How to Remove Noise & Grain from Photos: AI Denoising Guide
High ISO photos and old film scans are plagued by noise and grain. Learn how AI denoising removes noise while preserving sharp detail and natural texture.
What Causes Image Noise?
Noise appears as random speckles of color or brightness that degrade image quality. The main causes:
Digital Noise (High ISO)
When your camera amplifies the sensor signal in low light, it also amplifies random electronic noise.
ISO 100-400: Minimal noise on modern cameras
ISO 800-3200: Noticeable noise, especially in shadows
ISO 6400+: Significant noise across the entire image
Phone night mode: Heavy processing hides noise but adds artifacts
Film Grain
Analog film has physical grain structure:
ISO 100-200 film: Fine grain, barely visible
ISO 400 film: Visible grain, adds character
ISO 800+ film: Heavy grain, can obscure detail
Pushed film: Extreme grain from underexposure compensation
Compression Noise
JPEG compression and social media re-encoding introduce:
Blocking artifacts: Square patterns visible at high zoom
Banding: Smooth gradients become stepped
Mosquito noise: Halo-like artifacts around high-contrast edges
How AI Denoising Works
Traditional denoising (Gaussian blur, median filter) removes noise by blurring — which also destroys detail. AI denoising is different:
Noise pattern recognition: The AI identifies which pixels are noise vs. real detail
Selective removal: Only noise pixels are smoothed; detail pixels are preserved
Texture preservation: The AI understands that skin pores, fabric weave, and leaf veins are detail, not noise
Edge preservation: Sharp edges remain sharp while surrounding noise is removed
When to Denoise
Before Upscaling (Recommended)
Removing noise BEFORE upscaling prevents the AI from amplifying noise into larger artifacts. This is the recommended order:
Denoise
Upscale
Sharpen (if needed)
After Upscaling
If your source image is already clean, upscale first. But if you notice noise in the upscaled result, denoise afterward.