The histogram is one of the most important tools we have on our digital cameras and in our processing software. A histogram is a bar graph of the intensity of the light absorbed by the sensor in a digital image file. Like any bar graph, it can be read both vertically and horizontally. The horizontal axis spans the range of brightness from pure black on the left to pure white on the right. The height of the bars indicates how many pixels on the sensor were recorded at that brightness level. Also, note that unlike normal bar graphs that have spaces between the each bar segment, histograms do not have these spaces. All of the spaces have been removed and that is what makes them hard to understand at first.
In this sample histogram, the different areas of the histogram are labeled according to the tones related to those areas. The horizontal axis spans the range of brightness from black (shadows) on the left to white (highlights) on the right. The height of the bars indicates how many pixels on the sensor were recorded at each brightness level. Below the histogram shown here are the range of black and white tones from black to white. We will come back to this illustration later in the blog post and discuss it further.
For the uninitiated, learning to read histograms can seem like a daunting task. In reality, they are quite simple to understand. It just takes some time to understand what a histogram is telling you and how it describes an image. By looking at a lot of images and their corresponding histograms you can start to understand why they look the way they do and which parts of the image correspond to the different parts of the histogram. We’ll discuss the histogram and how it can be used to help adjust our exposure while shooting, as well as how it can inform us about the tones in our image while processing them in Lightroom. While I don’t go into the entire post-processing workflow in the Develop module, we’ll discuss the histogram and how it can be manipulated in the Lightroom Develop module in upcoming blog posts.
Technically, digital camera sensors only record the intensity of the light that hits them in black and white. These black and white tones are converted to color images via a grid of red, green, and blue (RGB) tiles that are layered over the sensor and also by using an extremely complex algorithm that converts the image from black and white to color. Because digital cameras convert these tones to RGB values, there are three different histograms for every image – one for red, one for green, and one for blue. There is also a fourth histogram that shows up on the back of the camera, which is the averaged histogram. This averaged histogram is, as you might suspect, an average of the RGB histograms. Most DSLRs can show the averaged histogram and all of the RGB histograms. I have my cameras set up so that these histograms are visible on the back of my camera when I view the images. And I use these histograms to tell me if my exposure is off or needs to be adjusted.
Here is an example of an image and it’s corresponding histogram. Note that this is the final worked up image and the histogram for that final worked up image.
Digital cameras record light differently than a piece of film does. Digital cameras are linear devices, whereas film and our eyes are not—they are non-linear. What I mean by non-linear here is that with film there was some amount of roll off on both the highlights and the shadows side of the exposure, which in turn made it more difficult to completely blow out the highlights when shooting film. In effect, with film, and similarly with our eyes, the highlights were compressed so that the transitions from the bright to the extremely bright areas of the image were rendered with a smooth transition. Our eyes work in a similar manner. Our eyes compress both the highlights and shadows in a scene so that we can see detail in a very wide range of dark and light tones. We are able to see approximately 18 stops of light with our eyes, whereas the best digital camera sensors can see only 13 or 14 stops of light. The difference is that a digital camera is seeing those 13 or 14 stops of light in a linear fashion—with no compression of tones in the highlights or shadows. A camera’s sensor is just recording the number and intensity (amplitude and wavelength) of the photons that hit the sensor. With its inherent linearity, once the brightness level is beyond that which the sensor can record the chip just records it as white. This is why you have to watch your histogram and your exposure so that you don’t "blow out" the highlights when shooting.
The reality is some highlights are going to blow out—it is just a matter of if they are important or not. The other fact about digital and the linear relationship is that most of the information is recorded in the first two or three brightest stops of the sensor’s levels. For example, if you are shooting in raw mode with a 12-bit camera that means you have 4,096-segmented values of brightness per color channel. Hence 2,048 of those values are in the brightest part of the histogram—e.g. half of the histogram spanning from the midtones to the brightest highlights in the image (see the graph below). All this is to say that if you underexpose the image you are throwing away massive amounts of data from your camera’s sensor. And when you adjust the exposure of the raw image file in the raw processing software to make the image brighter you’ll start to see a lot of noise show up in the shadows. This is one of the major reasons why I recommend shooting in 14-bit mode, if your camera allows it, as the final image will have much more information recorded in the shadows than it would if the image was shot in 12-bit mode.
