Mobile Masking Guide: Selective Adjustments on Your Phone

Masking is the single biggest leap a mobile photo editor can make. The slider work that comes out of the box, brightening the whole image, warming the whole image, adding contrast to the whole image, will get you part of the way there. But the edits that look professional, the ones that feel finished and intentional rather than just filtered, are almost always the result of applying a change to part of the picture instead of all of it. This guide is about that part. It explains what masking actually is, the categories of mask tools you will find in any serious mobile editor, when each one is the right choice, and how to combine them into edits that hold up at full resolution. The aim is capability, not buttons. The names and exact gestures change between apps and updates, but the underlying ideas have been stable for years and are not going anywhere.

What Masking Is, and Why It Matters

Every adjustment you make to a photo has a scope. The simplest scope is the entire image. When you drag the exposure slider in any editor and the whole frame gets brighter, that is a global adjustment. Global adjustments are useful for normalizing a file: setting the overall brightness, fixing the white balance, dialing in a baseline contrast curve. But they have a built-in limitation. They cannot tell the difference between the parts of the image you want to change and the parts you want to leave alone.

A local adjustment, by contrast, is an edit applied only to a chosen region of the image. The chosen region is defined by a mask. A mask is just a way of saying “this part counts, that part does not.” When you draw a mask over your subject’s face and then move the exposure slider, only the face gets brighter. The rest of the image stays exactly where you left it. That is the entire trick. The slider is the same slider. The change is the same change. The mask is what gives the change a target.

Global vs Local: A Practical Example

Imagine a backlit portrait taken at the end of the day. The sky behind your subject is gorgeous, full of color and gradient. Your subject’s face is in shadow because the sun is behind them. If you push exposure globally to fix the face, the sky blows out into a flat white smear. If you pull exposure globally to save the sky, the face goes inky and detail-free. Neither version of the picture is what you saw or what you remember. The scene contains two separate problems that need two separate answers.

Masking is what lets you give two answers to one image. Mask the face, lift the exposure inside that mask, and the face brightens while the sky stays exactly as captured. Mask the sky, pull the exposure down a stop and warm it up, and the sky deepens while the face stays exactly where you set it. Done. That is the simplest possible mask job, and it covers a huge number of real-world photos: backlit anything, mixed light interiors, sunsets with foreground, anyone standing in front of a window. Once you start seeing your photos as a collection of regions that each want a different edit, you stop fighting global sliders and start solving each region on its own terms.

Why Masking Separates Beginner Edits from Professional Ones

Almost every editing tutorial that promises “ten preset looks” or “one slider for moody photos” is selling a global solution. Global solutions are easy to teach and easy to apply. They are also the reason a lot of edited mobile photos look the same. There is only so much variation you can get out of moving the same five sliders for the whole image. The minute you can target a single eye, a single window, a single patch of sky, your edits stop looking templated and start looking responsive to the photograph in front of you. The professional finish is not in the tools. It is in the willingness to localize.

The Three Categories of Mobile Masking Tools

Every mobile editor with serious masking gives you tools that fall into three families. Knowing which family you reach for first is the difference between a five-minute edit and a thirty-minute one. The categories are brush masks, geometric masks, and automatic or AI-driven masks. Each one is good at something specific. Each one is bad at something else. The strongest editors blend all three.

Brush Masks

A brush mask is exactly what it sounds like. You paint, with your fingertip or a stylus, over the parts of the image you want to affect. Wherever the brush touches becomes part of the mask. Brush masks are the most flexible tool you have. There is no shape, edge condition, or weird selection problem that a patient brush will not solve. The cost is time and steadiness. Brushing fine edges by hand on a phone is the slowest and most error-prone way to define a region.

Geometric Masks

Geometric masks are predefined shapes with built-in fades. The two you will see everywhere are the linear gradient (a straight line with one side fully selected and the other fully not, fading smoothly between them) and the radial gradient (an oval or circle with one side selected and the other not, fading from center outward or vice versa). Geometric masks are fast. You drop one in, drag the handles, and you are done. They are perfect for skies, foregrounds, and broad areas of even falloff. They are not good at irregular subjects.

