A Histogram Question

fotogenik

TPF Noob!
Joined
Jan 16, 2006
Messages
345
Reaction score
1
Location
Louisville, Kentucky
Ok I am sitting here at work this evening reading through all the entries from blogs in my newsreader because I have nothing better to do for the next 10.5 hours and I come across this statement:

If your camera will display a histogram after shooting an image, then you need to take advantage of this feature. By looking at the histogram of an image you can quickly tell if the image was properly exposed or not. With a digital SLR camera you want the histogram to be a nice curve with the curve being just to the right of center

Is this true?

Take this image for example. (Not my image, just a random image I picked off this particular website)

The histogram for the image looks like this:

histogram.jpg


There is not curve slightly right of center in this histogram. Does this mean the image is not properly exposed?
 
In general, the quote you posted is correct. For example, for a normal day-lit scene, or for a portrait, there is not a large dynamic range involved. In those cases, most of the light you want to be in the middle range of your histogram, otherwise the scene is underexposed (if the histogram peaks to the left) or overexposed (if it peaks to the right).

The example you link to is not that normal kind of every-day shot. It appears as though the photographer did it in black and white and purposely made it high-contrast to emphasize the black vs. white nature of the chess game. For that image, it works. When it was originally photographed, the histogram probably didn't look like that, but rather more "normal," with a peak in the center, that was then stretched to give that high-contrast look.
 
It's a 100% correct statement to say the histogram can tell if an image is correctly exposed or not. But it does come down to personal preference as to what is correctly exposed. Say I took the exact same picture as the one you linked to and there was a lot in the middle then it would be incorrectly exposed. Clearly some light hit the front element and shot the contrast on the image, the histogram would show that.

But if I take a photo from a leaf if the histogram appears anything other than centre weighted for a neutral middle green type image then again it's not properly exposed.

Use the histogram as an indicator to tell you if you got your image right. If I'm taking a photo of a lake with a white boat on it the histogram can tell me straight away if the boat has clipped highlights or not. Or if I have underexposed the image, how much latitude I have to increase the exposure before something clips.

It's a statistical feedback tool, but you need to know how it applies to an image. Use it to tell you what you shot, not to tell you how to shoot.

When it was originally photographed, the histogram probably didn't look like that, but rather more "normal," with a peak in the center, that was then stretched to give that high-contrast look.

Unlikely given the lighting. The lighting itself looks very contrasty and if it's controlled you could make that with no post processing at all.
 
so...in all reality...the histogram is nowhere near a hard fast rule for determining whether you got a good exposure or not.


I personally shoot a lot of dark images with dark backgrounds and my histograms are typically skewed very far to the left.

Take this image for example (my work, but linked because it may be NSFW)

The histogram looks like this:

histogram2.jpg


Skewed very far to the left, but I consider it to be a great image. Is that wrong?
 
You can also throw the histogram out the window when dealing with 95% of available light shots at night - depending upon subject matter. It is a very useful tool, but depends upon the application (subject).
 
Somebody here posted an awesome website the other day on this subject. No, a histogram doesn't have to display a typical bell curve to be a correctly exposed shot. Here is the link again:

http://sekonic.com/images/files/HistogramsLightmetersWorkTogether.pdf

This is a great tutorial because it shows both the images and their accompanying histograms. I think Double H originally posted this link on the other thread, but don't remember for sure now.
 
so...in all reality...the histogram is nowhere near a hard fast rule for determining whether you got a good exposure or not.

Yes it is. The point I was trying to make is that the definition of good exposure changes depending on the picture. If you know what you're trying to do the histogram can give you statistical feedback to completely eliminate the guess work associated with shooting under different conditions and thus getting inconsistent results from looking at your camera's LCD. I hate the LCD on the D200 it looks nothing like the prints or the result on my monitor so on more important shots I always look at the histogram.
 

Most reactions

New Topics

Back
Top