wanted. Camera settings ended up being 1/100^th, f/8, ISO 500.

Now for the key light; it’s simply a mater of dialing the power of the key light up or down to get the proper exposure on the foreground elements. I lucked out here, as that ended up being exactly where I had the flash set, 1/16^th power.

In the case of this shot, the aperture was limited by the performance of the lowest power flash. If I had a second 580Ex II back there, I could have stopped down a stop further due to the power difference.

a lesson in balancing dis-similarly powered flashes and being aware of the transmission factors of your gels if you’re using them.

This goes as a lesson in balancing dis-similarly powered flashes and being aware of the transmission factors of your gels if you’re using them. The key light was set four stops below full power, while the background flash was set at full power. One of those stops is accounted for in the difference between flash powers, the other three are due to the Rosco Storaro red gel that was used to color it. It’s my favorite color so far for getting a rich saturated red; unfortunately, it eats three stops of light.

Sculpting the Scene

With the lighting setup, it was time to get to the actual fun part, placing individual popcorn kernels. I knew ahead of time I wouldn’t have enough popcorn to make the whole scene I wanted form popcorn alone. So I had to add a lot of filler. In this case, the filler was printer paper.

I started by crumpling sheets of paper to fill the bottoms of the containers and form the cores of the hills of popcorn. Then I covered the tops of the hills with crumpled but smoothed out paper.

Adding Popcorn

The final step in preparation was to add and arrange the popcorn. It had cut off only a corner of the bag of popcorn so I had some control while pouring it. However, it turns out where ever I didn’t have a lot of tabletop or a paper dam installed popcorn inevitably fell off the table and went everywhere, “cleanup on isle 1”.

I shot this tethered, with Lightroom auto importing the images. The big preview with a full RAW histogram was definitely helpful in insuring I had the right exposure dialed in, as the in-camera histogram is often not quite as helpful as it would seem.

The scene was shot at 33mm, on my 1.3x crop 1D mk.3 that worked out to 43mm or just a smidge on the wider-than-normal side. I was a little surprised at this, but the tests at longer focal lengths didn’t work out nearly as well.

The remainder of the time was spent moving individual kernels around until I had no obvious holes in the field of popcorn, the right patterning in the “hills” and a properly full container.

All told, I shot about 60 frames to get everything dialed in and nailed, including some with the focus adjusted to varying positions just in case (this actually was very important).

How Things Could Have Been Easier

What could have made things easier? Well for starters, a second 580Ex for the background light, it would have given me a stop narrower aperture, so I would have been shooting at f/11 and not f/8.

Moving up a step from that would have been the jump to full studio strobes with modeling lights. With the limitations I had on shutter speed, controlling the ambient light levels was important to keeping them from influencing the scene unexpectedly. The modeling lights and flash power has two advantages. More flash power, especially for the background light, would directly translate to a narrower aperture and more depth of field. Second, the modeling lights would allow the scene to be brightly illuminated for focusing and composition and then only lit by the strobes for the actual exposure.

All in all though, I’m happy with the results given the time, conditions and equipment I had available to work with.

Final Product

Next time you’re shooting fill flash with on-camera bounce flash, try this. Rotate the flash head 90° so that the normal up-down bend is pointed in the direction you turn your camera when you shoot a vertical. If you rotate the camera so that the grip is at the top, turn the flash head towards the grip. If you do it the other way, with the grip pointing down, turn the flash head away from the grip. This way you can quickly rotate the flash to keep it pointing up.

It doesn’t make a difference to the light bouncing off the ceiling and it’s one less angle you have to rotate your flash though so it’s quicker.

Like this, if you rotate the grip towards the top for a vertical shot.	Or this, if you rotate the grip towards the bottom for a vertical shot.
Not Like This

Last time we looked at focusing using a ground glass and the problems it presents as frame sizes and view finders get smaller. This time we’ll look at how to find the distance to something with out leaving the camera or using a ruler.

Finding the Range

What’s needed is a way to measure the distance to the subject with enough accuracy to allow for correct focusing. A tape measure would do, but probably isn’t going to work well for photography. Even if it did work, in many ways, it would only be marginally better than guessing with out a guide. Just imaging asking the important public figure you’ve been hired to photograph to hold one end of the tape measure while you setup the camera.