Automatic and AI Masks

Automatic masks use computer vision to identify common image regions for you. Tap once, and the editor returns a mask of the subject, the sky, the background, a specific person, or even smaller features like eyes, hair, lips, and teeth. Under the hood these are semantic segmentation models, the same family of techniques that powers portrait mode and background blur. They have transformed mobile editing because they collapse what used to be five minutes of brushing into a single tap. They are not perfect. They miss, especially on tricky edges and ambiguous scenes. But on a clean shot of a clear subject, they get you ninety percent of the way to a finished mask before you have done anything yourself.

Brush Masking in Detail

Brush masking comes down to three controls and one habit. The controls are size, feather, and flow. The habit is zooming in farther than you think you need to.

Size is the diameter of your brush. Big brushes cover ground quickly and create soft, generous masks. Small brushes are for fine work along edges. The instinct is to pick one size and use it for the whole mask. The better habit is to switch sizes constantly: a big brush to fill the interior of a region, a small brush to walk the edges, the big brush again to clean up any holes in the middle. Feather controls how soft the brush edge is. A high-feather brush blends seamlessly into the surrounding image, which is what you almost always want for tonal edits like exposure or contrast. A low-feather brush has a hard edge, which is what you want when you need to cut along a sharp boundary like the line between a building and the sky.

Flow controls how much of the brush stroke registers per swipe. Low flow lets you build up a mask gradually, layering passes until the area is fully selected. High flow stamps the full strength of the brush in one stroke. For tonal work where subtlety matters, a moderate flow is forgiving: you can paint over an area twice for full effect, once for half effect, and stop where the falloff feels natural. For hard-edged work, max flow is faster.

Auto-Mask Edge Detection

Most modern mobile brushes include some form of edge-aware mode, sometimes labeled auto-mask or smart brush. With this enabled, the brush refuses to paint past hard tonal or color boundaries it detects under the cursor. You can drag a wide brush along the edge of a building and have it stop cleanly at the skyline without spilling into the sky, even if your finger swerves. This is one of the most useful productivity boosters in mobile editing. It does not work on every edge. Soft edges, low-contrast edges, hair, and fine detail confuse it. But on most architectural and clean-subject shots, it lets you brush at half the precision and still get a clean line.

When Brush Is the Only Option

For all the convenience of geometric and AI masks, there are jobs where brush is still the only sensible choice. Irregular shapes that no rectangle, oval, or semantic category will fit. A specific patch of skin on a forehead. A single ribbon of light through a forest. The strap of a bag. Hair edges in tough conditions, where the AI hands back a mask that is technically a person but loses every flyaway hair into the background. A small distracting object that you want to dim without darkening anything around it. In all of these cases, the answer is the brush, possibly combined with the patience to zoom to two hundred percent and walk the boundary by hand. Some of the best mobile edits in the world are still finished with a finger on a small region for ninety seconds.

Geometric Masks: Linear and Radial Gradients

Geometric masks deserve their own section because they solve so many problems so quickly that they should usually be your first instinct, before brushing or before reaching for AI selection. If a region of your image is roughly aligned with a straight axis (sky above horizon, foreground below horizon) or radiates outward from a point (subject in the middle, distractions around the edges), a gradient is almost always the right tool.

The Linear Gradient

A linear gradient mask is a band that fades smoothly from full effect on one side to no effect on the other. It is the digital descendant of the graduated neutral density filter, a glass filter that lens manufacturers have been making for decades to balance bright skies against dark foregrounds at the moment of capture. The mobile editor version does the same job in post. You drop a linear gradient with the dark side over the sky, pull the exposure down inside the mask, and the sky balances against the foreground without any visible seam.

