Early in my retouching career, I got a note back from a client that I still think about. She looked at the portrait I’d delivered and said, “She looks great, but… kind of like a video game character?” She wasn’t wrong. I had smoothed every pore into oblivion, cloned out every shadow, and in trying to make the skin look perfect, I had made it look like it belonged on a wax figure. That note stung, but it taught me the single most important thing about skin retouching: texture and tone are not the same problem, and treating them like they are is exactly what creates that plastic look.

The Two-Layer Problem Nobody Explains Clearly

Skin is carrying two distinct types of information at the same time. There’s the tonal layer: the uneven patches of redness, the dark circle under an eye, the slight discoloration around the nose. And then there’s the texture layer: the pores, the fine lines, the subtle peaks and valleys that catch light and tell your brain “this is real skin.”

When we blur or heal without a plan, we collapse both layers into the same destructive edit. We fix the redness, sure, but we also sand away the texture. The result is skin that reads as smooth, but not as human. Frequency separation is the technique that separates these two layers so we can work on them independently, and once it clicks, it changes everything about how you retouch.

Setting Up Frequency Separation in Photoshop (The Numbers That Actually Matter)

Duplicate your background layer twice. Name the bottom copy “Low Frequency” and the top copy “High Frequency.” On the Low Frequency layer, go to Filter, Blur, Gaussian Blur. For a full-resolution portrait shot at 24 megapixels or higher, I use a radius between 6 and 9 pixels. For a lower-res file around 12 megapixels, I bring that down to 4 or 5 pixels. The goal is to blur until the texture disappears but the tonal shapes of the face are still readable.

Now click on the High Frequency layer. Go to Image, Apply Image. Set the layer to your Low Frequency layer, the blending mode to Subtract, Scale to 2, and Offset to 128. Click OK. Then change the blend mode of the High Frequency layer itself to Linear Light. What you’re looking at now is a gray, strange-looking layer that contains only the surface texture. The Low Frequency layer beneath it holds all the color and tone.

Group both layers and add a group mask if you want to keep things tidy. I have a Photoshop action that does all of this in one click. I named it “Thelma and Louise” because it separates two things that were stuck together and sends them somewhere new.

Working the Low Frequency Layer Without Destroying the High

This is where most of the actual retouching happens, and it’s also where people still go wrong. On the Low Frequency layer, use the Lasso tool with a feather of around 15 to 25 pixels depending on the area, select a patch of uneven skin tone, and then go to Filter, Blur, Gaussian Blur again, just enough to blend the selection into the surrounding skin. You are blending color, not destroying structure.

For more stubborn areas like redness around the nose or shadows under the eyes, I use a soft brush set to around 10 to 15 percent opacity and sample a nearby neutral tone using Alt-click. Then I paint very lightly over the problem area, building up gradually. Because the texture is living on a completely separate layer above, none of this painting affects the pores or surface detail at all. The skin still looks like skin.

What to Actually Do on the High Frequency Layer

This layer is for true structural problems: a blemish that’s raised, a scar with a hard edge, or a stray hair casting a shadow across the skin. I use the Clone Stamp tool here at 100 percent opacity, because on this layer you’re moving texture, not color, and soft edges tend to smear the texture into a blur. Sample from a clean area of skin nearby, and clone directly over the problem spot. Done carefully, the result blends invisibly because the tone underneath on the Low Frequency layer remains untouched.

One thing I’ve learned after years of doing this: less time on the High Frequency layer is almost always better. If you find yourself doing heavy work up here, you’re probably trying to solve a color or tone problem with a texture tool. Go back to Low Frequency.

The Realism Check Nobody Builds Into Their Workflow

Before I flatten anything, I toggle the group on and off and look at the original for a full five seconds. Not a quick flash. Five seconds. The eye needs time to adjust and actually compare. I also zoom out to 25 percent and look at the full frame, because plastic skin always reveals itself at distance before it reveals itself close up.

The other check I do is look at the shadow side of the face. Smooth lighting forgives a lot of over-retouching. But where the light falls away and shadow starts to build, that’s where fake skin announces itself. If the shadow side looks like it has the same smooth perfection as the highlight side, I’ve gone too far.

The whole point of frequency separation isn’t to make skin look untouched. It’s to make your corrections invisible, so the viewer’s eye travels to the person’s expression, their eyes, the story in the photograph, rather than stopping to wonder what’s wrong with the skin. If someone looks at your retouch and sees a technique, the technique failed. If they just see a person, you got it right.