There is, however, another way to measure the distance to an object with out physically extending a measuring stick to the object. This works because of the trigonometric relationship between angles and the lengths of the sides of a triangle. It doesn’t take long to realize that using trigonometery as a base, the distance to the subject could be computer with out ever leaving the camera.

Figure 2: Overview of the formed to determine the range to the subject when using a rifle scope to calculate range.

Figure 3: Measuring the angle by reading the scale in the scope.

Figure 4: The equation for ranging. (d = distance to subject, s = subject size, α = angle)

To make this trigonometric solution work you do need to know one thing ahead of time, the size of the subject. In addition to that you need a way to measure the angle of the subject and a formula to convert that into a distance. One approach, and what’s used in long distance rifle shooting, is to use a scale in an viewfinder or scope to measure the angle the subject covers from your position. The scale is marked at positions that correspond to known angles. Commonly, at least in rifle scopes, these marks are made in milliradians (mils) or minutes of angle (MOA).

If the hypothetical reticle in figure 3 is marked using milliradians then the subject covers an angle of 4 mils¹ and the subject is known to be 6″ tall. Solving the equation shown in figure 4 for the values given, and a distance of 41.6 inches² is calculated.

Whether or not this is more practical than a tape measure is probably up for debate, what isn’t debatable is that this still isn’t practical for photography. Accuracy is dependent on how well you know the size of the subject and how accurate you can measure the scale and the whole thing is predicated on being able to work out an equation on the scene as a table is still problematic because of the errors in subject size. Fortunately for people shooting rifles, an error of a few feet has little impact on a shot that’s covering 100s of feet. However, an 85mm f/1.4 lens used to make a portrait has a total depth of field of only 4.1 inches at 10 feet.

For what it’s worth, I’m not aware of this style of ranging ever being applied to common photographic equipment. It’s certainly not accessible to most people. However it does introduce the trigonometric strategy for calculating distance rather clearly and that’s what all focusing aids based on range finding use one way or another.

The Coincident Image Rangefinder

The solution to the having to solve equations is to cleverly build that into mechanics of the rangefinder and present it to the user in a simple, easy to understand manner. The basis for doing that is to turn the triangle around so that the length of the subject in the previous figure now becomes a fixed length inside the camera/rangefinder.

The ease of use problem is solved by presenting the user with a pair of superimposed images where adjusting the alignment of the images corresponds with setting the focus. This all comes together in a device known as a coincident image rangefinder (CIR).

In a coincident image rangefinder, the user knows they’ve found the correct distance because the two images in the viewfinder are perfectly aligned. Even better, because of the way the CIR works, the user knows (not that it’s strictly necessary in this case³), which way the lens needs to be adjusted to focus the image.

Figure 5: Reversing range finding triangle.

Reversing the range finding triangle presents us with two solvable  problems, measuring the angle and combining the two images. The second problem is addressed by a beam splitter used as a beam combiner. A beam splitter, like most optical elements, it can be used in both directions. That is, when light is fed into two of the sides of the beam splitter, it will superimpose them over each other and send them out a third side.

Figure 6: Parts of a basic coincident image rangefinder focused at infinity

Figure 7: Distance equation

Figure 7: Simplified light paths though a coincident image rangefinder

The second problem, measuring the angle formed between the two rays form the subject. This appears more challenging at first, however even that is relatively straight forward.

The law of reflection states that the angle of incidence is equal to the angle of reflection. From this two things become clear, first that when the mirror is rotated 45° as shown in figure 6, the camera will be focused at infinity. Second, we know that the angle between the two light rays is equal to twice the angle the mirror will rotated.

From the second point, the equation shown in figure 7 can be derived; where d is the distance to the subject, b is the rangefinder’s base length and θ is the angle the mirror is rotated. Further, this makes measuring the angle α (in figure 5) directly unnecessary to calculating the distance as it was done in the previous method.

The final piece of the puzzle is coupling the rotating mirror to the lens’s focusing ring, so the whole system can be driven by the operator simply by adjust the lens. The actual math is hidden from the user in this mechanical coupling.

How it all fits Together

When the lens isn’t focused properly, the reflected light from the mirror isn’t aimed at the center of the beam splitter. Thus the combined image is misaligned, the direction of the misalignment is also a queue to the direction of the focus error. Figure 8 shows an exaggerated schematic of a coincident image rangefinder at various focus positions.

