Equalize to recover a sky

[Charles2]

The top image came out of a raw converter. I reduced exposure in the converter to avoid blowing much of the sky, but it has a dull uniform overcast.

Stumbled on the idea of applying Special Effects | EqualizeHSV through a mask of the sky. The bottom image is the result - amazing recovery of clouds. I have no idea how equalization works here.



Not perfect. Could not mask the really blown blue at the top left and on the water, and part of the sky still looks blown.

A related use of Equalize: Sometimes the equalized image shows ugly exaggeration of light and dark. Good! If you Composite | Subtractive filter the equalized over the original image and adjust the amount, you often get a pleasingly smoothed image. Again, I have no idea how this works.

[jsachs]

Equalize creates and applies a brightness curve that causes the result image to have roughly equal numbers of pixels at every brightness level.

[couman]

From the electronic manual:

The equalize transformation computes and applies a brightness curve to the input input so that each brightness level is represented equally in the output image. In other words, it attempts to adjust the brightness levels of the input image so that its histogram becomes flat. This transformation is often useful for bringing out hidden detail in an image.

Oops, Jonathan beat me to it:)

[Charles2]

Thank you for the reminder and quote.

Sometimes Equalize brings out detail, as in the example of the clouds. Sometimes Equalize brightens most of the image without bringing out a noticeable amount of detail. I'm trying to relate the histogram-leveling Equalize operation and the change it produces to insight into the exposure or other characteristic of the original image. Cuz then I might learn something about taking the photo in the first place, and I might develop a feel for when applying Equalize is likely to be helpful, instead of trying it randomly. Theory and practice.

[couman]

Charles,
Sometimes you can also use the color correction transformation to gently (and selectively) nudge saturation levels and thereby achieve enhanced saturation contrast in your skies.

[den]

For me... Histogram Equilize seems to work best for those images, generally landscapes, that have a tall maximum single peak around the 50% tone +/- 10%, with decreasing distributions to minimums at 0% [shadows] and 100% [highlights] tones... i.e., a very definite bell shape distribution with relatively narrow width.

For these images, an 'Invert'-ed mid-tone mask [...where mask white represents the shadow/highlight tone ranges and mask black represents the mid-tones...] used with the Equalize transform will allow me to vary to preference the amount of equalization that will occur in the shadows/highlights while maintaining/minimizing changes to the mid-tone distributions which may already contain preference brightness and contrasts...