Two things make a linear gradient feel natural. The first is angle. If your horizon is tilted, the gradient should follow the tilt. A perfectly horizontal gradient over a sloping skyline produces visible edges where the falloff line crosses the actual horizon, and the eye picks up on the mismatch immediately. Most editors let you rotate the gradient with two-finger gestures or by dragging an angle handle. The second is the width of the falloff zone. A narrow falloff makes the gradient look like a band, which is wrong unless you are deliberately stylizing. A generous falloff makes the gradient invisible, which is what you almost always want.

The Radial Gradient

A radial gradient mask is an oval or circle of effect with a fade from inside to outside, or from outside to inside. The classic use is a soft spotlight on the subject: drop a radial over the face, invert it so the effect applies outside the oval, and pull the exposure down. The face stays bright, everything around it darkens slightly, and the viewer’s eye is drawn straight to the center without any obvious dark frame around the edges of the picture.

This is also the right way to add a vignette that does not look like a vignette. The traditional vignette tool in any editor darkens the corners uniformly, which can look heavy-handed and which always centers on the geometric middle of the frame. A radial gradient lets you put the lightest spot wherever your subject actually is, off-center, even off-frame, and shapes the falloff to fit the composition rather than the rectangle. The same technique works for selective softening, selective warming, and selective desaturation around a subject. Anything you can do globally you can do as a soft halo around your point of interest.

AI and Auto Subject Masks

Automatic subject detection is the largest single change in mobile editing in the last few years. The capability is built into all of the serious mobile editors now. The names vary. The function is consistent. You tap a button labeled select subject, select sky, select person, select background, or one of the body part variants, and the editor returns a mask that traces what it found.

How These Masks Work, in General

The underlying technique is semantic segmentation. The editor runs the image through a model that has been trained on millions of labeled examples to recognize categories like person, sky, vegetation, building, water, and so on. For each pixel in your photo, the model outputs a probability that the pixel belongs to each category. The mask is the set of pixels with high enough probability for the requested category. You do not need to know the math to use the tool, but it is worth understanding that you are not getting an exact geometric match to anything. You are getting the model’s best guess, and like all guesses it can be confidently wrong.

This is also why the more specific selectors (eyes, lips, teeth, hair) are useful for portrait work. Each one is a separately trained category. Selecting eyes does not mean the editor is finding two oval shapes near a face. It means the model has learned what eyes look like in context and returns a mask of those pixels. The result is dramatically more precise than any brush a human could draw on a small phone screen, especially at full resolution.

When AI Masking Saves Almost All the Work

The cases where AI masks are nearly perfect are the cases where the subject is well separated from the background, well lit, and uncluttered. A portrait taken outdoors with the subject standing in front of a softly out-of-focus tree line is a near-ideal case. A product shot of a phone on a clean tabletop is another. A landscape with a clear horizon and an unbroken sky is another. In these cases, what would have been ten minutes of careful brushing becomes a single tap and a five-second visual check.

When AI Masking Fails

AI masks struggle in predictable ways. Low contrast between subject and background is the biggest one: a person in dark clothing standing in front of dark vegetation is a hard case for any segmentation model, and the resulting mask will have ragged edges where the model gave up. Wisps of hair and fur are perpetual trouble; the model will usually catch the main mass of hair and lose every flyaway, so the masked subject ends up with a bald-looking outline. Transparent and translucent objects (glass, water, fabric) confuse the model because the same pixel partly belongs to subject and partly to background. Reflections and shadows are similarly ambiguous; the model has to choose, and you may not agree with the choice. Heavily compressed source files give the model less information to work with, so the same scene shot in raw will produce a cleaner mask than the JPEG version of itself.

Refining Auto Masks

The most useful idea in mobile masking is that the AI mask is a starting point, not a finished selection. The workflow that produces consistently good edits is iterative. You generate the auto mask, look at it carefully (most editors let you view the mask as a colored overlay or as a black and white matte), and then refine.

