Frequency separation, the studio way

Frequency separation is a retouching technique that splits a photograph into two layers — one carrying the colour and tone, the other carrying the fine texture. Once they're separate, you can soften skin or even out a backdrop without losing any of the grain that makes a photograph feel real. The trap most people fall into is going too soft on the texture layer; that's when faces start looking like wax. The version we run in the studio keeps the skin pores readable at print resolution.

Our default recipe is a high-pass-radius around 4 px on a 24 MP file. We work with two duplicates: a Gaussian-blurred low-frequency layer for tone, and an Apply-Image-derived high-frequency layer for detail. Healing and clone are restricted to the low-frequency layer; we never touch texture with a stamp tool. If a blemish has carried through to the high-frequency layer, we use a small soft brush at 25–40% opacity instead of erasing — the goal is to reduce noise, not delete it.

The last 10% is colour. Once tone and texture are clean, we always finish with a Selective Colour adjustment to pull the warmth back into mid-tones. Skin reads as plastic the moment you crush the reds; a +3 / +5 nudge on cyan-to-red on neutrals is usually enough. Print the final at A4 and look at it from arm's length — if anything reads as smooth instead of soft, you went too far on the high-frequency layer.