Since that early beginning, more than 50,000 photographers have seen my exposure program over the years and many who persevered became excellent at using the method. I was glad I could help them. I know hundreds attended my field programs primarily to learn to expose difficult slide film. Little did they know that I would have loved to use autoexposure, if only it was accurate enough for my purposes. When the Nikon F5 first came out, my wife Barbara was thrilled to get the new camera. I was shooting Canon. Some of the folks who first brought the Nikon F5 to my workshops said they didn’t need to learn the system I used because the color matrix meter generated perfect exposures every time. I am always up to a challenge, so during the workshop we tested the Nikon F5 on automatic against figuring the exposure on manual with the in-camera spot meter. Of course, I got to pick the test subjects. I thought I could outfox this new color meter, so I was careful to select subjects where I suspected auto exposure would have trouble. For example, one scene I selected was a beautiful autumn color forest that was nicely reflected in a still lake in golden dawn light. However, in the foreground, I had three trunks of white cedars framing the scene in the background. There was no light on the cedar trunks which filled about 25% of the frame. How would the meter know if we wanted the autumn color in the background properly exposed or the silhouetted trees in the foreground properly exposed? Using manual exposure and the in-camera spot meter, I told my student to meter the bright orange leaves in the background and add 2/3 stops of light to the exposure which turned out to be a perfect exposure. The autumn colors in the background were perfectly exposed while the cedar trunks were in silhouette as we wanted. On automatic, the color matrix meter was thrown off by the 25% black in the foreground and overexposed the autumn color by about 1 stop. The Nikon F5 meter did recognize colors though and knew enough to add more light to a white snow scene. I tested it a few times and it added about 1 stop of light. But, my method was far more precise because I knew snow on a sunny day needs about 1 1/3 more stops of light and a full 2 stops of extra light on a cloudy day. If a dark moose or bison is large in the frame and surrounded by snow on a cloudy day, then I even add 2 1/3 stops to get some detail in the dark fur without overexposing the snow too much. The F5 didn’t know what I wanted and didn’t give me the results on automatic that I wanted, so I was never a fan of the accuracy of the Nikon F5's meter. It was still a great camera, though, but Barbara continued to use the camera on manual and spot metered to get proper exposure.

I realize I got a lot of workshop and seminar business because of my ability to explain exposure in great detail and back it up with plenty of examples. The system I developed has meant a lot to me, giving me control over exposure and earning me a lot of dedicated fans. Still, I only use the system because it works so well. The only time I ever used matrix meters with an automatic mode like aperture priority was to test it.  Otherwise, I found it to be far to inaccurate to use for capturing real images in the field.

For many years, I never really bracketed slide film either. However, when AEB (auto exposure bracketing) and shutter speeds and f-stops in 1/3 stop intervals became available on camera bodies, I began setting my Canon cameras that I used manually to automatically bracket the shot by 1/3 stop lighter and 1/3 stop darker. I did this because I might be off slightly when I guessed subject tonality, the film speed may actually be somewhat different than listed, or the processing could vary as well. Even shutter speeds and f-stops can vary a bit from what they are supposed to be. I call 1/3 stop exposure bracketing "tweaking the exposure". This system worked well for me over several years. Any slides that I thought were more than 1/3 off the exposure I was looking for were sent to the circular file except for some moody images of fog or sunrise/set images where I wanted an exposure different than how it looked in reality.

When I shot my digital Canon 10D for the first time, I set the camera on manual and exposed subjects just like I was shooting Velvia 50. I remember metering a yellow gray-headed coneflower and adding one stop of light. With the camera set on manual, I metered the bright yellow and the cameras meter balanced at 1/4 second at f/16 for example. I slowed the shutter speed down to ½ second to add one stop of light. Instead, I could have set the aperture to f/11 to get one more stop of light, but I wanted the depth of field that was available at f/16. I looked at my histogram and saw that I had a nice bell curve slightly to the right of the middle of the histogram. There was no clipping on either the left or right side of the histogram. And just to learn about histograms, I shot another frame where I metered the yellow and made no exposure compensation. Then I tried another where I metered the yellow and added two full stops of light. Checking the histogram in both cases showed I had moved the histogram to the left when I made no compensation and farther to the right when I added two full stops of extra light. But, in both cases, I still wasn’t clipping either side. I was shooting RAW digital files that offer the maximum control in Photoshop. Barbara showed me that by moving an exposure slider one way or the other, I could make the two images that I purposely improperly exposed look identical to the properly exposed image where I added one full stop of light. This demonstration was shocking to me! A Velvia slide that was over or underexposed by one stop was nearly always useless in my opinion. With digital, I could be a stop off and still make it a perfect (well almost perfect) exposure after the fact.