Refinement comes in three flavors. The first is addition: anywhere the mask missed pixels that should be included, you brush in. The second is subtraction: anywhere the mask grabbed pixels that should not be included, you brush out. The third is feathering: anywhere the edge of the mask is too sharp or too soft for the tonal effect you are about to apply, you adjust the feather amount or run a smoothing pass. The recipe most editors expose, even if they call the operations different things, is “subject mask, plus brush in this missed shoulder, minus brush out this background detail it grabbed by accident, with edge feathering increased a notch so the exposure change does not show a halo.”

This iterative refinement is what separates good edits from rough ones. The temptation, especially when AI selection works well most of the time, is to trust the first result and move on. The discipline that produces edits that hold up at full resolution is to spend a few extra seconds zooming in on the mask edges, especially at the edges where mistakes are most visible (around faces, around hairlines, around hard skylines), and clean up by hand.

Combining Masks: Union, Intersection, Subtraction

The masks you have looked at so far are simple shapes from a single tool. The next layer of capability, supported in most modern mobile editors, is combining masks together with set operations. Union (or “add”) combines two masks into one that includes everything either of them covered. Intersection (or “intersect”) keeps only the pixels both masks covered. Subtraction (or “subtract”) starts with one mask and removes whatever the other one covered.

The reason this matters is that real images often need masks that no single tool can produce on its own. The classic example is a sky-only mask in a cityscape. The select-sky tool returns the sky, but if the model is uncertain about a few bright window panes or a glass-fronted building, those windows can end up included in the sky mask, which means your sky exposure adjustment will also dim the windows. The fix is a combined mask: select sky, then subtract a separate mask of buildings (often available as its own AI selector, or paintable with a brush). The result is a sky-only mask that dodges the building entirely.

Intersection is less common but powerful when you need it. The pattern is “this region, but only where it is also that region.” Imagine you want to brighten only the lit side of your subject’s face, leaving the shadow side alone. The select-person mask gives you the whole face. A luminance mask of the brighter pixels covers all the bright areas in the image, including the sky and any reflective surfaces. Intersect the two, and you get the bright side of the face only. Add a touch of exposure inside that combined mask, and you have a directional dodge that no single mask could have produced.

Range Masks: Luminance and Color

Range masks are the most powerful and most underused tools in mobile editing. They define a mask not by shape or category, but by a property of the pixels themselves. A luminance mask selects only pixels within a specified brightness range. A color mask selects only pixels within a specified color range. Both are sometimes labeled simply as “auto mask by tone” or “auto mask by color,” depending on the editor.

Luminance Masks

A luminance mask might select only the brightest highlights, only the deepest shadows, only the midtones, or any custom band in between. They are the cleanest way to color grade tonally separated regions of a photograph. The classic move is to mask the highlights, push them slightly cooler and bluer, then mask the shadows and push them slightly warmer and orange. The result is the split-tone look that drives most cinematic color grading, applied surgically to the parts of the image where it actually changes the feel, without ever leaving the editor’s masking panel.

Luminance masks are also the right answer when a global tonal adjustment is too blunt. If you want to recover blown highlights without touching anything else, mask the brightest pixels and pull the exposure down only inside that mask. The midtones and shadows stay exactly where they were. The same logic works in reverse: lift the deepest shadows without raising the rest of the image, and you get a clean shadow recovery that does not flatten the photograph the way a global shadow slider tends to.

Color Masks

A color mask selects pixels of a specified hue, with adjustable tolerance for related hues. Tap the eyedropper on a red shirt and the mask covers everything in the image that is roughly the same red, ignoring the green grass and the blue sky. Color masks are the cleanest way to neutralize a distracting accent in a photo: a road sign, a piece of equipment, an article of clothing in the background that is pulling attention. Mask the offending color, drop the saturation inside the mask, and the distraction fades into the rest of the scene without touching the colors you wanted to keep.

