can anyone teach me ,or anybody else interested all there is to know about histograms? thank you
can anyone teach me ,or anybody else interested all there is to know about histograms? thank you
That's a big question. Here's what I understand so far:
1) The black & white histogram represents every tone in the photo. From black on the left to white on the right. The X-Axis
2) The vertical (Y-Axis) represents the number of pixels for each tone.
3) There is no such thing as the perfect histogram, it depends what you are trying to create.
4) But for general picture taking: A) A histogram that has a hump in the middle tones and trails off to either end is probably a good sign that you are on the right track. B) If the histogram extends past the extreme left or right sides, you will lose all detail in black and white, respectively. C) If the entire histogram is shifted towards the left or right, the entire image is trending underexposed or overexposed, respectively.
I hope this helps some.
yes it does, thank you steve
Glad I could help!
Originally Posted by Steve Eisenberg
Steve's explanation is an effective one. So with that as the baseline, let me add some more...
His second point is very important. You typically want to avoid bars from touching either end with the right side being more important to avoid than the left. A good healthly spread of the bars across the histogram indicates a lot of tonal quality in the shot, a narrow spread indicates less tonality. Either is ok depending on what you are shooting but the graph should visually represent the scene.
- If you're shooting a duck in the snow expect a narrow graph to the right of center. In this example, if the histogram was centered you're shot is underexpose because snow is white and the middle means gray. Hope that makes sense.
- If you're shooting a black bear in a cave, expect the histogram to be left of center. If it's in the middle your shot is overexposed because it made the black bear gray.
- If you're shooting a black bear chasing a deer, through the snowy woods on a beautiful blue sky day you want a wide histogram graph that peaks in the middle but doesn't touch either side. Why, the blue sky, the deer, the woods are all mid-tones and should represent most of the scene. Yet the bear is black (left side) and the snow is white (right side).
Three other quick points:
- The height of the bars mostly just indicates how much of that tonal value there is in the scene and not something I pay too much attention to.
- Many people will tell you to expose to the right. Which means, in an average scene, use enough exposure to prevent too much of a left sided histogram. The logic is simple, in post processing if you have a picture that is too dark overall or too dark in the shadows and you increase the brightness it will cause the noise signals in the shadows to show. On the other hand if you shot is a little overexposed (no blown highlights though) and you decrease the brightness you will have no ill effects from noise. So an slightly overexposed shot is more forgiving than an underexposed shot.
- The RGB (red, blue, green) channels each have their own histograms. I personally, don't use them much unless I'm photographing a highly saturated or difficult color like red and then it's just to ensure I've got wide enough tone in the color channel the the detail will show up in the final result.
In addition to Don's good comments:
Everyone should have a grey card that's easy to take with them. I like the Lastolite Ezybalance. It folds up, and is weatherproof. Meter off that, and you will be on your way to good exposures in sketchy situations (use the "Exposure Lock" button, the "asterisk"). It's also handy for White Balance. Shoot a picture with the card in the frame, and use it to adjust your white balance later. If I shoot a bunch of bugs, as soon as I'm done I'll toss the card into the frame, and shoot it for later use.
Another important point: if you shoot raw, your histogram is lying. It is based on the JPEG, which is often one, two, or even three stops over- or under-exposed relative to the true raw data. The reason for this is that the saturation, curves, space conversion, contrast, and especially white balance processing all affect the histogram, even though they don't affect the raw data.
Here's an example:
http://forums.dpreview.com/forums/read.asp?forum=1037&message=26905476
Until manufacturers add a raw histogram feature, the only workaround is to build a custom white balance file that reverses the effect of the in-camera processing to give you a *real* histogram. Some call that a "Uni-WB". The downside is that the metadata, preview, etc. are all useless (and very green!), so you can't check an image for color tones *and* histogram at the same time.
thank you very much guys for all the answers, from now on i'll check the histograms ,i've never tought they were important.
thank you
I like to understand more about this...
I think I understand the post processing white balance part as the card become the neutral gray target to balance against but I'm not sure how or why you are using the card relative to setting the exposure.
Thanks in advance for the help...
Adrian,
Below is a link to the Cambridge in Color website that has a couple ofvery good tutorials on histograms, how to read them,and how to use them to improve your images.
http://www.cambridgeincolour.com/tutorials/histograms1.htm
http://www.cambridgeincolour.com/tutorials/histograms2.htm
By the way this Cambridge in Colour website is an excellent resource.
Kyle
Posting Permissions
Reply With Quote