UNDERSTANDING HISTOGRAMS

If you're new to photography, then pretty much everything associated with it is confusing as hell. With so much technical jargon and nomenclature to absorb, it's very easy to feel intimidated and quite overwhelmed. Don't panic! There are altruistic individuals like myself out there, that impart their knowledge for free, to help. I do this because hopefully karma will one day reward me with a big lottery win, although knowing my luck I'll be the wrong side of 70 and have a heart attack and die of the excitement and shock.

Anyway, one of the most common things newbies try to comprehend is the histogram - primarily because they've been told you can gauge accurate exposure from it. Personally, I would contest this, as there can be so much going on in a photograph that you cannot identify specifically where the skin tones are in a histogram, unless of course you literally fill the frame with someone's face. It's useful as a general guide to overall exposure for a scene but not necessarily individual elements. For that, spot-meter your subject's face if doing a portrait, or use a light meter that is profiled to your camera/lens combination.

It's also important to remember that understanding your histogram will do nothing to physically solve exposure issues. All it will do is tell you if your highlights are blown or your shadows are black. Most beginners are guilty of taking photographs outside on a sunny day - which of course is the most challenging high contrast lighting. Bright highlights, dark shadows. It's real easy for a beginner to screw up with either their exposure or the way they light their subject. 

It is because of this notion of gauging accurate exposure with the histogram that a lot of people get muddled up. They assume that what they should be be seeing is something resembling a Bell Curve, on their LCD screens. The reasoning for this is that correctly exposed things are generally in the middle between pure black and pure white. The truth is that a correctly exposed image can vary from one extreme to another, depending upon what the tonal value of the elements in the frame are. I'll explain more about that later. 

 

Firstly we should identify how the histogram works. It's actually pretty simple if you imagine it to be a bar chart made up of 256 separate columns. We all remember those at school and understand how they work. On the far left we have pure black, which is represented by the "0" in blue. This is where shadow clipping occurs and we are unable to see any detail beyond this point. On the right, the opposite is true. Here we have pure white denoted by "255" and is where we lose detail in the highlights. In the middle we have mid tones at 128. To be accurate, the numbers should be "0,0,0", "128, 128, 128" and "255, 255, 255" which is their RGB Hex Colour Code.

The more pixels there are at a certain tonal value, the higher the corresponding spike.

 

The diagram above is the histogram from the image next to it, which I have broken into 5 sections. Blacks (also referred to as shadows), darks, mid-tones, lights and whites (also known as highlights). Looking at the histogram we can determine that there are no overexposure issues in the corresponding photograph, as there are virtually none of the tones venture into the "whites" section. The majority of the tones are spread out between darks, mid tones and lights. There are also a fair amount of shadow areas too but there is still detail in these areas, as the spike isn't all the way to the left.

Below is another histogram with the corresponding image next to it (click to expand). To a lot of newcomers it would be blatantly apparent in their mind that the corresponding image is grossly underexposed. They would of course be wrong. The model in the photograph is absolutely correctly exposed against the black background. It's a high contrast image but, it's important to note that the big spike on the left and other dark tones relate to that black background and the shadows in the photo. If the histogram above was of a cornfield on a sunny day, with blue skies, then yes, we could deduce that the image was underexposed, because it wouldn't correlate to what we were meant to see.

Despite appearing underexposed, the resulting photograph is exposed correctly.

 

Below is an histogram example of the complete opposite. A model shot against a pure white backdrop. This accounts for the spike on the right side of the histogram and the relatively muted tones spread through the rest of the image. Notice though that the highlights on the subject's face are not blown to pure white and there is still detail in those areas. You will not be able to discern this information from the histogram on your camera as is. The only way to determine that would be to fill the frame with the subject's face and take a photo under the same lighting conditions. This will then give you an area specifc read out.

You will often see the histograms in their combined RGB or luminosity states, which the above examples are. However they can also be viewed in their component RGB channels, either individually or grouped. By changing the settings on the camera you can alter the way you view the histogram.

You may ask why this is important? Well, the reason being is skin tones. If you aren't using a light meter to get accurate exposures, or using an incorrect metering mode to get correct exposure of your subject's face, the histogram can help. 

What we are looking for is the red channel specifically. This is because skin tones live in the red channel and so to avoid overexposure and loss of detail in the skin, we want to avoid a spike on the far right side of the histogram.  A handy tip if you are unsure if your metering/exposure of the scene is correct.

 

In summation, the important thing to remember is that the histogram is a reflection of what the camera is seeing and can vary dramatically according to what you point your camera at and the lighting conditions. If your scene contains a lot of dark elements, expect a concentration of peaks on the left side of the image. If there are a lot of bright areas expect the opposite, with a spike on the right of the histogram. 

If there is very little in the way of highlights or shadows, expect the histogram to be grouped somewhere in the middle. 

The amount of the tones will determine the height of the spikes. It really is that simple. We only need to concern ourselves if the histogram doesn't translate to what we would expect. Then we know there are exposure issues - often caused by high contrast lighting, incorrect metering or using the camera in manual mode and selecting inappropriate aperture/shutter/ISO combinations. 

Is too much emphasis put upon the histogram by newcomers? For my mind, yes - for the simple reason it cannot tell you what specifically the exposure values of certain elements are. For exposure issues, you are better off relying on the flashing "zebras" setting on your camera, which will flash any areas or overexposure to let you know you have loss of detail. Use the histogram as a rough guide but don't fret too much about them. Review your images on the camera's LCD screen and so long as all elements look relatively good on there, you're probably in the ball park.