Coming back to this histogram diagram we looked at before, if we are working with a 12-bit camera, with 4,096 levels, and which has a 6-stop dynamic range, then half of the information recorded by the sensor—a total of 2,048 levels—is captured in the brightest half of the histogram. Only 1,024 levels are captured in the next stop of light, and so on. As you can see only 64 of the 4,096 levels are captured in the darkest stop in this example. This means that only 1.5% of the data is captured in the darkest areas of the image. This graphic representation shows very clearly that with a linear device most of the information is captured between the midtones and highlights in the histogram. Hence, if you underexpose an image you are losing a lot of critical image data that could be captured by pushing the histogram to the right as described in the text. This goes a long ways to explain the "Expose to the Right" (ETTR) philosophy.
The tendency when you first start to shoot digital is to underexpose to preserve highlights. You should avoid underexposing because of all the reasons we have laid out above and also because highlights can be recovered if you are shooting in raw and process your images with Adobe Lightroom, Adobe Camera Raw (ACR), or any of the top raw processing software options. If you are shooting in raw mode, overexposing the image or metering as your camera suggests on your subject will serve you much better. This concept is known as "Exposing to the Right” (ETTR), meaning that you shift the histogram as far to the right as possible without blowing out important highlights in the image. This is counterintuitive, and trust me, it will take some time to get comfortable with. When using this technique, your raw images in Lightroom or ACR will look overexposed and washed out. In other words it will look like you really messed up, but that is not the case because the contrast, saturation, and brightness can be corrected in the raw processing software for optimum image quality. This discussion is why you really want to shoot in raw mode and why you need to understand the entire digital workflow process while you are out shooting to obtain the best quality final image.
To reiterate this important concept, two keys to getting a good exposure are remembering that you can recover up to two stops of highlight information using Adobe Lightroom, Adobe Camera Raw, Apple’s Aperture, or Capture One software, and that you can control the brightness in the post-processing and bring it back down to where it should be when post-processing the raw image file. Please understand that this ETTR method of exposing for your image only applies to RAW image capture. If you are shooting JPEGs, do not use this method, as you will end up with a bunch of overexposed images that can’t be corrected.
Below are some examples of histograms and the images they represent.
These are some sample histograms (from Lightroom) and the images they describe. None of these histograms are wrong as they represent each image and the tones recorded in those images. The left image has clipped highlights, which is accurate given that it was shot on a white background and the background is intended to be purely white. The middle image preserves both highlights and shadows with no clipping on either side of the histogram. And finally, the right image has some clipped shadows, which is correct for this image since it was shot right before nightfall. [Please note that these histograms include the averaged histogram, which is the grey area, as well as the RGB histograms.]
My shooting workflow in the field for getting the exposure dialed in using the histogram is as follows:
- Take a photo of the subject at the camera’s recommended exposure setting.
- Preview the image and histogram—at this point I am not looking at the image—I am just looking at the histogram.
- Looking at the histogram I can tell if I need to adjust my exposure (using either the shutter speed or the aperture). If the histogram is mashed up on the left then I open up and overexpose to push the histogram to the right; conversely if I have seriously blown out important highlights I stop down (i.e. underexpose), take another shot and check the histogram again to make sure I have the optimum exposure. Once I get the exposure dialed in, I then start shooting in earnest.
In terms of using the histogram during post-processing, everything we do in our post-processing is basically adjusting the histogram of the image. When we move the sliders around in the Develop module we are essentially tweaking the histogram. Hence, knowing how to read the histogram and understand what effect you are having on the tones of the image is critical. I always have the histogram visible when I work on images so that I can tell if I am crushing the blacks or blowing out highlights in the image. Depending on the effect I am going for when I process the raw image file I might want to blow out the highlights to create pure white areas in the image or crush the blacks so there is solid black parts of my image. Having the histogram visible and understanding it tells me exactly what is going on in the image and allows me to have more control over the final output. Hence, this is why understanding the histogram is useful not only for adjusting your exposure while shooting, but also when processing your images on the back end.
This blog post is a modified excerpt from my e-book, Adobe Photoshop Lightroom: A Professional Photographer’s Workflow. For more information on this e-book or to purchase the e-book please visit my website.