They are also useful for skin tone work, sky color, and foliage. A color mask of skin lets you adjust the warmth and saturation of skin without affecting clothing or background. A color mask of sky blue lets you intensify the sky without saturating any blue clothing in the foreground. A color mask of foliage lets you shift the green of leaves toward a richer or more muted look without touching anything else. Once you start treating color as a region you can mask, every distracting hue and every dull patch becomes addressable on its own terms.

Real Use Cases

Brightening a Backlit Face

The single most common masking job in mobile editing. The subject is in front of a bright background. The face is too dark. The fix is to mask the face (auto subject works for the whole person, auto person works in some editors, brushing is fine if those fail), lift the exposure inside the mask by up to a stop, lift the shadows a little to bring out detail, and warm the mask very slightly to restore skin tone. The rest of the image stays exactly where you set it.

Adding Contrast to a Sky Without Crushing the Foreground

Mask the sky, push the contrast and clarity inside the mask, drop the highlights to recover any bright clouds. The foreground keeps whatever tonal balance you set globally. A clean sky-only mask, possibly refined with a subtraction of buildings, makes this a one-minute edit on most landscapes.

Toning Skin Separately from Background

The mobile beauty work that drives a lot of social photo editing is built almost entirely on masked skin work. Mask the person (or the face specifically), warm the temperature very slightly, lift the shadows on the face, and drop the saturation outside the mask to push the background away. The skin glows. The background recedes. None of the changes look like filters because each one is happening only where it belongs.

Dropping Hot Windows in an Interior Shot

Real-estate and interior shots almost always have at least one window that has blown out completely. Mask the windows (a luminance mask of the brightest pixels often catches them in one tap, refined by a small brush touchup), pull the exposure inside the mask down by one to two stops, and the windows return to a believable view of the outside world. Without the mask, the same exposure cut would darken the whole room.

Color Grading Shadows and Highlights Separately

Two luminance masks, one of the highlights and one of the shadows, with each one carrying a different temperature and tint shift. This is a classic split-tone grade applied surgically. It works on landscapes, portraits, and street photography. The strength of the look is dialed by the magnitude of the temperature shifts; subtle shifts produce a “this looks expensive” feel, larger shifts produce explicit cinematic styling.

Sharpening Only the Eyes

Global sharpening looks crunchy and over-processed. Sharpening only the eyes makes a portrait pop without making any other surface look harsh. Mask the eyes with the dedicated eye selector if your editor has one, or by brushing in two small strokes if it does not. Push the texture and clarity sliders inside the mask only. The face stays soft and the eyes draw the viewer in.

Selectively Desaturating a Distracting Color

A bright orange traffic cone in the corner of an otherwise calm street photograph. A neon-yellow jacket on a bystander in a portrait. The fastest fix is a color mask of the offending hue, with the saturation pulled down sharply inside the mask. The cone goes gray. The jacket fades. The rest of the image stays vivid where you wanted it.

Performance, Resolution, and the Desktop Ceiling

Heavy masking on high-resolution files is genuinely demanding work. A modern mobile sensor produces images at twelve, twenty-four, or even forty-eight effective megapixels. Each mask you create is essentially a full-resolution alpha channel that has to be generated, stored, and recomputed every time you change a slider inside it. Stack five or six masks on top of one another, run the editor on a phone that is several years old, and you will start to feel the lag.

There are tradeoffs available. Most editors let you flatten or merge an edit so that intermediate masks are baked in and dropped from active memory, which speeds the rest of the session at the cost of being unable to revisit those masks later. Working at a smaller preview resolution while you place masks, then letting the editor render at full resolution only on export, helps on weaker hardware. Closing other apps before a heavy mask session is not a joke. And if you are coming back to a session for the second or third time, doing a global render pass to flatten older edits is the fastest way to get the editor responsive again.