In practice it’s not quite this simple and there are more elements included to insure that both images are right side up and to project framing lines into the viewfinder, but the concept is the same for all coincident image rangefinders. Visual alignment is the key here, and what makes a CIR easily accessible to all users. The complicated math is handled by the people designing and building the cameras.

This system is used in cameras that are collectively called rangefinders. Arguably, the most successful example of this type of camera is Leica’s M series of rangefinder cameras.

Now that rangefinder provides an easy way to determine when the lens is focused it would seem like this should be the solution to all focusing needs. Simply stick a coincident image rangefinder on top of a camera and it’s good to go.

Not so fast…

Problems with a Coincident Image Rangefinder

There are a few issues the coincident image rangefinder presents. First since a rangefinder doesn’t work though the imaging lens, calibration of the lens-mirror link becomes very important. Small errors in mounting distance, lens construction or even rangefinder construction can throw the focus off enough to cause problems and they aren’t visible until after the image is made.

The second, and more fundamental, issue is accuracy. The rangefinder’s base length dictates the angle the mirror has to rotate for a given focus distance, because of that it controls the accuracy of the range finder. To make the measurement more accurate, a longer base length is needed to increase the angle the mirror has to rotate though. To give an idea how much rotation is involved, the mirror in a hypothetical rangefinder with a 60mm base length will rotate just over 0.5° from the infinity position when focusing on a subject 3M away; at 30m it will have been rotated just over 0.05°.

The objective for the rangefinder designer is to choose a rangefinder base length that is sufficiently long to provide enough accuracy with the focal lengths and working distances that will be used with the system. On the other hand the size of the camera places hard limits on the maximum length of the range finder base length. In turn the limits of the rangefinder place limits on the maximum focal length that the system can reasonably support.

This method of range finding works well, in practice, for wide-angle, normal and short-telephoto focal lengths because the depth of field at long distances grows quickly enough to mask errors in focus. However, telephoto and super telephoto focal lengths require much greater accuracy at long distances. That, coupled with the fixed external viewfinder makes long telephot lenses unwieldy at best on this type of camera, and in general they aren’t available for rangefinder systems.

Conclusions

The coincident image rangefinder solves some of the problems of focusing a small format camera but not all of them. This is no slight of the design though. One things a rangefinder camera offers that was impossible to achieve, before digital at least, is a way to build a very compact camera that doesn’t have any large moving parts. In fact when the exposure is being made, the only moving part in a rangefinder camera is the shutter. This makes rangefinder cameras very quite, as well as eliminating mirror slap induced vibration reducing image quality.

Next time we’ll look at how to make a rangefinder thin enough that it can be placed in a through the lens situation as used in an single-lens reflex camera.

What makes a camera, and by extension an image produced by one, different from how we perceive the world with our eyes is focus. Our eyes and brain work together to insure that the world is always in focus regardless of where and what we look at. However a camera and lens cannot simply reproduce the world as we perceive it, nor is that in general desirable. Focus and depth of field are inherent artifacts of lenses and using them is one of the ways a photograph can fame the world in a unique perspective when making a photograph.

Focusing, the actual process of adjusting the lens’s position, has changed little over the history of photography. Even modern autofocus lenses still shift the position of one or more lens groups.  However the way we measure and control those changes has. Hopefully this series of articles will shed some light on the evolution of focusing.

What’s perhaps not immediately obvious is that one doesn’t actually look though the lens when working with a camera. What, in fact, you’re looking at when looking though the viewfinder is an image projected on a focusing screen. This may seem a bit odd at first, but it’s necessary because the lens is actually focusing light over a large area at a fixed distance to form the image on the film.

The Ground Glass, How To See Though Film

Figure 1: Focusing on a ground glass.

Imagine for a moment the problem that’s posed when one uses a view camera.  A view camera provides no viewfinder at all, and even if it did it would be rather useless given the amount of adjustments that can be made. Instead the operator looks thought he same lens they’ll be imaging with and makes their adjustments to movements and focus. However the problem should be pretty obvious, if the film is in place, there’s no way to see though the lens. If the film is removed, the photographer’s eye would have to be in exactly the right spot to form a proper image.

The solution is a ground glass, the most basic focusing screen. A ground glass is exactly what it sounds like, a sheet of glass that has been ground on one side to a matte finish. The matte finish is key, that’s what allows the image to form on the surface of what would otherwise be a transparent piece of glass.