While I previously rarely used autoexposure because I required such exposure precision, perhaps it would be a possibility with digital. My trip to Kenya in 2004 was a good time to test digital. Since I knew that virtually nothing in Kenya that I would be photographing was all black or all white, I though the exposure latitude of digital which is greater than slide film and the ability to adjust raw files to get the precise exposure I wanted in the computer might make a case for trying autoexposure. But, why try autoexposure anyway? I am very quick at metering on manual, meaning I set the shutter speed and f-stop while monitoring the exposure meter in the camera, but it does take me a couple of seconds and is a distraction. If I didn’t have to turn the dials to set the exposure compensation I wanted, I would be slightly faster which isn’t critical when photographing a landscape, but most critical when wildlife action is the subject.

On the second day of my Kenya tour while in Samburu National Park, I set my Canon 10D digital camera to ISO 100, the meter to matrix or evaluative metering and exposure mode to aperture priority. I set the camera to f/5.6 depth of field to make certain my shutter speed was fast enough to be certain of getting sharp images. But, I didn’t overdo the shutter speed. When shooting lenses longer than 200mm, the minimum shutter speed I felt confident in shooting at is 1/125 second even when the camera is mounted on a tripod or cushioned with a bean bag. I actually prefer 1/200 second if I can get it. But, I don’t need faster shutter speeds such as 1/500 second or faster. If I found the camera was setting such shutter speeds, I adjusted the aperture to f/8 to get more depth of field, knowing I still had plenty of shutter speed. Of course, I did want fast shutter speeds in the 1/500 to 1/1000 second range when photographing wildlife action like vultures flying to a lion killed wildebeest that was lying in front of me.

I found not metering to be a real strange way to photograph wildlife. Naturally, I was suspicious at first, so I checked the histogram after each exposure to make sure no part of the scene was underexposed and especially not overexposed. The histogram graph makes this easy to determine. My camera even uses "blinkies" to show areas of the image that are overexposed meaning there is no detail in the pixels at that position. When I first used my digital camera, I was so used to shooting in low contrast light that I never saw any blinkies. When I learned my Canon 10D had this feature, I deliberately overexposed a scene by a few stops just to see the blinkies. Blinkies are a great aid to making sure overexposure is avoided in the image.

Let’s look at histograms closer. You don’t have to be a math whiz to understand what a histogram tells you. It is a very simple bar graph that even I can understand. A simple histogram shows you a graphic representation of the gray scale values from 0 (black) to 255 (white). Black is on the far left, pure white on the far right, and shades of gray that gradually lighten to the right are in-between. There is no perfect shape to a histogram since it merely shows you the range and amount of tones (shades of gray) in the image. You could have a nice bell curve if the majority of your image was around middle tonality, but some whites and blacks also were present. On the other hand, you might get a spike that goes off the top of the histogram. A subject that would do this is a set of cottontail rabbit tracks in the snow. Since virtually everything in the image is close to white, a properly exposed image should produce a histogram where the vast majority of the values are near the right edge of the histogram graph where the near white values are found. You could even get a histogram where you get a spike at both the dark and light portions of the histogram. Photographing a dark moose standing in a field of snow on a cloudy day will produce this result since nearly all of the tones in the scene are either very dark or very light with little else between these values. There is absolutely no problem with a spike of tones that go off the top of the histogram chart, as long as the spike isn’t on the far left or right side of the histogram. It just means a lot of pixels have that tonal value.

Histograms do show you when you have problems, though. Any histogram that produces a spike on the far left (black side) or far right (white side) indicates potential exposure problems. While you may have some pure blacks in the photo that you want to be black, it’s usually not a good idea to have a large spike on the left side of the histogram because you may get digital noise which looks like grain in a print. However, digital cameras are quickly reducing the noise problem with each new generation of cameras. Above all, it’s important to avoid clipping the highlight or right side of the histogram because overexposed pixels have no detail. An excellent histogram typically comes close to the right side of the histogram without clipping any of the highlights. But, you must be careful because most histograms only show you an average of the red, green, and blue color channels. A histogram might not show a spike on the right hand side when it averages the three color channels, but a single color could actually be overexposed without this being shown. Many cameras today offer an RGB histogram that does show you the histogram to each individual color channel. Many cameras are designed to make overexposed areas blink when you examine the image by looking at the LCD on the back of the camera. These overexposed areas that blink are typically called the "blinkies" by digital photographers. Looking for the blinkies is a fast way to see if part of the image is overexposed. If you do get blinkies, then you must reduce the exposure, so they are no longer present. If you are photographing a landscape, it might also be possible to use graduated neutral density filters to reduce the contrast in the scene, so the digital camera can handle the entire dynamic range of light or use the fabulous HDR technique (explained in our Digital Landscape Photography book.