When to Bring It to Desktop

There is a ceiling. Some masking jobs are still better on desktop and probably always will be, because the problems they solve benefit from a precision pointer, a large screen, and unconstrained processing power. Compositing, where you cut a subject out of one image and lay them into another, demands tighter selection edges than any mobile editor can comfortably produce. Complex hair selection, where you need to keep the wisps and fly-aways across a contrasting background, is still best handled by desktop tools with dedicated hair refinement features. Frequency separation, the technique behind high-end retouching that splits an image into texture and color layers, is conceptually possible on mobile but practically belongs to the desktop. Heavy local color work that requires viewing a large monitor for accurate calibration is the same. None of this is a knock on mobile editing, which has come further in five years than the previous twenty. It is just an honest map of where mobile is great and where it gives way.

The Right Workflow Order

Masking should not be the first thing you do in an edit, and it should not be the last. The order that produces the cleanest results, every time, is global first, masks second, refine third, globals to taste fourth.

Start with global edits. Set the exposure roughly where you want it. Set the white balance correctly. Apply any baseline contrast curve. The goal is to normalize the file so that the rest of the work is built on a stable foundation. Then add the masks for emphasis. Brighten the face. Deepen the sky. Tone the shadows. Sharpen the eyes. Each mask is a deliberate intervention in a specific region.

Then refine. Walk every mask edge at high zoom. Clean up the spots where the AI got lazy. Adjust feathers where the effect is showing a halo. Strengthen or soften the masked adjustments based on how they read in the full image. Finally, revisit the global sliders one more time. Now that the local work is in place, you can re-evaluate the overall image and make small global adjustments to taste. Often the global exposure that felt right at the start now feels a touch too high, because your masked dodging has lifted the subject naturally; pulling the global exposure down a hair brings everything into balance. This last step is fast and almost always improves the final result.

Common Mistakes

Trusting auto masks without refining the edges. The first result is rarely the right result. Zoom in. Look at the boundaries. Brush a few additions and subtractions. Thirty seconds of refinement saves an edit from looking like an obvious AI cutout.
Masking too hard so the local edit looks pasted on. If you push exposure inside a mask by two stops, the masked region will not just look brighter, it will look detached from the rest of the image. The eye reads the mismatch as artificial. The fix is restraint: most masked adjustments should be subtler than you think, and almost always less than the equivalent global adjustment would be.
Brushing huge regions when a gradient would do. If your mask is going to fade smoothly across half the image, you should be using a linear or radial gradient, not painting it. The brush will never produce as clean a falloff as a built-in gradient, and it will take twenty times as long.
Ignoring feather. A hard-edged mask applying a tonal change creates a visible seam wherever the mask edge falls. Feather softens that transition. Most masked adjustments want generous feather. The exception is hard architectural boundaries where the seam matters.
Not zooming in enough. The most common reason a mobile edit looks rough is that the editor never zoomed past the fit-to-screen view. At full resolution, mistakes that were invisible become obvious. Always zoom to one hundred percent on every mask edge before you commit.
Applying contrasting color to a masked region without a transition zone. A warm color shift inside a tight mask, against a cool background, creates a hard color line that screams “edit.” A wider feather and a smaller color shift produce a result that the eye accepts as natural.
Stacking too many masks without merging. Six masks each doing a small thing can make for a file that lags on the phone and is a nightmare to revisit. If a group of masks together produces a settled effect you are confident in, flatten them and move on.
Forgetting that masks compound with global adjustments. If you push exposure globally by half a stop and then push exposure inside a face mask by another half a stop, the face is now a full stop brighter than baseline. Watch the totals.

Try This

The backlit portrait recovery. Find or take a photo where the subject is standing in front of a bright window or a setting sun. Without touching any global slider, mask the face and lift it until it reads naturally. Notice that the sky behind has stayed exactly as captured. This is the single most useful masking habit you can build, and once you can do it from instinct you will use it on a quarter of your photos.
The sunset with linear gradient only. Find or take a sunset where the sky is gorgeous and the foreground is too dark. Use only a single linear gradient mask to fix it. Practice the angle, the falloff, and the magnitude of the exposure change inside the mask. Nothing else. No global edits, no second mask. The exercise teaches you how much one tool can do when used well.
The split-tone luminance grade. Take any photograph with a clear range of highlights, midtones, and shadows. Apply two luminance masks: one of the highlights, pushed slightly cool and blue, one of the shadows, pushed slightly warm and orange. Compare to the original. Notice how the image feels more cinematic without anyone being able to point at exactly what changed. This is the exercise that will permanently change how you think about color grading.