The ground glass does offer some advantages, even though it might not seem that way at first. First, there aren’t any special optics between the lens and the image used to focus. This means that there is nothing to mask the detail allowing the photographer to insure focus is placed exactly where they want it. However as much as that is an advantage for focusing, it doesn’t make for a bright surface to look at. In fact, the basic ground glass focusing screen is the darkest focusing screen.

The other advantage has more to do with the size of early cameras. Most early photographic work was done with what we’d now call large format cameras, with frames larger than 60mm on a side.  When the focusing area is that large, it’s practical to employ another aid such as a loupe to magnify the image yielding even more accuracy when focusing.

The Need for a Focusing Aid

As photographic formats became smaller the problems with focusing changed. No longer was there room for a large glass screen that could be used with a loupe against and inspected while hiding under a curtain to reduce glare. Compounding things, the smaller film size allowed for new compact camera designs with the objective of being easily portable. Even if it was desired to use a loupe it would defeat half the purpose of the smaller camera.

A view of the focusing screen from a pre-war Zeiss-Ikon Ikoflex II

To illustrate the problems with a simple ground glass, the image to the right shows the view through an pre-war Ziess-Ikon Ikoflex. The illustration is a bit deceptive, as I actually ended up use a flash to illuminate the scene after focusing under as much light as I could get and guessing some. With out the flash the metered exposure for the screen was about 2 EV. The actual light levels metered at about 7 EV.

The easiest solution, if it could be called that, to the focusing problem is to simply not focus. In fact there have been several cameras made that used hyper-focal focusing (most cell phone cameras do this today) in an attempt to get around the focusing problem. However hyper-focal focusing removed the ability to use focus and depth of field to creative ends. It also be necessity limits the maximum focal length that can be used and fixes the aperature, usually to a small one around f/11 or f/16.

The next easiest solution is to simply guess. That is, mark the lens at points that correspond to different distances and turn the lens to the distance that seems about right. While it might make distances scales sound completely useless, they aren’t. In fact, they persist on many lenses even today though often in a simplified for.

Guessing works okay in some situations, but proves to be almost useless out side of them. The two biggest problems are when used with fast lenses, lenses with long focal lengths or worse both. Fast lenses, necessary for selective focus and blurred backgrounds, have very shallow depths of field necessitating more accurate focusing. Long focal lengths too have very shallow depth of field, and require accuracy at quite great distances before the depth of field is large enough to matter. The combination of a fast long lens compounds the problems even more.

Clearly a better way is needed. Next time we will look at how to find the distance to the subject without using a ruler or guessing, or even leaving the camera.

The key to changing the position of the bracketed shot, when shooting 2-frame brackets.

For quite some time Nikon’s mid- and pro-level bodies have had a novel and somewhat useful option for bracketing, a 2-frame auto exposure bracket. With the introduction of the EOS 1D Mark 3 in the spring of 2007 Canon brought that feature to their EOS-1 series digital bodies (the 1Ds Mark 3 will do this as well).

The catch is, while 3, 5, and 7 frame brackets are symmetrical around the middle exposure, with an equal number of over and under exposures, a 2-frame bracket is not. Nikon got around this by having two bracketing modes, +2F and -2F. Canon, however, didn’t seem to think it was necessary to include such an obvious way to control which direction the 2-frame bracket would be biased.

By default a 2-frame bracket has a non-biased normal frame and a negatively-biased bracket frame. In fact reading the manual one wouldn’t be faulted for assuming that the Canon 2-frame bracket only works as a negative bias.

The most obvious workaround, is to dial in a stop of positive exposure compensation; and in many cases that works fine. However, doing so limits the range of your exposure latitude from +/- 3 stops to+2 to -3 stops. If you want to keep the same +/- 3 stops of of exposure compensation but bias the bracket positively you’re seemingly out of luck

The truth is, there’s a side affect to another custom function that’s not obviously documented.

The trick to getting a “0,+” 2-frame bracket is actually one custom function away, at Custom Function 1-5. Custom Function 1-5 sets the AEB brackets sequence, giving us three options; 0, -, +; -, 0, +; and +, 0, -. What isn’t clear is that if you set C.Fn 1-5 to 2 (+,0,-) the 2-frame bracket becomes a positive-biased bracket not a negative-biased one. Either of the other two options (“-,0,+” and “0, -, +”) result in negative biased bracket frames.

