Histogram Use Advise Needed

I'm new to digital cameras and like to compose and optimize exposures in the camera as opposed to post-process manipulation. Can anyone expain in laymans terms how I can use the histogram function of my S404 to optimize "as taken" photos, for instance by EV compensation, adjusting contrast settings, or using flash fill. I'm just looking for basic rules of thumb (if their are any). Anyone out their smart enough to reduce such a complex subject down to a paragraph or two?

There's been a few threads on this try a search on histogram.

Jack Dahlheimer said:

I'm new to digital cameras and like to compose and optimize
exposures in the camera as opposed to post-process manipulation.
Can anyone expain in laymans terms how I can use the histogram
function of my S404 to optimize "as taken" photos, for instance by
EV compensation, adjusting contrast settings, or using flash fill.
I'm just looking for basic rules of thumb (if their are any).
Anyone out their smart enough to reduce such a complex subject down
to a paragraph or two?

Click to expand...

Geraint Morgan said:

There's been a few threads on this try a search on histogram.

Click to expand...

wgm-cp4500 said:

Jack replied:

Click to expand...

Thanks, I sometimes do forget to run a search, but in this case I did. Their was nothing of substance, or that people agreed on regarding using histograms to optimize in-camera exposures. Perhaps (1) nobody else wanted to reveal they really didn't know how to use histograms, (2) it's a "gee-wiz, look at that!" feature like an altimeter, rate of climb, turn and bank, or air speed indicator in an automobile [interesting but of little use] or (3) nobody has mastered the subject sufficiently to provide an easy expanation how to use it.

wgm-cp4500 said:

Jack originally said:

I'm new to digital cameras and like to compose and optimize
exposures in the camera as opposed to post-process manipulation.
Can anyone expain in laymans terms how I can use the histogram
function of my S404 to optimize "as taken" photos, for instance by
EV compensation, adjusting contrast settings, or using flash fill.
I'm just looking for basic rules of thumb (if their are any).
Anyone out their smart enough to reduce such a complex subject down
to a paragraph or two?

Click to expand...

Click to expand...

--
new s404 owner

I thought it best to check ... because nobody had answered you and my response would bring it back to the top of the list.

I haven't got one on my S304 ... looking from the perspective of the Photo Editors ... and using it for levels ... I'd use it to check the shape of the histogram, i.e. is it off to one side and are there lots of zeros.

Geraint Morgan said:

I thought it best to check ... because nobody had answered you and
my response would bring it back to the top of the list.

I haven't got one on my S304 ... looking from the perspective of
the Photo Editors ... and using it for levels ... I'd use it to
check the shape of the histogram, i.e. is it off to one side and
are there lots of zeros.

Click to expand...

--
Jack wrote,

Thanks again. As an engineer I'm familiar with using a histogram to, for instance, analyze a Weibul curve for regularity or "center" a manufacturing process, but I suspect the one in my camera is really just showing amount and distribution of black to white objects (and gradients between) captured in a given exposure. A picture of a black rock sticking out of a snow covered field on a sunny day would have two "bumps" (how high depending on size of rock versus amount of snow included), one at either end of the graph. But of what value is that? It doesn't provide any information I already didn't know simply by looking through my view-finder.

It is important where those two bumps are in the histogram. If the black is pure black you want that bump bunched on the left edge and if the white is pure white you want it on the right. Since you know basically how it works it will take you just a very few minutes to relate it to images.

I like this explanation. It relates to scanning but the principle is the same: http://www.scantips.com/simple1b.html

Here are some simple rules I found in a PSP tutorial:

“If the graph is spread out, the image has a balanced composition and enough detail to make corrections to it.

If the lines are compressed into a narrow range you probably need to rescan the image or take another picture.

If the grayscale graph is mostly on the left, the image is too dark and you will need to increase the image's lightness.

If the grayscale graph is mostly on the right, your image is too ligt and you will need to decrease the image's lightness or make it darker

If the grayscale lines are not spread out enough, you may need to stretch the Histogram or increase the contrast”

Use them as a guide without understanding what you are doing and you are likely to ruin more photos than you help.

The thing to look for in the histogram is saturation at either the white or black grayscale levels. For example, say you take a picture of a white flower against a dark, shadowy background. The camera's auto-exposure system might make the white flower a pure white blob, losing all the subtle shades of white in the flower. This would be evident as a a sharp spike at the far right end of the histogram. The histogram will actually be truncated at the right side. In this example, a little negative exposure compensation would shift the histogram to the left, assigning more grayscale levels to those previously-truncated near-white values, and allowing you to actually capture the subtle shades of white in the flower.

In general the goal is to use in-camera adjustments like exposure compensation to spread out the histogram across the available range of grayscale values without truncating either end of the curve. But for most of the pictures I take, truncation at the right side (white) is more likely to result in a disappointing photo.

....Wolf

Follow on from slipe's comments...

Try taking some pictures in Multi-mode manual exposure and make them obviously under or over exposed.

Jack Dahlheimer said:

Jack wrote,
Thanks again. As an engineer I'm familiar with using a histogram
to, for instance, analyze a Weibul curve for regularity or "center"
a manufacturing process, but I suspect the one in my camera is
really just showing amount and distribution of black to white
objects (and gradients between) captured in a given exposure. A
picture of a black rock sticking out of a snow covered field on a
sunny day would have two "bumps" (how high depending on size of
rock versus amount of snow included), one at either end of the
graph. But of what value is that? It doesn't provide any
information I already didn't know simply by looking through my
view-finder.

Click to expand...

