Question about understanding histograms

[Aleksandar]

First off, just want to say I'm new to photography and am trying to learn as much as possible. Now, I've read a lot of threads here where people often say "check the histogram" when someone has a question about exposure. Knowing nothing about those funny little graphs, I started reading links people offered. What I got out of it is that there's no such thing as as a "perfect" histogram since they're all unique to the picture they're representing. The only hard guidelines is not to have any "lines" hugging the left and/or right edge, which would signify under or overexposure, respectively.

That being said, look at this graph:

http://www.xcelluploader.com//files/21/Histogram.jpg

Knowing what I know, I can't see any signs of over or underexposure. Yeah, there's not much on the right side but that would just lead me to believe the picture was composed of more dark "objects" than lighter ones. So when I saw this when I was shooting earlier today, I thought I had a good exposure. When I got home though, the pic this graph represented was a different story altogether:

http://www.xcelluploader.com//files/21/Post/Peterson.jpg

Considering this was taken on a very sunny day, it looks completely underexposed to me. Granted, I wasn't in the best position since I was shooting into the sun but I have other pics from around the same time where I can see the players faces. And coincidentally, their histograms show a slight overexposure.

So does this mean that not only should you take into account the subjects you're shooting when evaluating histograms, but also the conditions they're shot in? For example, does the above mean that I should overexpose a bit in order to see players faces when shooting into the sun?

Also, this is probably a question best left for the Post-Processing forum, but is this pic salvageable? Being knew to photography, I know next to nothing about PS. Below are the high-res version of the pic if it helps:

RAW/CR2 Image (http://www.inat.org/files/Pictures/IMG_7441.CR2)
JPEG Image (http://www.inat.org/files/Pictures/IMG_7441.JPG)

[DDan]

Seeing that histogram, I would have increased exposure. I would do so until I started seeing signs of clipping on the right side. This would be a spike against the right edge. This would probably be accompanied by some blinkies in the thumbnail.

[atholg]

Not sure why but the histogram in CS looks more like the image than the one you have shown

[Glenn NK]

There are various difficulties with photographing some colours; white and black immediately come to mind.

You wouldn't be able to crank up the exposure much before there were signs of clipping on the player's right arm and foot.

His helmet creates a shadow, and a face with very dark skin is not going to yield any detail in this shadow unless you blow some highlights.

There is too much dynamic range in this image to capture everything.

[MissSage]

There is not much salvageable info in the picture as far as working on it in PS.. his face has almost no data to it as it is in shadow and he is african-american as well. You could lighten up the other areas and such but the data on his face is not good enough... to get his face visible you would blow out most of the rest of the picture.

[viet]

The histogram takes some time to get used to. Just make sure you don't have a lot of clippings & running offs to either sides.
Here's what I got from your raw file less than 2 minutes tweaking in Lightroom.

http://img112.imageshack.us/img112/6688/img7441tinymn2.jpg

Atholg, the difference might be because you are showing a RGB histogram and his is in grayscale. Although you can change it to show RGB in his Canon 30D, but I know a lot of people are more comfy with b&w histogram.

[Aleksandar]

The histogram takes some time to get used to. Just make sure you don't have a lot of clippings & running offs to either sides.
Here's what I got from your raw file less than 2 minutes tweaking in Lightroom.

Wow. You got that in just two minutes? I definitely need to check some books out.

Atholg, the difference might be because you are showing a RGB histogram and his is in grayscale. Although you can change it to show RGB in his Canon 30D, but I know a lot of people are more comfy with b&w histogram.Hmm. Didn't know that was customizable. I'll have to check that out.

[In2Photos]

The histogram takes some time to get used to. Just make sure you don't have a lot of clippings & running offs to either sides.
Here's what I got from your raw file less than 2 minutes tweaking in Lightroom.

http://img112.imageshack.us/img112/6688/img7441tinymn2.jpg

Atholg, the difference might be because you are showing a RGB histogram and his is in grayscale. Although you can change it to show RGB in his Canon 30D, but I know a lot of people are more comfy with b&w histogram.
After seeing this shot I wonder if the histogram in the first post is for this crop or for the crop in the original post.

[viet]

Wow. You got that in just two minutes? I definitely need to check some books out.

Actually, my laptop took about a minute to start up Lightroom and imported the picture, would have been a little faster with my PC at home.

It doesn't really take too long to tweak things in LR once you get used to the interface and you can generally tell if an image is salvagable or not at a glance.

LR is probably the best thing since slice bread for me because I process a tons of pictures after every wedding heh.

[StewartR]

Spot the difference.

195337

195338

[Aleksandar]

Spot the difference.

Yeah, huge difference.

Now I just got to figure out how to do the same thing myself. ;)

[StewartR]

Yeah, huge difference.

Now I just got to figure out how to do the same thing myself. ;)I think you might be missing the point.

In your original post, you showed the histogram from the entire picture. It had an attractive shape with data towards both ends, and you concluded that the exposure was probably OK.

However, most of the data from the right hand end of the histogram came from the sky, which you're not interested in. The part of the picture which you were interested in contributed mostly to the left hand end of the histogram. And when you looked only at that part of the picture, you saw it was underexposed.

The point is, I think, that when you look at the histogram you should also consider what's causing it to be the shape it is. To do this you have to think about what elements are in the picture and where they are in the histogram. For example in your original histogram I think the big spike on the right is the sky, the broader spike slightly left of centre is the grass (possibly a bit underexposed IMO) and the very broad area to the left is the dark blue shots and helmets, the players' skin, etc. And thinking about it that way suggests that the parts you care about are going to come out a bit on the dark side, and preserving the highlights on the right isn't necessarily that valuable.