Hey all,
I'd like to start by saying that I'm not a photoshop expert, and the following is my opinion. All of the below refers to shots that are intentionally exposed perfectly, in contrast to being intentionally over or underexposed.
I see a lot of histograms on internet forums (including this one) with funny shapes - mostly slight overexposures and underexposures with areas lacking information, so I thought I'd take a minute to explain my method of optimizing images for print and internet display.
A perfect exposure of the perfect compostion will completely cover a histogram. Every exposure should be perfect. In a perfect world, every exposure's darkest point will be black, and lightest point will be white. A sensor only detects light from black to white, so it is important that it sees black and white in every shot. Whether or not every tone inbetween is present is irrelevant - if the lightest point is grey, you cannot have tones lighter than grey!!
This histogram is capable of, and contains, every tone that jpgs are capable of containing. If i printed the photo that this belongs to, it would use every tone (not hue, that's different!) the printer is capable of printing.
This is a shot I took last month, the exposure is pretty good for the light, there isn't much I could have done in camera to change it for the better. It's a bit underexposed, but to increase the exposure would have blown the sky out and risked motion blur. To decrease the exposure would lose detail in the shadows. Since the light was very bright, contrast would also be increased. look at how the information in the histogram is bunched up in the center:
Since the shot was exposed well enough, it looks OK with no adjustments made at all - a little dark, thats because the information is mostly darker than the tone middle grey - but since there is no black and no white, I haven't lost any detail to blown highlights or clipped shadows. This is the state I see a lot of photographs presented at. The histogram is bunched up because the camera sensor did not capture values for the far and near end of the graph. The raw shot contained no black and no white, but lots of information in the middle. Here's the original, no PP except resize.
A really rip-roaring bird shot will contain black AND white values in most situations - My shot uses a total of 194 tones in varying quantity, but I have 255 to play with, thats a loss of 1/5 of the information my photo could contain. There are, of course, exceptions - but most wildlife photos benefit from having at least one pixel black and one pixel white. The adjustment, of course, is a new levels adjustment layer.
The left and right sliders determine what value black and white have. Pure black has a value of 0, and pure white has a value of 255 - assuming you are working with Jpg. Raw is a bit different, although the premise is the same. Here I've told photoshop to make value 204 - which is a very light grey, into value 255, which is white. All of the pixels in the shot that are 204 have become the threshold of white (255). I've done the same for black, telling photoshop that the photo's value of 10 is actually value 0. Photoshop now averages the pixel values between my two thresholds, adjusting the tones accordingly. Now I've made use of every tone available. Hint: you can see how many pixels are black or white by clicking and holding on the appropriate slider and pressing the ALT key - this will show you the threshold for that particular tone.
This shot looks much more vibrant - but the reds are oversaturated, and the robin is still pretty dark. To lighten the robin, this is your chance to use the fancy quickmask to extract it and create a new adjustment layer - for the sake of this quick post, I've just moved the middle gray slider to the left.
Moving the middle gray slider controls the value of middle gray in proportion to the photo. Middle gray has a value of 128, but Levels displays it as a ratio - it starts at 1. The farther to the left the slider goes, the darker middle gray becomes. The opposite holds true for the right side of the histogram. Changing the value of middle gray in a photo has the opposite effect on percieved lightness. Moving the slider to the left makes more tones in the photo brighter than middle gray, and fewer tones darker. Thus the photo has more light tones and is lighter overall.
I've just eyeballed the adjustment, here middle grey is set to 1.46.
Now I can adjust the curves - a slight S curve gives more contrast, but one or two tones goes a LONG way. Curves controls the magnitude of tones. You can make blacks blacker and whites whiter. You can change the value of white and black also. Of course, the opposite is true also. If a shot has too much contrast, a reverse S curve can help even it out.
And because my reds are oversaturated, I desaturate the reds. Here to about -20. I know the photo shows master saturation, but really I only adjusted the red saturation...just too lazy to reshoot the screenshot.
and finally, the shot is ready for sharpening. I could have spent a lot longer working on this shot, it is a good candidate for an extraction of the bird to adjust it's levels individually, and some cloning might be nice too...but I've already covered that aspect of post processing.
original
edited
I hope that this brief explanation of exposure and optimization helps, there are so many great shots that I see here that do not utilize output the way they should - a shot should be perfectly exposed, unless it is not intended to be. When we fail to expose properly, sometimes photoshop can restore (or create) the information we have failed to capture.




LinkBack URL
About LinkBacks







Reply With Quote
