How to Use Frequency Separation for Portrait Retouching in Photoshop

What Is Frequency Separation and Why Should You Learn It?

Frequency separation is a retouching technique that splits an image into two layers: one containing color and tone information (low frequency), and another containing texture and fine detail (high frequency). By separating these elements, you can smooth out blotchy skin tones, under-eye circles, and color irregularities without destroying the pores, hair follicles, and skin texture that make a portrait look like a photograph of a real person.

The problem with simpler retouching methods, such as blurring a duplicate layer and painting it back with a mask, is that they smear texture and color together. The result is skin that looks waxy, plastic, or unnaturally smooth. Frequency separation solves this by letting you work on color and texture independently, giving you surgical control over exactly what you change and what you preserve.

This technique is a standard tool in the kit of professional retouchers working in beauty, fashion, and editorial photography. Once you understand the underlying concept, the actual execution in Adobe Photoshop takes only a few minutes per image, and the results are dramatically better than any one-click skin smoothing filter.

As an Amazon Associate, ExifGrabber earns from qualifying purchases.

How Frequency Separation Works (The Theory)

Every digital image can be thought of as containing information at different "frequencies," borrowed from the same concept used in audio engineering.

Low frequency information includes the broad color transitions, overall skin tone, shadows, highlights, and large-scale gradients in an image. Think of it as the image seen from a distance: you can tell the general color and brightness of each area, but individual pores and hairs are invisible.

High frequency information includes the fine details: skin texture, pores, individual hairs, fabric weave, eyelash strands, and the crisp edges between light and dark areas. This layer, when viewed in isolation, looks like a gray, almost embossed version of the image.

When you stack these two layers back together with the right blending mode, the original image reappears exactly. But because the layers are separate, you can paint on the low-frequency layer to fix a red blotch without touching the skin texture above it, or clone stamp on the high-frequency layer to remove a blemish without altering the surrounding skin tone.

Before You Start: Preparing Your Image

Frequency separation works best as a mid-stage technique. Before you begin, complete these steps:

1. Basic exposure and color correction in Lightroom or Camera Raw. Get the white balance, exposure, and contrast close to final. Use ExifGrabber to check your image's color space and white balance settings in the EXIF data if you are unsure what your camera recorded.

2. Spot healing of obvious blemishes. Use Photoshop's Spot Healing Brush to remove any temporary skin imperfections like pimples, scratches, or stray hairs. These quick fixes are faster to handle before frequency separation.

3. Crop and straighten. Work at your final composition to avoid wasting time retouching areas that will be cropped out.

4. Work on a flattened duplicate. If your PSD has multiple adjustment layers, create a stamped visible layer (Ctrl+Alt+Shift+E on Windows, Cmd+Option+Shift+E on Mac) and work from that.

Good studio lighting reduces the amount of retouching needed. Even so, frequency separation remains essential for polishing skin tones and addressing the subtle color irregularities that even professional lighting cannot eliminate.

Step-by-Step: Setting Up Frequency Separation

Step 1: Duplicate the Background Twice

Select your working layer and press Ctrl+J (Cmd+J on Mac) twice to create two duplicates. Name the bottom copy Low Frequency and the top copy High Frequency. Having clear names prevents confusion when you are switching between layers later.

Your layer stack should now look like this (top to bottom): High Frequency, Low Frequency, Background.

Step 2: Apply Gaussian Blur to the Low Frequency Layer

Click on the Low Frequency layer to select it. Go to Filter > Blur > Gaussian Blur.

The blur radius is the most important setting in the entire process, and it depends on your image resolution and the distance between the camera and the subject:

Portrait type	Recommended blur radius
Close-up headshot (face fills frame)	6 to 10 pixels
Medium portrait (head and shoulders)	3 to 6 pixels
Full-length portrait	1 to 3 pixels
High-resolution file (45MP+)	Scale up by 1.5x to 2x

The goal is to blur away the fine texture (pores, hairs) while preserving the broad color and tonal transitions. If you blur too little, texture remains on the low-frequency layer and you will not be able to smooth tones effectively. If you blur too much, you will lose the natural tonal transitions and the final image will look flat.

A useful test: toggle the layer visibility on and off. If you can still see individual pores on the low-frequency layer, increase the radius. If the features of the face are starting to blend into each other, you have gone too far.

Click OK to apply.

Step 3: Apply Image on the High Frequency Layer

This step is where most tutorials get confusing, because the process differs between 8-bit and 16-bit images. Here is how to handle both.

Click on the High Frequency layer to select it. Go to Image > Apply Image.

