Peter Oshkai photography

When doing artistic image processing, knowing when to stop is important. In a non-digital world, for example, when painting, continuing to paint beyond a certain point could actually ruin the picture. The colours would become dirty and lose vibrance. With a digital image, there is always a possibility to undo the last action, so as long as the versions are saved regularly, there is no danger of losing all the work. However, there is certainly a point of diminishing returns, when doing a lot of work results in progressively less and less noticeable changes in the image.

I find that some image adjustments are almost always worth doing. They are cropping, colour balance and exposure adjustments. Playing with them does take some time, but the results are significant and there is a good chance of drastically improving an image by playing with these tools.

On the other hand, more detailed work, such as local dodging and burning, using a Liquify tool in Photoshop, skin smoothing in portraits, etc. is not always necessary. Of course, the more work you put in the image, the better it ultimately becomes, but the return on the time investment decreases. There is one counter-argument to this, though, which is quite significant to me. When I spend time working on an image, I am not only investing time in improving it, but also in learning new skills and techniques. In the long run, this makes me more efficient overall. I learn which methods work or don’t work for certain types of images. I also learn to apply various methods faster.

Generally, whether an image is over-processed or not is a very subjective question. I believe that in order to make a photo unique, I have to touch most of its pixels with a brush (I use a Wacom graphics tablet) at some point in the editing process. Having said this, time is a very precious resource, and while I am doing fine tweaks, which might not be noticed by anyone, except myself, I am not doing something else, for example, shooting more photos or learning new techniques. In my experience, the “optimal” stopping point in post-processing is actually not a true optimum, in the optimization theory sense of the word. Rather, it is a bit of a moving target. Finding the balance between under- and over-processed image is like riding a bicycle: sometimes, I lean more one way, sometimes – the other. The important thing is doing it consistently and avoiding extremes.

Retouching children’s portraits requires a surprising amount of skin smoothing. Baby’s skin is a proverbial reference for smoothness. This is why it is expected that skin in a child’d portrait would be impeccably smooth. In reality, children have very thin, quite translucent skin, so an unretouched portrait shows variations of colour, as well as small blood vessels under the skin. While this image is technically accurate, it appears unnatural, because of our expectations of smooth baby skin.

I have just finished post-processing a series of photos that my associate Armando and I took last week for an advertisement campaign of a dance school. I used Portrait Pro Studio 12 – a retouching software that can be called as an add-on from Photoshop. I found that the default settings for a “girl under 12 years old” result in too much processing for my taste. Through some trial and error, I found that these global tweaks give more reasonable results:

• Turn off face sculpting and lighting effects. Children’s faces are rather round and flat; the Portrait Pro algorithms for enhancing the shape of the face detract from these features.
• Reduce the default amount of skin smoothing (controlled by the Master Fade slider under Skin Smoothing Controls) by half.

Before applying the Portrait Pro filter, which appear on a separate layer, I make a stamp visible layer underneath it (create new layer, then press Shift-Option-Command-E). This way, if at a later point I decide that the retouching effects are too strong, I can reduce the opacity of the layer containing them to tone them down.