Understanding Histrograms

We've spoken a lot about exposure over the course of last year, and another element of this is knowing how to read a histrogram to see whether your image is correctly exposed.

A lot of people can be intimidated by histrograms, but they actually aren't that complicated. Think of a histogram as just a simple bar graph (although admittedly with an awful lot of bars) showing how the light and dark tones across the range of your image are distributed. The extreme left hand of the histrogram represents all the completely black tones in your image (at 0) , and the extreme right hand side shows all the pure white (at 255)  Everything else falls in between with a value from 1 to 254.

We can use this information to tell us whether the image is correctly exposed, and whether we are losing any details in the image.  In a histogram, the edges are by far the most important part. If you see a spike at 0 or at 255 (pure black or pure white) then you have lost information in your image. No amount of post processing can recover this information, as if there was no detail recorded, you simply can't bring it back.

It's probably easiest to show you some different images and read their histrograms to show you how this detail is recorded, and how this relates to your image. Here is the first image which looks to be pretty well exposed. If you look at the histrogram for this image you can see that the range of tones looks like a hill, rising gently in from the left to a peak in the middle then tailing off gently on the right.  There is a narrow spike on the right hand side at 255 - this represents the clipped highlights in the image, meaning all detail there is lost.  I know by looking at this image that this spike is caused by the light coming through the trees in the background (you can see the areas that are clipped as they just show as pure white on the screen) 

Here is the same picture but this time I have manipulated it in Photoshop to make it under exposed. You can see now that the histrogram is heavy on the left hand side, showing that this particular image is underexposed.  There is no gentle tapering off at the end - it's abruptly cut off on the left and there is a spike at 0. (You see that although the image is now under-exposed I still haven't got any detail in the white area on the right - as it was clipped when I took the photo it has gone forever!)

Same image, now over-exposed, and you can see that the histrogram now has a peak at the right hand side, showing that the image is indeed over-exposed with most of the tones in the white or nearly white range. There is no gentle roll-off to the right like we see in the properly exposed image.

Those were a relatively easy to read as there was a good range of tones within the image. However, the distribution of the histrogram also depends on the tonal range of the subject matter. If you have an image where most of the tones are in the shadows (a dark image) then you would expect to see your histrogram lean to the left. Similarly, if you have a image where most of your tones are in the highlights range, you will see it slope to the right. This is why it is important to be able to assess each particular  image to see if it should fall into these ranges - there is no such thing as a perfect histogram, it depends on the image.  I'm only going to show one example - white flowers on white background. This is an image where all of the tones fall into the highlight range and this is represented on the histrogram.

As all the tones fall into the white area the histogram has all it's graph on the right.  Because the scene is all white, this is correct. Watch the difference in the histrogram if it were over-exposed - the range is now climbing the wall on the right.

So when do you look at a histogram and how do you use it?

If you are out and about taking pictures and are worried you might be clipping some highlights, you can take a quick check at the histogram in your camera to be sure.  This is more accurate than looking back at the image in the LCD screen, particularly if you are in sunlight where it is more difficult to see.  If you see you have spikes over at the 255 range, you can reduce your exposure. (If you do this and your scene then looks under-exposed, there is no correct exposure as the range of tones in your image is too great. As an example, look again at the very first image in this post. If I had tried to get rid of the clipping in the background it's possible my subject would have been under-exposed because I could not expose for both the highlights in the background and my subject)

I must say I rarely do this (no time with a three year old, I'm lucky if I get enough time to press the shutter) and I usually let my eyes be the guide. However, if I were setting up some special shots (or had paying clients and needed to make sure it was right on the day) and wasn't sure, then I would.

When I am most likely to use a histogram is during editing.  I always evaluate exposure by eye, but check this with the histogram.  This is because sometimes when editing you can have a tendency to over-expose your image as this can make your subject's skin smoother and brighter.  Plus if you are layering a lot of edits, these can have a cumulative affect on how bright your image becomes. So, to make sure that I am not going overboard, I keep one eye on my histogram.

Think of your histogram as simply a guide to the tones in your image.  You will be able to determine whether most scenes are correctly exposed without using a histogram (I bet you didn't need the histogram to show you that the above images were over or under exposed!) but it can be handy to know how to read one for times when getting the correct exposure is a touch more difficult, or during editing.

Yes, I know, not the most exciting of topics, but you never know when this information might come in handy!


estelle doidge said...

Thanks so much for taking the time to explain this so clearly. I'm a beginner desperately trying to learn & find exposure can be a guessing game so this is so valuable to me & definitely something I will pay attention to :)

rsorey said...

Thanks so much for this!! I have always wondered how to read the histogram and this makes so much sense!

Audrey said...

Glad you both liked it!

Gayatri said...

So well explained.. Thanks a lot. I have just begun to try to figure out how to use my camera correctly to get sharp pictures..Your blog is very helpful. Thanks.

Post a Comment