For 8-bit images, use these settings:

• Layer: Low Frequency
• Blending: Subtract
• Scale: 2
• Offset: 128

For 16-bit images, use these settings:

• Layer: Low Frequency
• Blending: Add
• Check the "Invert" box
• Scale: 2
• Offset: 0

Click OK. The High Frequency layer should now look like a flat gray image with only the fine texture detail visible, similar to a high-pass filter result.

Step 4: Set the Blend Mode

Change the High Frequency layer's blend mode to Linear Light. The image should now look identical to the original. If it does not, something went wrong in the Apply Image step. Go back and check your settings.

Why Linear Light and not Overlay? Overlay works but it is not mathematically accurate for this technique. Linear Light provides a precise reconstruction of the original image when combined with the Subtract/Add method used in Step 3. Overlay introduces contrast shifts that can lead to subtle tonal errors.

Step-by-Step: Retouching on the Separated Layers

Now that your layers are set up, the actual retouching begins. The key rule: work on one layer at a time, and be intentional about which layer you are on.

Smoothing Skin Tone (Low Frequency Layer)

Select the Low Frequency layer. This is where you fix uneven skin tones, redness, dark circles under the eyes, and splotchy color.

Technique 1: Mixer Brush. Select the Mixer Brush tool (nested under the Brush tool). Set it to: Wet: 30-40%, Load: 30-40%, Mix: 70-80%, Flow: 30-40%. Use a soft brush at a size slightly larger than the area you want to smooth. Paint in small, circular strokes to blend adjacent skin tones together. The Mixer Brush literally mixes the colors under the brush, creating smooth gradients.

This is the bread-and-butter technique for evening out skin. It works beautifully for:

• Smoothing redness around the nose and chin
• Blending the transition between foundation and bare skin
• Softening dark circles under the eyes
• Evening out patchy tans or color differences between the face and neck

Technique 2: Clone Stamp at Low Opacity. Set the Clone Stamp to 15-25% opacity and sample from a nearby area with the desired skin tone. Paint over the problem area. This method gives you more control over the source color than the Mixer Brush.

Technique 3: Lasso + Gaussian Blur. For larger areas of uneven tone, use the Lasso tool with a feather of 15-30 pixels to select the problem area. Apply a gentle Gaussian Blur (5-10 pixels) to smooth the tones within the selection. This is a faster method for broad fixes.

Common mistake: Going too far. Check your work by zooming out to 50% view regularly. Skin that looks perfectly smooth at 200% zoom often looks plastic at normal viewing size. Subtle is the goal.

Fixing Texture (High Frequency Layer)

Select the High Frequency layer. This is where you remove texture-level blemishes: acne scars, prominent pores, stray hairs across the face, and small skin imperfections.

Technique 1: Clone Stamp. Set the Clone Stamp to 100% opacity and Current Layer sampling. Alt-click (Option-click on Mac) to sample an area with clean skin texture, then paint over the blemish. Because you are only working on the texture layer, the underlying color and tone remain untouched.

Technique 2: Healing Brush. The Healing Brush works on the high-frequency layer but can sometimes pull in tonal information from the edges of the sampled area. Use it for larger blemishes where the Clone Stamp would leave visible pattern repetition.

Technique 3: Patch Tool. Draw a selection around a blemish with the Patch Tool and drag it to an area with clean texture. This works well for larger imperfections like scars.

Key principle: Only remove things the subject would want removed. Moles, freckles, and natural skin characteristics are part of a person's identity. Discuss with your subject (or follow the brief) before removing permanent features. Frequency separation gives you the power to change anything, but restraint is what separates good retouching from bad.

Advanced Frequency Separation Techniques

Multi-Band Frequency Separation

Standard two-layer frequency separation covers most retouching needs, but some high-end retouchers use three or more frequency bands for finer control. The idea is to split the image into low (broad color), mid (medium details like wrinkle shapes), and high (fine texture). This is accomplished by creating additional blur/apply image layers at different radii.

For most portrait work, two-band separation is sufficient. Add a third band only if you are working on large-format prints or extreme close-ups where the distinction between wrinkle shapes and pore texture matters.

Dodging and Burning After Frequency Separation

Many retouchers apply frequency separation first and then perform dodge and burn work on a separate layer above the stack. This ordering makes sense because frequency separation fixes the color baseline, and dodge/burn refines the three-dimensional light shaping of the face. See our guide to dodging and burning portraits in Photoshop for a detailed walkthrough of that technique.

Creating a Photoshop Action

If you use frequency separation regularly, recording the setup as a Photoshop Action saves time. Here is the process:

1. Open the Actions panel (Window > Actions)
2. Click the Create New Action button and name it "Frequency Separation"
3. Click Record
4. Perform Steps 1 through 4 from the setup guide above
5. Click Stop Recording

Now you can apply the setup with a single click. Note that the Gaussian Blur radius in the action will be fixed at whatever value you used during recording. For maximum flexibility, insert a stop in the action before the blur step so Photoshop pauses and lets you enter a custom radius for each image.

