I used the Clarity slider in Lightroom and liked the way it separated her from the background, in Lightroom 4 it seems to introduce more contrast. For example, take a quick read through these two threads: and. However, a quick read through even a few of the many threads on pertaining to color correction should convince anyone that there is a lot more to color correction than using highly simplified methods. Another "mechanical" approach is to adjust the image to make the skin tones come out close to the published average CMYK values for a person of that race, eg. If one is going to try to get by with simple mechanical approaches to color correction, just shoot a frame with a grey (or underexposed white) card to get a decent color balance. The attached image illustrates how different the average color would be if the OP simply had cropped or framed his image differently. For example, the results can depend on the cropping and framing, the color of clothing worn, etc. Many things besides the background unduly influence the averaging technique. However, if the same subject is standing in the same place but it's now closer to sunset and the sky is magenta/pink, the technique will move the entire image towards the green or cyan. So, for example, if a perfectly exposed, perfectly color corrected person is standing in front of a nice blue sky, this technique will make the entire image (including the person) more yellow. The reason for the inaccuracy is because the results depend on all the colors in the image, not just the skin tones. Sometimes it improves things, but unfortunately, it can give wildly incorrect results, and IMHO, should never be trusted. Obviously, if there is resolution to spare, one can use the patch tool, content-aware fill, spot and other tools instead of the pasting-in-texture approach I just outlined.Īntonio, the average color technique (that you described) has become well known primarily because it is easy to understand and simple to perform. I find this technique of replacing skin texture to be extremely general and useful. If too much of the dirt texture is still showing through, and I don't have the resolution to work on it pixel by pixel, I might apply a bit of a blur and brightening to the entire dirty area, and then put the "soft light" high-pass filtered layer on top of that. The approach I take to the luminosity on low rez dirty images is to "steal" a patch of good skin (ie, skin with a similar texture, lighting, and brightness), put it on a new layer, run it through a high pass filter, and then change the blending mode of that layer to "soft light". Obviously, one shouldn't just increase the brightness of all the pixels in that general area because you will almost certainly brighten "clean", adjacent pixels that don't need brightening. So, now that we've got hue and saturation taken care of, this leaves the big question of what to do about the lower luminosity of those pixels where there is dirt. Then you can adjust the opacity of each for the best overall effect. Set one to "Hue" blend mode, and the other to "Saturation" blend mode. So, to clear this up, I often will make two copies of the layer with color. Since the areas in question are a bit dirty, they probably have less saturation than desirable. The only problem is that if you paint with hue (or put the desired color on a new layer with the blend mode set to "Hue"), the result will take on the saturation value of the underlying image at each point. FWIW, New Haven's suggestion to paint with a better hue is a very good technique that I use myself on occasion.
0 Comments
Leave a Reply. |