The limitations of skin processing techniques
This article is written by a member of our expert community. It expresses that member’s views only. We welcome other perspectives. Here’s how to contribute to MM EDU.
I had a conversation with a fellow photographer recently and I realized a few things about skin-processing techniques that I wasn’t consciously aware of.
Imagenomics Portraiture, frequency-separation, high-pass, even Gaussian Blur all do the same thing… they attempt to make skin more evenly toned through blurring.
This is true of every technique above. Even though you might not know what’s happening in Portraiture, it’s really just automating the frequency-separation technique, and that includes blurring (I use the Surface Blur with frequency-separation).
What each of these techniques fails to accomplish is to discriminately attack the problem areas. What it does is provides you a more even skin tone across the board (face). By definition these techniques behave indiscriminately.
Some may argue, “But then we can mask in only the effects where we want them.”
That will make your overall image better, but it doesn’t change the fact that these techniques are blurring luminosity differences without consideration to content. It doesn’t know (or care) that the problem skin is right next to the fragile details of the eyeball and (depending on the parameters) can very well blur the details of the eyeball into the skin. What now?
“Well we can always set more specific parameters.”
You sure can. But with those parameters, now it’s not nearly as effective with larger patches of problem skin.
“But we can do 2 passes on the blurs, one for more detail and one for less detail.”
The band-aid arguments go on and on…
Don’t get me wrong. These techniques have their place. With careful manipulation they can allow you to retouch more images with less time, assuming the “effect” is not detrimental to the image and the client is tolerant of the results. Furthermore, if you “clean up” the skin before applying these techniques you can apply extremely stringent parameters and achieve amazing results. But regardless of the parameters, what ends up happening (to various degrees) is that the middle frequency of skin tones disappears. You get great overall skin tone (and great pore detail if you use frequency-separation or Portraiture) but the local skin tones (middle frequency) get lost.
It’s something of a conundrum really. It’s exactly why until programmers develop an “intelligent blur” that understands the human face, retouchers will always have a place in this industry.
The decisions I make when I’m dodging/burning, healing, or clone-stamping are all dependent on the area of the face I’m working on. I operate on the “conditionals” provided by the way the light interacts with that particular human face (including the bone structure) so I can intelligently manipulate uneven skin tones. Thus, I am incredibly discriminate. I darken what should be darkened and lighten what should be lightened. I don’t apply an “average” across the board.
I don’t shoot as much beauty anymore, but even when I do, skin-techniques such as the above are the last thing I reach for in my bag of tricks.