spacer spacer
spacer

Tangents

using the histogram to determine exposure

July 31, 2007

Histograms display the relative levels of the darker to brighter tones. As the histogram stands, it isn’t of much direct use to us, since the tonality of the scene that was captured will dictate what the histogram shows us .. without a direct indication of whether exposure is correct.

Some will say that a histogram should have an even bell-shaped curve, but this is too simplistic.   A light toned subject against a white wall will show a much different histogram that a dark toned subject against a dark wall .. even though the exposure might be correct in both instances.

In both those cases, the actual histogram display might be interesting to look at, but of no real direct use to us. But, here’s how I use the histogram to determine correct exposure …

For all that the histogram shows us, there is only one thing that is of direct consequence to us .. where the brightest relevant tone lies on the histogram. And I have to stress the word ‘relevant’.

It is no use looking at the histogram to determine exposure if there are bright patches of sky or highly reflective surfaces .. for this will skew the histogram display in making it appear like we are over-exposing, when it fact, we might very well have correct exposure.

With weddings (and portraits), the brightest relevant tone is most often the bride’s dress or groom’s shirt cuffs. So what I do is point the camera to an area that only contains the white dress / shirt, and no other bright areas. Then the most right-hand point on the histogram will be the white area. And then we can place that white tone correctly on the histogram .. and all the other tones will fall into place, whether skin tones or clothes or surroundings.

There is another way to approach this, and that is to spot-meter off only the relevant white area, and place it around 1.3 stops over the zero mark on your camera’s meter display. You don’t want to zero the meter reading then, since we don’t want the white to appear as grey. We want the white to appear as white .. and that will be around 1.3 stops to 1.7 stops over the zero mark. You will have to figure out the specific value for your specific camera, since there are some differences between the camera makes.

But back to using the histogram …

Here is what the Canon’s histogram will look like.

 

In my experience I have found I will get the best exposure if the brightest relevant area has the edge of the histogram appear half-way through the middle of the 5th block on the display, or 1/3rd from the edge of that display.  It would seem that this varies a little between the various Canon D-SLR models, so it might be necessary for anyone who wants to use this method of calculating correct exposure, to interpet and apply this idea to the specific camera being used.

If I start over-exposing, then that kind of spike will start to appear.
With Canon D-SLRs (that I have experience of), the blinking highlights display won’t blink yet, but the image will appear too bright.

.
.
The Nikon histogram looks slightly different:
In my experience, I get an optimally exposed image, if I have the edge of the histogram just barely not touching the corner of the histogram display … for the brightest relevant tone. (That term again. It’s an important distinction.)

.
Here is the same image, but 2/3rd stop over-exposed. You can see the spike on the right-hand side. It looks different than the spike on the Canon histogram.

.
.
From the above images, it should be obvious how I use the histogram to achieve correct exposure … I take a close-up view of the relevant brightest tone (some white area) .. and I place this at a very specific point on the histogram.

I lock this exposure value by using manual exposure mode, and all my images taken under the same lighting *will* be correct.

I’ve found this method to be consistent to the extent that I don’t use my flashmeter any more when using studio type strobes.  (I use the Quantum flashguns.)

       Comments (40)

 

 

 

 

 

 

All Rights Reserved © 2006-09 | Client Login