Frequently Asked Questions

Which mobile editor has the best masking?

The honest answer is that several editors are excellent and the differences come down to specific use cases rather than a clear overall winner. Adobe’s mobile editor leads on AI-driven subject and sky selection and on combining masks. Snapseed is unmatched for its history-of-edits brush approach and runs well on older hardware. VSCO has improved its masking dramatically and is the right pick for editors who want a more visual and less technical interface. The best masking app for you is the one whose tools you can find quickly and use without thinking. See the Lightroom Mobile guide, the Snapseed guide, and the VSCO guide for deeper dives on each.

Can I save and reuse masks across photos?

Mostly no. Masks are tied to the specific pixels of the specific image. The exception is parametric masks like luminance and color masks: those rules can be saved as part of a preset and will reapply themselves to a new image, masking the new image’s highlights or reds the same way. Geometric and brush masks do not transfer cleanly because the geometry would land in the wrong place. See the mobile presets guide for what does and does not travel between images.

Does masking work better on raw files than on JPEGs?

Yes, noticeably. Raw files carry more tonal data, which means the AI segmentation models have more information to draw on and produce cleaner masks at fine edges. More importantly, the local adjustments you apply inside a mask have more headroom on raw: you can push exposure further, recover more highlight detail, and pull more shadow detail before the image breaks. JPEG masking still works for moderate edits, but for any photograph you plan to push hard, raw is the correct starting format. See raw vs JPEG and the mobile raw workflow guide.

Is there a limit to how many masks I can stack?

Practically, yes, although the limit is performance rather than software. Most mobile editors will let you create more masks than is comfortable to work with on a phone. Past about six to eight active masks on a high-resolution file, you will see lag, especially on older devices. The discipline is to consolidate. If three small masks together accomplish what one larger mask could, do the latter. If a group of completed masks is settled and you are not going to revisit them, flatten or merge them so the editor can drop them from active memory.

Can I mask still frames pulled from a video?

Yes. Once a video frame is exported as a still image, it behaves like any other photo file in your editor. The only caveat is that video frames are typically lower resolution and more compressed than dedicated photo captures, which means the AI masking tools have less to work with and you may need to do more brush refinement. Color masks and luminance masks work just as well as on a normal photo.

When should I move an edit to desktop instead of finishing on mobile?

Move to desktop when the edit involves precise hair selection that the mobile editor is not getting clean even with refinement, when you are compositing one image into another, when you need frequency separation or other high-end retouching, when color accuracy matters and you are not on a calibrated screen, and when the masking workload is so heavy that the phone is lagging. For ninety percent of routine editing, mobile is sufficient and faster. The desktop is for the cases where the ceiling on the phone is genuinely lower than what the photo deserves.

Do I need a stylus, or is a fingertip fine?

A fingertip is fine for ninety percent of mobile masking, especially when you are leaning on AI selection and gradients and using the brush only to refine. A stylus helps for precision brush work on small features, eyes, lips, fine architectural lines, and for editors who do a lot of work at high zoom on small phones. If you are masking on a tablet, a stylus is more comfortable and noticeably faster. It is not necessary for the workflow described in this guide.

How do I know if my mask edge is good enough?

Apply the strongest version of the adjustment you plan to make inside the mask, even if you intend to dial it back. Look at the edge at one hundred percent zoom. If you can see a halo, a hard line, or a color shift along the boundary, the mask edge needs work, more feather, or a different combination of tools. If the strong version still looks seamless, dial back the adjustment to where you actually want it and you are finished. The strong-version test catches edge problems that hide at subtle settings and only become visible later in print or on a larger screen.