However, it is possible to make an excellent exposure and still have a small spike on the right side of the histogram. If your image has specular highlights that might come from light shining off water or metal, then it could be fine since, no detail appears in the specular highlights anyway.

It’s clear digital photography is rapidly changing the way most photographers do business. Due to the wider exposure latitude digital has over slide film and the control you have over exposure in the digital darkroom, getting an excellent exposure isn’t as difficult as it was with Fuji Velvia 50, the last film I shot. I welcome this change even if it does cost me some clients who no longer need my expertise in learning to expose slide film. There are many exacting details I have learned about exposure that I know I won’t need anymore. If I can forget a few things, which is easy for me to do, perhaps there will be more room in my brain for the new information I must know such as monitor calibration, printing, and even how to use a digital projector.

Even though I have found I no longer need all the details that I learned over the years about exposing slide film, I am glad I have the background anyway. I think it is still worth knowing why we have to open up or add more light to the exposure when metering very light or white subjects like snow. Conversely, it is worth knowing why I had to subtract light when metering a black or very dark subject. When I have the time, I still plan to meter manually and compensate for reflectance (I am leaving this in for historical reasons. We now do things differently which is described at the end of this document). In fact, just yesterday I was photographing Barbara riding her horse among snow-covered aspens. Since I had plenty of time and the light wasn’t changing, I metered the snow with my digital camera on manual, added two stops of light, and fired a shot. I checked the histogram to determine that the white values in the scene were close to the right edge of the histogram, without clipping anything and no blinkies appearing on my LCD display. It was important to be close to the right edge of the histogram without clipping anything because I wanted to get as much detail as possible in the dark red fur of the horse and the dark clothes Barbara was wearing. I used that manual exposure setting to fill my CF card with 156 shots and all the exposures were excellent.

It’s clear that getting a great digital exposure is much easier for my students than it was with slide film. There is nothing like getting instant feedback to see that you got it right. I still think metering with the spotmeter in the manual mode and compensating is a viable option (again, we don’t do this anymore, but it is still viable). But, the method I have taught since 1978 will be simplified greatly (it is now even easier). Here’s a summary of how I would compensate for subject reflectance when metering with my digital cameras thru-the-lens spot meter in the manual mode.

Black or very dark subjects-subtract 1 ½ to 2 stops of light

Dark subjects-subtract 1 stop of light

Middle tone-no compensation

Pink, yellow, light blue, etc.-add 1 stop of light

White-add 2 stops of light

After applying these compensations and taking the image, always check the histogram to make sure nothing is clipped on the right edge of the graph to avoid overexposure problems. But, don’t get carried away. There really is no need to keep rechecking your histogram if you are shooting a series of similar shots where you are photographing the same subject and the light isn’t changing as long as you use manual metering. Any autoexposure mode has trouble maintaining a good exposure if the background changes or the size of the main subject changes from image to image. When photographing Barbara riding her horse in the snow, I metered the snow and compensated for the snow by adding two stops of light. Checking the histogram proved that everything was fine. I rechecked the histogram again after about every 40 shots just to be certain nothing had changed. It didn’t. The variable that was most likely to change is the amount of ambient light illuminating Barbara and her horse. But, the cloud cover remained steady, so nothing changed.

To sum everything up, the light sensors in digital cameras can record a larger dynamic range of light than slide film handles. With adjustments that are available in software, exposures that are fairly close to being right on can be made near perfect, if no data is lost from being clipped at the dark end or especially the light end of the histogram. This makes the possibility of effectively using automatic exposure for subjects that aren’t all dark or all light in reflectance quite possible. I found using aperture priority auto exposure with the matrix meter to be quite effective on my last Kenya safari where nearly everything was some shade of brown, tan, yellow, or gray. Virtually nothing was all dark or all light. On the other hand, photographing bison or elk in Yellowstone during the winter was another matter altogether. I found the best way to get a perfect exposure was to meter the white snow on manual, and add 2 1/3 stops of light.

When photographing closeups or landscapes or anything else that doesn’t move, I still prefer to meter manually with my thru-the-lens spotmeter, compensate for subject reflectance, and shoot a three frame exposure bracket using AEB which stand for autoexposure bracketing. By the way, AEB on my Canon cameras can still be used when shooting manually. All it does is automatically bracket the exposure I think is right by giving me a second exposure 1/3 stop darker and a third at 1/3 stop brighter. I have my cameras set for 1/3 stop intervals, the EOS 5D Mark II could be set for 1/2 stop increments, too. My 1D Mark III can be set for 1 stop increments as well. I find 1 stop increments to be too large for my needs (except in the case of shooting HDR sets of images) and 1/2 stop increments aren’t quite precise enough with the RGB histogram we use for obtaining superb exposures.

