How do I read a histogram?

I'm struggling with exposure. You might ask why aren't you looking at your pictures when you take them? Well I am! They seem to be good in my playback, but on computer often wrong, even when the camera sets a "correct" exposure. So I feel like I need to learn how to read a histogram. Does anyone know, or have any good links for me? Thanks!!!

You might start with the five threads at the bottom of this page having nearly identical titles.

or type 'understanding histogram' in googles seach window - some good sites with good tutorials.

http://www.luminous-landscape.com …standing-histograms.shtml

Federkeil wrote in post #4702339
I'm struggling with exposure. You might ask why aren't you looking at your pictures when you take them? Well I am! They seem to be good in my playback, but on computer often wrong, even when the camera sets a "correct" exposure. So I feel like I need to learn how to read a histogram. Does anyone know, or have any good links for me? Thanks!!!

I don't find that histograms help me much. Histogram assumes that your scene is "normal", and since you'll never get such scene in real life, relying on histograms for your exposures will not give you desired photos.

It sounds as though you need to understand camera's built-in light meter and what determines proper exposures first. Google "Understanding Exposure" by Bryan Paterson - a great book. Good luck!

bbbig wrote in post #4703955
I don't find that histograms help me much. Histogram assumes that your scene is "normal", and since you'll never get such scene in real life, relying on histograms for your exposures will not give you desired photos.

I only use the meter for a starting point and a rough one at that. I use the histogram for fine tuning.???

If you understand what the histogram is telling you (and there are many good tutorials online which will help you with that), it can be a great help with the exposure.

DDan wrote in post #4704036
I only use the meter for a starting point and a rough one at that. I use the histogram for fine tuning.???

For example, if you're shooting a subject in a big snow field, unless you calibrate your camera to a proper white balance first, then your image will come out underexposed. The histogram in that case, may show a pretty high singular peak in the middle region. However in "reality" the exposure should've been taken so that the histogram shows the high peak towards extreme right.

If you were "correcting" your exposures just based on what the histogram tells you, in above example instance, you will get severe under-exposure. Opposite is true when shooting against a dark background, etc.

This is why histogram really isn't very useful, other than to show the distribution of bright pixels once the photo is taken, properly exposed or not.
See what I mean?

I think the histogram provides good information but I dont think its something that you should follow religiously. A good picture may not (and often does not) have a perfect histogram.

This one helped me...

http://www.sekonic.com …ghtmetersWorkTo​gether.pdf

bbbig wrote in post #4708558
For example, if you're shooting a subject in a big snow field, unless you calibrate your camera to a proper white balance first, then your image will come out underexposed. The histogram in that case, may show a pretty high singular peak in the middle region. However in "reality" the exposure should've been taken so that the histogram shows the high peak towards extreme right.

If you were "correcting" your exposures just based on what the histogram tells you, in above example instance, you will get severe under-exposure. Opposite is true when shooting against a dark background, etc.

This is why histogram really isn't very useful, other than to show the distribution of bright pixels once the photo is taken, properly exposed or not.
See what I mean?

I don't think you see what I mean. I didn't explain very well though. I try to "expose to the right". I keep turning up the exposure until I see pixels piling up against the right hand side. This usually coincides with the blinkies showing up somewhere in the preview indicating blown highlights. If the blown highlights are in acceptable positions, that is OK. I then get the exposure correct during PP.

AdamLewis wrote in post #4708654
I think the histogram provides good information but I dont think its something that you should follow religiously. A good picture may not (and often does not) have a perfect histogram.

I do not care about the shape of the histogram. I just try get the meaningful data as far to the right as I can without blowing highlights as explained in the post above this one.

DDan wrote in post #4708814
I don't think you see what I mean. I didn't explain very well though. I try to "expose to the right". I keep turning up the exposure until I see pixels piling up against the right hand side. This usually coincides with the blinkies showing up somewhere in the preview indicating blown highlights. If the blown highlights are in acceptable positions, that is OK. I then get the exposure correct during PP.

Ok I hear you then - yes, historgram can be helpful in deciding whether your highlights are overblown. I thought you were talking about the shape of all things - my mistake.

A histogram is a graph that goes from black on the left to white on the right, with 18% grey in the middle.
It shows only how many pixels are at each tonality point along its length.
By looking at the peeks you will get an idea whether any aspect of the image is over or under exposed, but only that.
It is possible to take several images of the same subject illuminated in several different ways and get practically the same histogram for each. Have a look at this link and you will see what I mean.

http://www.sekonic.com …ghtmetersWorkTo​gether.pdf

kato1 wrote in post #4710094
By looking at the peeks you will get an idea whether any aspect of the image is over or under exposed, but only that.

Not really. The peaks just show how many pixels are at that level. Overexposed is on the far right and underexposed is on the far left.