AEB exposures with oEV, -3Ev and +3Ev of exposure compensation for a 2-frame bracket when Custom Function 1-5, Bracketing Sequence, is set to 0 or 1. Note the arrow on the vertical meter for the -3EV frame, indicating another exposure exists outside of what can be displayed.

AEB exposures with 0Ev, -3Ev and +3Ev of exposure compensation for a 2-frame bracket when Custom Function 1-5, Bracketing Sequence, is set to 2. Note the arrow on the vertical meter for the +3 Ev frame, indicating that another exposure exists outside of what can be displayed

It would have been nice if Canon had mentioned this instead of leaving it up to be discovered.

This one comes from DTown TV but works equally well on Canon bodes.

If you’re ever in a situation where you want to set the white balance in camera but you don’t have an appropriate target, there’s hope. If your camera has Live View it will show you in real time the affect changing the color balance has on your image. Of course this isn’t as accurate as creating a custom white balance. Fortunately, like most things in photography being close is more often than not close enough.

On an EOS 1D Mark 3 or EOS 1Ds Mark 3 you cycle though the preset white balance modes using the rear dial, and change the kelvin value (K) or switch between the 5 custom settings with the front dial while previewing the image. All all other dual dial cameras (40D, 50D, 5D-II), you can change though the preset modes using the rear dial.

Also if you’re a Nikon user with a dual dial camera with live view (D90, D300, D700, D3, D3x) you can use the front dial to fine tune your white balance in the amber/blue axis (in all but K and Pre modes) or change the kelvin value (in K mode) as well.

Of course, if you shoot RAW you can just as easily leave your camera in Auto WB and make any fixes in post production.

Creative uses aside, accurate white balance is critical to insuring the proper reproduction of colors in a photograph. In cases with difficult lighting conditions frequently we create and turn to custom white balances to insure the color in their images is correct.

However using the common copy or printer paper, which usually contains optical brighteners-chemicals that change the color response of the paper-can lead to bizarre and incorrect results under some conditions.

Background

The white balance controls how colors are reproduced in an image by changing the way the 3 additive colors (red, green and blue) are mixed. Any photographer who’s taken a picture in a room lit by incandescent lights with daylight film or even with most digital cameras and had everything look very orange should have realized that different light sources have different colors. For example, incandescent lights produce a very warm orange light with a lot of amber or orange color and most florescent lamps produce light with a green cast.

Our brains adjust for this based on experience and intuition (we know something white should be white and perceive it that way). However a camera doesn’t have that luxury, it needs to be told how the light illuminating a scene changes the color of things in the scene. In many cases, the camera is intelligent enough to figure this out on it’s own. This is the Auto White Balance mode.

However in very complex lighting situations where lights with several different color balances are mixed or when more accuracy is necessary. Most digital cameras allow you to create a custom white balance by photographing a target in the light you intend to work in order to work around these problems.

The requirements are simple, you need a way to show the camera a neutral gray target illuminated by the light that you want to balance for. Serious photographers will use a photographic gray card, a color check card or one of the myriad of white balance targets to insure accuracy.

Fortunately, or unfortunately, the target doesn’t specifically need to be mid-tone gray either, and many amateur photographers exploit this by substituting a sheet of white typing paper as their target.

This however is problematic.

Typing/printer paper isn’t designed with photography in mind. In fact it’s primary objective, besides holding the ink or toner that’s printed on it, is to make the print look good. One thing that definitely doesn’t make the print look good is not so white paper. One of the ways that paper manufacturers make paper whiter is to add optical brightening agents (OBAs) to the paper when it’s produced. OBAs are fancy dyes that absorb UV light and emit visible blue light. The result is that the paper appears bluer, which for some odd reason makes it also appear whiter and brighter.

The OBAs are the problem that makes using regular paper unsuitable for creating custom white balances. The paper is not necessarily neutral under all kinds of lighting.

The White Balance Test Setup

The plan was to use some different papers and generate white balances under both indoor flash conditions and outdoor sunlight conditions.