The nice thing about digital cameras is you know you are getting an image and you can check the histogram to make sure the exposure is what you are seeking. If it isn’t, then in most cases you can adjust the exposure. In the manual metering mode, just change the shutter speed, f/stop, or ISO a bit to change the exposure. In an automatic mode, use the exposure compensation button to adjust the exposure. Properly exposing images with digital capture is easier than slide film and we all welcome this change. But, great images demand much more than excellent exposure. The digital revolution has made exposure easier, so now we’ll all have more time to devote to the other key factors that make up a great photograph. These factors include developing excellent shooting habits to make sharp images, finding the most photogenic subjects, learning to be able to quickly see a unique situation, composition, using flash, camera viewpoint, and the effective use of light in terms of color, contrast, and direction. That’s a welcome change that digital has made possible. We’ll all have more time to spend on the creative side of image making.

Dec. 2009 Update

We provided the previous discussion to show you how we used to meter. Things have changed now! We have finally abandoned old exposure ideas for film and adopted leading edge new digital exposure tactics. Everything is so much easier now. Forget about spot-metering, compensation for subject reflectance, and the rest. Let’s keep it simple. Here is how we determine exposure in December of 2009.

We use multi-segmented metering patterns and manual exposure. Canon calls their multi-segmented metering pattern "Evaluative" while Nikon calls theirs "color matrix". We like the multi-segmented metering patterns over spot-metering because they tend to get you closer in more situations. With manual exposure, changing the ISO, aperture, or shutter speed will change the exposure. This is not true with autoexposure, since you have to use the exposure compensation control. We also use the RGB histogram in the camera. Most cameras offer this choice, but the default histogram is the luminance (called brightness by Canon) one. The luminance histogram averages the color channels together and then plots the brightness values on the histogram chart. It works fairly well, but any scene where one color predominates (such as a bright red sunset) can be a problem because the averaging histogram may not show any clipping, even though the red channel is clipped very badly.

Setting the RGB histogram, the display shows a separate histogram for each of the color channels. It is simple to use. Look at the histogram for each color channel and find the one that has data (however small) farthest to the right. Ignore the others. Now adjust the exposure until the rightmost data is snuggled up to the right edge of the histogram. That’s it in most cases, you are done! It takes us about one minute to teach our students how to master this technique. If you do have any specular highlights or detailess objects such as a bright sun, go ahead and let the pixels that measures the light from these areas clip a bit (climb the right edge of the histogram). By exposing in this way, you preserve highlight details while allowing the pixels that correspond to the dark portions of the image measure more photons, so the signal-to-noise ratio is better for capturing good color and detail.

This technique is even easier if your camera has a live histogram option. Of course, your camera must offer live view and let you display a live histogram. Then, you can adjust the exposure before shooting the image. Otherwise, we manually adjust the exposure level indicator you see in the camera’s viewfinder to line up at the zero position. (Do this by changing the ISO, shutter speed, aperture, or a combination of these.) Shoot the image and check the histogram. If the rightmost data is climbing the right edge of the histogram, subtract light until it is just touching the right edge. If the rightmost data doesn’t touch the right edge, add light until it does. In other words, forget about sunny 16 rules, hand-held meters, spot metering and compensating, and any other method. Just shoot the shot and check the histogram to see what you must do to properly expose the image. Use the histogram to guide you, use the histogram to guide you, use the...............well, you get the idea. When in doubt about exposure, shoot an image, consult the histogram, and now you know what to do. It’s easy!!!!

Please take the time to fully understand the following sentence!

The worst sin you can do to your digital image is to overexpose important highlights that should have detail! Once you overexpose them, there is no way you can fuss with the file and get them back.

Question! How does digital get overexposed anyway? How do you see detail? If you look at a fresh sheet of white copy paper, is there any detail? No. All parts of the paper have the same color and brightness, so no detail. If parts of the paper had different colors or different brightnesses, then you could see detail. Imagine photographing a white egret. Most of the feathers are white, but the feathers have slightly different brightnesses, so you can see detail. If you overexpose the egret enough, all of the photodiodes (usually called pixels) that make up the sensor that should have varied in brightness values, say from 200 to 255, might be pushed all the way up to the maximum brightness value at 255. If all of these feather areas that should have varied from 200 to 255 become 255, you can’t see the difference between them, so no detail. We all know we can adjust the exposure by changing the levels with software, but if you lower the maximum brightness with the levels slider in Photoshop to 230, you do darken the maximum brightness of the image, but all of these feather areas are now 230, they still look alike, so still no detail.