How to Fix Grainy Photos Without Losing Detail

A grainy photo has too much noise: the speckled, sandy texture you see most in shadows and skies, worst in shots taken in low light or on an old phone. Moderate grain is very fixable in 2026 — AI noise reduction cleans it up while keeping faces and edges intact. But every fix also softens real detail, so the honest goal is less grain, not a plastic, zeroed-out image. Here's how to do it right, and how to tell grain apart from the two problems people confuse it with.

First, confirm it's actually grain

Three different problems all get called "grainy," and they need different fixes. Zoom to 100% and look:

1. Noise (real grain). Random speckles — gray dots, or colored red/green/blue flecks — scattered across smooth areas like sky, walls, and shadows. This is what denoising fixes.
2. Blur. Detail is smeared or soft, not speckled. That's a different guide entirely: see how to fix blurry photos for motion and focus problems.
3. Low resolution. The image is small or blocky, with stair-stepped edges and mushy detail. The fix there is upscaling, which rebuilds pixels rather than removing speckle.

You can have more than one at once — a dim phone photo is often noisy and soft — but you treat them in the right order, and it starts with knowing which is which.

The two kinds of noise

Noise reduction works better when you know what you're removing:

• Luminance noise is the gray, sandy speckle. It comes mostly from a high ISO in low light. It's the more natural-looking of the two and the one to remove gently — push it too far and skin turns to wax.
• Color noise is the random colored dots, often from a sensor working hard on a long or low-light exposure. It looks unnatural and can be removed more aggressively without hurting the photo, because those colored flecks are never real detail.

Treat them separately: clear the color noise firmly, then ease the luminance noise back only until the speckle stops distracting.

A workflow that keeps detail

1. Start from the best original. The camera file or original scan, never a screenshot or a messaging-app re-save — compression bakes noise in and adds its own blocky artifacts.
2. Fix exposure first. A lot of "grain" is really an underexposed photo brightened after the fact. Lifting the shadows correctly can reduce the visible noise before you denoise anything.
3. Remove color noise, then luminance noise. In that order, each only as far as it needs to go.
4. Upscale if it's also small. If the cleaned photo is low resolution, upscaling rebuilds detail — and it works far better on a denoised image than a grainy one, because the model isn't trying to reconstruct the speckle as if it were texture.
5. Sharpen last, lightly. A gentle sharpen recovers the bite that denoising softened. Overdo it and you'll bring the grain right back.

Grainy old and scanned photos

Film grain and scanner noise are their own case. A grainy scan of a decades-old print is noisy and usually damaged, so denoising is one step inside a larger restoration — repair scratches and stains first, then reduce noise, then sharpen faces. Our guide to restoring old photos covers that full order of operations, and a good restoration tool bundles denoising with the rest.

43frames handles this as part of photo restoration: it reduces noise, repairs damage, and sharpens faces in one pass, tuned to keep a real face looking like a real face rather than a smoothed-out mask.

FAQ

What causes grainy photos? Digital noise — mostly high ISO in low light, small phone sensors, or film grain in scans. Underexposing and brightening later makes it worse.

Can grain be removed completely? No, and you shouldn't try. Every denoise pass softens real detail, so aim for less grain, not a waxy, zeroed-out image.

Is grain the same as blur? No. Grain is speckle; blur is smearing. Denoising fixes grain; sharpening or upscaling fixes blur and low resolution.