Wolf1 said:

The thing to look for in the histogram is saturation at either the
white or black grayscale levels. For example, say you take a
picture of a white flower against a dark, shadowy background. The
camera's auto-exposure system might make the white flower a pure
white blob, losing all the subtle shades of white in the flower.
This would be evident as a a sharp spike at the far right end of
the histogram. The histogram will actually be truncated at the
right side. In this example, a little negative exposure
compensation would shift the histogram to the left, assigning more
grayscale levels to those previously-truncated near-white values,
and allowing you to actually capture the subtle shades of white in
the flower.

In general the goal is to use in-camera adjustments like exposure
compensation to spread out the histogram across the available range
of grayscale values without truncating either end of the curve.
But for most of the pictures I take, truncation at the right side
(white) is more likely to result in a disappointing photo.

....Wolf

Click to expand...

--
Jack wrote:

Thank you one and all. Wolf's answer especially to me seems to make a lot of sense. I can see where the histogram might provide more insight than simply looking at the image just shot on the camera's little 1.8 inch LCD. screen.

....Jack

Since you are somewhat familiar with histograms, this should be easy :

The graph counts how many pixel have a luminance of 0, then how many are a 1, then a 2... up to 255. The dark pixels are at the left and the clear ones at the right.

From there you must use your brain.

If you were to take a shot of a medium gray card, the perfect histogram should look like a high peak at the value 127 and 0 everywhere else : this is never going to happen since we live in a world of many shades. Nevertheless, something would be very wrong if the histogram was flat with uniform values for each luminance level, as you can imagine. There should be a sharp peak in the middle and almost nothing at a relatively short distance on each side of that peak.

If the card was pure white, there should be a peak at the far right and if the card was pure black there should be a peak at the far left. One may want to avoir the values 0 and 255 as they are meaningless to the eyes, except in very very special and rare cases.

Let's imagine a practical circumstance : if you were to photograph some light clouds in the sky there wouldn't be any chance that you'd get many dark pixels. Using a good technique with a polarizing filter you may get a fair number of medium luminance though. On the other hand, if you get a whole lot of pixel at value 255, there is every chance that all detail in the clouds will be lost into overexposition.

If you were to shoot in a dense forest where no direct sun penetrates, a large number of pixel with the value 0 would mean that a lot of shawdow detail is lost in under exposition. On the other hand it would be quite normal to have no pixel in the 200 and over.

As a rule, I avoid completely to have pixel that are 0, and rarely want many that are close to 0, although some clair-obscur effects may require pixel that are very dark. In the same manner I avoid pixel with a value of 255 and rarely want any that are close to that. To achieve these goals I play with the exposition and the contrast somewhat.

As you can see, the ideal histogram may vary a lot according to the subject : on thing for sure in almost every case is that pixels with extreme values means either underexposition or overexposition. If you get both there simply is more dynamic in the shot than the CCD can take : use an ND filter to compress that. Or take an underexposed and another overexposed shot of the same subject and combine them in Photoshop as to retain the highlights details as well as the information in the shawdows.

Long explanation, longer time to get the necessary experience but fun time too.

Jean

One other thing, concentrate on the picture comp, not the histogram. The histogram is a nice tool to have, but don't spend too much time with it. You can do some shifting in the photo tools after wards. basically, if it's all bunched up on one end, the tools can spread it out some. It's still not as good as having it right the first time, but just don't concentrate "too" much on it.

BTW to you 7i gurus, just what is the 7i histogram showing you??

Does it show just lumonosity or does it combine the RGB values as well? Can you select? Maybe I should go back and check Phils review again?
jack

Jack Dahlheimer said:

I'm new to digital cameras and like to compose and optimize
exposures in the camera as opposed to post-process manipulation.
Can anyone expain in laymans terms how I can use the histogram
function of my S404 to optimize "as taken" photos, for instance by
EV compensation, adjusting contrast settings, or using flash fill.
I'm just looking for basic rules of thumb (if their are any).
Anyone out their smart enough to reduce such a complex subject down
to a paragraph or two?

Click to expand...

--
I just want a camera that works.

Jack Steinhilper said:

BTW to you 7i gurus, just what is the 7i histogram showing you??
Does it show just lumonosity or does it combine the RGB values as
well? Can you select? Maybe I should go back and check Phils
review again?
jack

Click to expand...

The real time histogram is a small translucent box in the lower center of the viewfinder. There are no markings or modes. You can either turn it on or off and that consists of your only input. It just gives luminosity. It isn’t intended to be a complex tool. It just gives you a rough idea of what you are doing with the exposure, but I find it very useful.

The D7i manual is online now at Minolta if you would like to take a look.

The easy way to understand your photos and the histogram is to take the same photo and shoot it five time 2 stop underexposed 1 stop underexposede 1 right on 1 stop overexposed 2 stops overexposed. Then look at the photos and the histogram. You can see right away the differance in real photos and historgrams.

Ron

Jack Dahlheimer said:

I'm new to digital cameras and like to compose and optimize
exposures in the camera as opposed to post-process manipulation.
Can anyone expain in laymans terms how I can use the histogram
function of my S404 to optimize "as taken" photos, for instance by
EV compensation, adjusting contrast settings, or using flash fill.
I'm just looking for basic rules of thumb (if their are any).
Anyone out their smart enough to reduce such a complex subject down
to a paragraph or two?

Click to expand...

Ronald Gadrim said:

Ron

Jack Dahlheimer originally said:

I'm new to digital cameras and like to compose and optimize
exposures in the camera as opposed to post-process manipulation.
Can anyone expain in laymans terms how I can use the histogram
function of my S404 to optimize "as taken" photos, for instance by
EV compensation, adjusting contrast settings, or using flash fill.
I'm just looking for basic rules of thumb (if their are any).
Anyone out their smart enough to reduce such a complex subject down
to a paragraph or two?

Click to expand...

Click to expand...

-Thanks to everone who helped educate me on the use of the histogram feature on the Minota S404. I now appreciate it is quite a helpfull little tool.