The papers used were:

• Office Depot multipurpose bright white paper – 20 pound 108 brightness (US) paper
• Kodak Bright White inkjet paper – 24 pound 108+ brightness (US) paper
• Hammermill Color Laser Gloss paper – 32 pound 90 brightness (us) paper

The three papers were laid out on a solid color background (in this case I used a sheet of black foam core) indoors under direct flash illumination and outdoors under sunlight.

I choose direct flash for 2 reasons:

1. Canon’s flashes communicate the proper color balance to the camera so I’d have a point of reference.
2. Second, direct flash eliminates any color cast from a bounce surface or photographic umbrella.

An exposure was made indoors with flash and again outdoors under the sun, with the camera set to auto white balance. Remember, Canon’s flashes tell the camera what the color temperature of the light they emitted is.

Both exposures were checked to insure that the papers were not clipping.

The image was then duplicated in Adobe Photoshop Lightroom 2.4 and a custom color balance was set using each of the white papers.

All images were rendered using the default Lightroom settings and the Adobe Standard camera profile.

Indoor Flash Results

Note: The images used in this article are rendered in the ProPhoto RGB color space. If your not using a browser that supports color management they will render incorrectly, however the differences in color should still be visible on most displays.

The table below shows each of the corrected white balances from the indoor photographs. The camera reported white balance was 6200K +6 magenta.

Indoor White Balance Images
Office Depot Bright White Multipurpose (108 brightness)	Kodak Bright White Inkjet, 108 brightness
6300K +3 magenta	6250K +5 magenta
Hammermill Color Laser Glass (90 brightness)
6150K +7 magenta

As you can see, indoors there is still some difference in the white balance but the papers all fall within 100K of the white balance the flash reported.

As far as I’m concerned the variations aren’t significant and could be due to any number of factors including any other dyes that the paper contained. Clearly under flash lighting there is not enough UV light to active the OBAs and affect the paper’s color and almost any of the color balances would be acceptable.

Outdoor Sun Results

The next table shows the color balances using the same 3 sheets of paper taken outside at approximately 1 PM EDT under clear sunny Florida skies.

Outdoor White Balance Images
Office Depot Bright White Multipurpose (108 brightness)	Kodak Bright White Inkjet, 108 brightness
6200K, -12 green	6400K -15 green
Hammermill Color Laser Glass (90 brightness)	Auto White Balance
5950K -6 green	5550K -7 green

The camera calculated the white balance at 5550K -7 green, and balancing using the black foam core board gave me a value of 5350K -2 green.

To my eye the two images balanced on the 108+ brightness paper have a yellow/amber cast. This is consistent with OBA’s causing the paper become bluer under UV light.

Further to my eye the AWB result looks fairly consistent with the way my eye perceived the color of the grass. Unfortunately, I didn’t have a gray card handy to get an accurate measurement.

Conclusions

The results are pretty clear to me, even though this wasn’t very scientific.

Most common office papers, especially bright white ones, have some level of OBAs in them. This makes them unsuitable as white balance targets under any condition where they may be exposed to UV light. Basically anything outdoors. The indoor tests showed that they could be used in a pinch if necessary as long as your lighting doesn’t emit a lot of UV (some halogen lamps are strong sources of UV light).

Outdoors, common office papers should be avoided completely.

The best solution of course, is to use a photographic gray card, the are inexpensive and guaranteed to be both neutral and not shift in color in any meaningful way.

Today concluded round 2 of serious focus testing. What’s different this time from the last? A new target, a new alignment strategy, and some new results. Last time I looked at autofocus, I had enough problems focusing fast lenses that I was growing concerned that there was either a systematic design flaw or similar error in phase detection AF systems. I was seeing behavior with fast lenses on both Canon and Nikon bodies that would inconsistently focus depending on a verity of situations.

The good news, the new target dramatically improved auto focus consistency on both platforms and the new alignment strategy shortened setup time and aided accuracy. The not so good news is that I’ve determined that there is definitely a problem with my EF 50mm f/1.8 II and it may in fact be a much wider issue; though I guess you could say, what do you expect from a $100 lens.

The Bad: AF with the Canon EF 50mm f/1.8 II

The Canon EF 50mm f/1.8 II–the lens that got me started looking at focus calibrating–is clearly a problem, at least mine is. I have over 500 frames shot with both a EOS 1D Mark3 and an EOS 40D, under carefully and not so carefully controlled conditions that lead me to believe, with quite a bit of confidence, that either my copy or the lens as a whole are problematic. Further, I’ve read a few reports with other photographers expressing similar similar behavior from this lens. In the end, I can only conclude that the chances are that it’s not my copy specifically. However, at only $100 it’s not really worth it for me to examine this further.

