Denoise

Filter

Reduces image noise while preserving edges and detail using Non-Local Means denoising.

Denoise uses Non-Local Means (NLM) filtering to reduce noise. Unlike simple blurring, NLM compares patches across the image to distinguish between noise and real detail, preserving sharp edges while smoothing noisy areas.

Strength controls luminance noise removal, and Color Strength handles chroma noise separately. The template and search sizes control quality: larger values produce better results but take longer.

8-bit processing

This node processes at 8-bit precision. In a 32-bit document, the image is converted to 8-bit before it runs: values above 1.0 are clipped, and each color channel is rounded to one of 256 levels. It's converted back to 32-bit afterward, but the clipped highlights and rounded detail don't come back.

Parameters

Pins

↳ImagePixels

InputRequired

↳MaskMask

InputOptional

↦ImagePixels

Output

Filter

StrengthNumber

Default: 10.0 Range: 1.0–30.0

Luminance denoising strength. Higher = more noise removed, but may lose fine detail

Color StrengthNumber

Default: 10.0 Range: 1.0–30.0

Color component denoising strength. Higher = reduces color noise (chroma speckles) more aggressively

Template Size (px)Integer

Default: 7 Range: 3–21

Size of the patch used for comparison. Larger = smoother result, slower. Must be odd (auto-corrected)

Search Size (px)Integer

Default: 21 Range: 7–81

Size of the area searched for similar patches. Larger = better quality, much slower. Must be odd (auto-corrected)

Mask

Mask FromChoice

Default: Luminance

Color image connected to the mask input

How the connected color image is read as a mask. Luminance: the image's brightness. Alpha: the image's transparency. Red / Green / Blue: a single channel

• LuminanceRead the image's brightness as the mask. The default.
• AlphaRead the image's transparency as the mask.
• RedUse the red channel as the mask.
• GreenUse the green channel as the mask.
• BlueUse the blue channel as the mask.

Mask FitChoice

Default: None (native size)

Mask connected

How the mask is sized when it doesn't match the image. None: center at native size, no scaling. Stretch: scale to exactly match image dimensions. Fit: scale to fit inside image, preserving aspect ratio. Fill: scale to fill image, cropping overflow

• None (native size)Center the mask at its native size, with no resizing. Unmasked regions are unaffected.
• StretchStretch the mask to exactly match the image dimensions. May distort the mask shape.
• Fit (letterbox)Scale the mask to fit inside the image, preserving aspect ratio. Letterboxes where needed.
• Fill (crop)Scale the mask to fill the image, preserving aspect ratio. Crops edges if aspect ratios differ.

Usage Tips

• Start with strength 10 and adjust up for noisier images.
• Increase search size for better quality at the cost of processing time.
• Use Color Strength independently to target colorful noise speckles.
• Connect a mask input to denoise only specific regions, targeting noisy areas while preserving detail elsewhere.

Related Nodes

• Blur
• Sharpen
• Local Contrast