The histogram, whether it’s on the back of your digital camera, or in Adobe Lightroom, or Photoshop – wherever you find it is the great tool for learning how to make a reasonable exposure. Ansel Adams spent many chapters in his fine books explaining what can easily be seen in the digital histogram.
The camera’s light meter measures the brightness of an object, it has no way of knowing what that object is supposed to look like. The classic example is a snowy landscape. No matter how many exposure points the light meter uses – it simply doesn’t know (at least until there is a way to tell it) that it is photographing snow. And so typically the meter imagines that this white expanse is 18% gray and exposes accordingly. The result is a grayish looking snow.
In other words, if the meter told you to expose at a shutter speed of 1/125th of a second at f5.6, you could either change the f-stop to say f4.0 (1 stop of exposure); or 1/60th of a second and f4 (2 more stops of exposure).
In the old days, when there weren’t as many tools for adjusting the exposure or the print – it was typical to start with the meter reading and then over expose by at least one stop, but you might go as high as 3 stops of additional light. That is, if you wanted the final print to show white snow with some detail.
If you overexposed it too much, you would lose detail in the snow banks.
Another common example from film days – would be how to expose for a predominantly black scene. Again, the camera wants to turn that black subject gray, so you might under-expose by one or two stops. This was (and still is) called stopping down, or closing down. So if the meter read the black building which filled the frame at 1/125th and f5.6, you could underexpose it one stop by changing the f-stop to f8, or two stops to f11. You get the idea. Then you had a good chance that your image was going to appear black, though still with detail in the final print. The idea was to get enough information on the negative to produce an “expressive” print.
This whole exposure business is a lot more important than you’d think – because as a general rule it’s the unusually lit shots that are most interesting. The world may be filled with areas and subjects that are 18% gray – but as a rule – not going to be very interesting (of course there are exceptions.).
Negative film was apt to lose detail in the shadow areas, and if you didn’t expose for the shadows you simply would have clear film with no data in those areas and you just couldn’t add it back.
This is called clipping. When you lose data, whether with a digital camera or a film camera, at either end of the scale, and there is simply nothing there – you’ve clipped the image.
Before I go further – let me say that “clipping” is not necessarily a bad thing. Unless you are a studio photographer with control over the lighting – you will always be clipping specular lights in a shot or the sky. Digital cameras, as a rule, are more prone to clipping highlights than shadows, which makes sense since they are giving you positive images, not negatives.
When I first began shooting with digital cameras, they were especially poor at preserving the right side (highlights) of the exposure. That changed for me with the 40D which had something called highlight priority mode. Whether this problem has been solved with other digital cameras – I don’t know but suspect that it is an important consideration for professional photographers.
The histogram is just a graph. I like to think of the vertical lines as piles of pixels. The higher the vertical lines, the more pixels fall into that area.
The left side is pure black; the right side pure white; and the center is gray.
Clipping is a term that means you’ve lost some pixels, either because you over- or under-exposed some area of the image. Almost every image has some amount of clipping. A bright light (specular lighting) is almost certainly going to go beyond the range of the camera unless you are exposing for that light.
If you see a spike that is showing clipping. Lightroom and most other programs will have a mode where they can show clipping with colors or blinking. Cameras will do the same thing. In this shot, you can see that the light fixtures and the white table cloth have been clipped. That is they are over-exposed to a point where data is missing. And if you look towards the right side, you can see that most of the data is left of the center gray point. In other words, a fair amount of it is under-exposed. But not clipped. So the image has enough data to brighten it up with your image editing tool of choice, but it is unlikely that you can bring back detail into the lights. This sort of clipping is called specular lighting – and frankly you will run into it all the time since you are not going to (as a rule) expose for the lights in the background because that would completely silhouette the figure in the foreground.
And surely there are times when you want to silhouette the figures. Look at the second thumbnail.
Here’s an example where exposure was made of the sky and the figures in the foreground were allowed to go black without detail. You can see this nice big spike on the left side. That’s the players. Clipping is not always bad – it’s just something that you need to understand in order to make your exposure.
Example #3 Here’s an example of a technically correct histogram with just a touch of clipping on the right (highlight) side. If you have a histogram which is just a bell curve – you are going to have a very gray low-contrast image. Again – not bad – just important to understand how to read it. You can take an image like that and turn it into the most contrasty shot in the world if you want. It’s simply giving you an idea of whether you’ve blown any data that you might need.
Now, even if you’ve under-exposed a digital image a bit – you can do wonders with it with an imaging program like Lightroom.
Thumbnail of Dog shows the original with histogram, and then the “fixed up” shot with new histogram.
Assuming that you begin with a raw file and then process it in Lightroom or Photoshop, you can never add any more original pixels. You can add pixels during post-processing that were there in the original raw file; and you can use various calibration techniques to get to pixels in the file, but for the most part, post-processing involves changes to the way your original pixels are displayed. This brings us to the concept of destructive and non-destructive edits and that is for another post. But I bring it up because the histogram can show you when you’ve made so many alterations (even non-destructive) to the original file that you are producing gaps in the histogram. And that’s is literally what you’ll see, between the vertical lines, you’ll see a space where there was once a white line.