Quite simply, while the EF 50mm f/1.8 II was somewhat of a deal when it cost less than $100 I can’t say that any more. At the current price, perspective buyers should strongly consider either spending a few hundred more on the EF 50mm f/1.4 USM (not that I’ve tested it) or be well aware that their lens may inconsistently focus on their camera. If you’re going to go ahead and get an EF 50mm f/1.8 II, I’ve found that shooting in AI Servo provides the most consistent results, but even then it’s a good idea to shoot a couple of extra frames as backup.

The Good 1: Consistency

The second issue that came up in the previous tests was the lack of repeatability with a few lenses other than the Canon EF 50mm f/1.8 II. Specifically two Nikkor primes an 85mm f/1.8 and a 50mm f/1.8. The odd part was that other than the Canon lens, the problem was slightly less predicabel with the Nikkors. It was bad enough that I though maybe the problem was due to play in the AF drive mechanism or worse that turned out not to be the case. However that appears not to be the case.

In the time since the last post and this post I had been working off and on with my glass trying to figure out what was up with the AF inconstancy issue. The biggest reason I had been using Jefferey’s target was for the light gray marks along the center of the target. The theory being even if the target is slightly misaligned they’ll still give an accurate indication of where the lens is focusing. However try as I might I still had issues with fast lenses miss-focusing more than they should using even the 5% target.

The confusing and frustrating issue was that I couldn’t generally get the cameras to focus on the light gray pattern in Jeffery’s target. Which is the test he recommends for determining which density to use.

In a fit of frustration I tried a different target, one that simply had a white field with a black bar. Surprisingly I had very few focus problems with this target. So I went back and tried focusing on the low contrast gray area to see if it really wouldn’t focus. Turns out that at least on the EOS 1D Mk3, and after today’s testing the Nikon D700 (and by extension the D300, D3 and D3x), under some conditions with a fast enough lens the AF system is capable of locking on the to light gray text in the target.

The result, a quick redesign with a white central field and the repeatability of focusing picked up significantly.

The Good 2: Alignment

The final change was in the arrangement of the test setup to eliminate some variables and control some others more easily. The most important aspect of these tests is squaring the camera to the target. Some minor variation is workable but major errors are not.

Typically, the target is placed on the ground, or on a desk, and the camera is angled down at it. In this case it’s difficult to insure that the camera is square to the target, as you have to deal with alignments in several directions. Instead I’ve taken to mounting my targets to a vertical surface, such as a wall or large window. Ideally the target would be cemented to a flat surface such as a piece of glass, to insure it is flat; but again there is some room for error.

With the target leveled and on a vertical surface the the camera can be aligned simply by leveling the camera (I use a Hot Shoe Level) and adjusting the height until the center of the lens is at the same height as the target. No more trying to square the target and the camera and any number of other things.

Conclusions

There’s no point trying to preform AF adjustments on a lens that can’t repeatedly focus at the same point. Before even attempting to do any kind of micro you need to insure your lens focuses consistently at the same point. I do this by manually focusing to both infinity and the close focusing distance (macro) on the focus ring and allowing the camera to refocus. At a minimum with a new lens I repeat this 4 times, twice from infinity and twice from macro. If there is any discrepancy I’ll repeat it often enough to be able to characterize the problem as either user error or an issue with the lens/camera. If I can’t get at least 95% of my shots to focus on the same place for both directions with a single black line AF target, I pretty confident that the lens is problematic.

Second, target design is important. It’s my experience that several things need to be addressed in the target. First, the actual autofocus point must be clear of any diagnostic markings. It’s not simply enough to assume that there is one predominate mark and any low contrast marks won’t be detected. In addition the target needs a visually clear system to show DoF. Most targets use some kind of simple ruler, that’s simply not adequate. The reason I like Jeffery Friedl’s target is the inclusion of hashed lines of varying sizes. The hash marks make determining DoF significantly easier than trying to guess at fuzzy numbers.

There’s certainly no reason not to test and adjust your lenses to achieve best focus. However the process is not something that should be taken lightly and there’s plenty of room to shake your confidence in your gear.