Common Mistakes and How to Avoid Them

The Plastic Skin Look

Cause: Over-blending on the low-frequency layer, especially with the Mixer Brush at high settings. Every facial feature has natural color variation, and removing all of it makes skin look like a mannequin.

Fix: Work at 50% zoom regularly to check your progress. Toggle the low-frequency layer on and off to see how much you have changed. If the difference is dramatic, you have gone too far. Use Ctrl+Z (Cmd+Z) generously.

Mismatched Texture

Cause: Cloning texture from one area of the face to a structurally different area. Cheek texture looks different from forehead texture, which looks different from chin texture.

Fix: Always sample texture from an area adjacent to the blemish, and from the same facial zone. Forehead to forehead, cheek to cheek, chin to chin.

Halo Artifacts at Edges

Cause: The Gaussian Blur radius on the low-frequency layer was too high, creating visible halos around the edges of facial features like the jawline, hairline, and lips.

Fix: Reduce the blur radius. If halos appear only in specific areas, use a layer mask on the low-frequency layer to mask out the retouching near high-contrast edges.

Working on the Wrong Layer

Cause: Accidentally cloning or painting on the high-frequency layer when you meant to work on the low-frequency layer, or vice versa.

Fix: Color-code your layers. Right-click the layer name and assign a color (blue for low, red for high, for example). Develop the habit of checking the Layers panel before every brush stroke.

When to Use Frequency Separation (And When Not To)

Frequency separation excels at:

• Smoothing uneven skin tones across the face and neck
• Reducing redness and blotchiness
• Softening under-eye circles
• Removing blemishes while preserving surrounding texture
• Fixing patchy makeup application

Frequency separation is overkill or inappropriate for:

• Quick social media edits (use Lightroom's masking tools instead)
• Landscape, product, or architectural photography
• Portraits that are meant to look completely natural and unretouched
• Images where the subject's face is small in the frame

For portraits destined for print or portfolio display, frequency separation combined with dodge and burn produces results that are noticeably better than any automated retouching tool. The technique gives you granular control that AI-powered alternatives still cannot match when precision matters.

Frequency Separation vs. Other Retouching Methods

Frequency separation vs. dodge and burn: These techniques are complementary, not competing. Frequency separation fixes color; dodge and burn shapes light. Use both on the same image.

Frequency separation vs. Lightroom skin smoothing: Lightroom's Texture and Clarity sliders affect the entire masked area uniformly. Frequency separation lets you smooth one specific blotch without touching the clean skin next to it. Lightroom is faster; frequency separation is more precise.

Frequency separation vs. AI retouching plugins: Tools like Luminar Neo's skin retouching, Retouch4me, and RetouchLab can produce good results in seconds. They work well for high-volume workflow where speed matters more than pixel-level precision. But for hero images, magazine covers, or any work where a retoucher's trained eye makes the difference, manual frequency separation still wins.

Putting It All Together: A Complete Portrait Retouching Workflow

Here is how frequency separation fits into a full portrait retouching pipeline:

1. Raw processing in Lightroom or Camera Raw (exposure, white balance, basic tone)
2. Spot healing in Photoshop (temporary blemishes, stray hairs)
3. Frequency separation (skin tone smoothing on low layer, texture cleanup on high layer)
4. Dodge and burn (contouring, enhancing catchlights, shaping facial structure)
5. Color grading (final color treatment, split toning)
6. Sharpening (output sharpening for print or web)
7. Export (check ExifGrabber to verify your color space and resolution metadata are correct for your output medium)

This workflow scales from a 10-minute quick retouch to a multi-hour beauty edit depending on how much time you spend on each step. The frequency separation step typically takes 5 to 15 minutes on a well-lit portrait and longer on images with challenging lighting or heavy makeup.

Practice Exercise

To build proficiency, try this exercise:

1. Open a portrait where the subject has visible skin texture and some color unevenness (most portraits have this, even well-lit ones).
2. Set up frequency separation using the steps above.
3. Spend exactly 5 minutes working only on the low-frequency layer. Focus on evening out the largest tonal differences.
4. Spend exactly 5 minutes working only on the high-frequency layer. Remove 3 to 5 blemishes using the Clone Stamp.
5. Toggle the frequency separation layers off to compare your retouched version with the original.

Repeat this exercise on 10 different portraits at varying distances and lighting conditions. By the tenth image, the setup will be muscle memory and you will have developed an intuition for how much smoothing is too much.

Frequency separation is one of those techniques that seems complex when you read about it but becomes intuitive once you have done it a handful of times. The payoff, portraits with smooth, natural-looking skin that still feels like real skin, is worth the learning curve.