Learn to read the histogram...

[martin_patek_strutsky]

I am a confused newbie in digital photography. Almost everywhere you

can read the mantra that you need to learn to "read the histogram"

for taking optimal pictures.

 

But no one seems to explain what this means in practical terms. How

does the histogram of a well exposed pic look like? What kind of

curve indicates that adjustments are necessary?

 

Thanks a lot

[jespdj]

<a href="http://www.luminous-landscape.com/tutorials/understanding-series/understanding-histograms.shtml">Understanding Histograms</a>

[matteo-delgrosso]

I am just reading Ron Reznicks e-book and from what he says, there is no specific curve shape that makes an image correctly exposed. Every image has of course a different and unique sort of distribution of brightness, luminosity, RGB-values or whatever exactly the histogramm shows. But if there is a significant gap on the left (white) or right (black) side, a compensation (stretching or shifting) of the curve might be required. Also if there are peaks jamming/touching either border, it indicates over or underexposure.

 

Matteo Del Grosso

[j.luis aranda]

<i>"How does the histogram of a well exposed pic look like?"</i><p>

There is not a correct standard curve. That depends on each image. But you can learn to identify problems instead, and with time you'll get accustomed to "reading the histogram". Some experimenting with different exposure settings to see how the histogram changes may help. It is like reading music, easy once you know how to do it. Well, the histogram is way simpler than music.<p>

There is a pretty good explanation in the dpreview glossary. Take a look: <a href="http://www.dpreview.com/learn/?/Glossary/Digital_Imaging/Histogram_01.htm">Glossary: histogram</a><p>

Jose Luis

[paul -]

In addition to Jesper's <a href="http://www.luminous-landscape.com/tutorials/understanding-series/understanding-histograms.shtml">Understanding Histograms</a> link, Luminous-Landscape also has another good article called <a href="http://www.luminous-landscape.com/tutorials/expose-right.shtml">Expose (to the) Right</a>.

[jbq]

A correct histogram has no specific shape. It typically hugs the right of the frame, though for highly-saturated colors it may make sense with some cameras to leave a bit of margin.

 

If there's a lot of space to the right of the histogram, you can expose more in order to reduce noise.

[martin_patek_strutsky]

Thank a lot. The links provided are exactly what I looked for.

[bobatkins]

Be careful. In camera histograms aren't very detailed and it's hard to see much on a 1.5" LCD screen. If you have only a few percent of the pixels blown out, you won't see it on the histogram. "Exposing to the right" is fine, but you do run the risk of blowing out small, bright highlights.

[ben_roubin]

How would looking at a histogram be better than simply looking at the picture to see if it's exposed correctly?

[Spearhead]

<i>How would looking at a histogram be better than simply looking at the picture to see if it's exposed correctly?</i><p>

 

I can evaluate the histogram in five seconds with the camera at waist level. Looking at the image requires zooming in and moving around to find out the same information. That takes a lot longer. In the studio or taking a photo of a bulding, it doesn't make as much difference. Working an event, it's the difference between watching and shooting.

[david arnold eugene, oreg]

To add to Jeff's comment, in addition to being easier and faster, reviewing the histogram rather than the image is much more accurate. When looking at the image you're trying to judge exposure by the brightness of the image. But that brightness is only partially determined by the exposure. The other determinant is how bright you've set the LCD panel. In low light it's typically set high so you can see it adequately; in bright sunlight low, so the image doesn't wash out. Thus the apparent exposure is really a combination of actual exposure and LCD setting. Any exposure judgement based on this is thus little more than pure guesstimate.