Image Noise Reducer
Clean up grainy photos from dark rooms, night scenes and small phone sensors. The noise reducer smooths luminance noise (the fine grain) and colour noise (the coloured blotches) separately with an edge-preserving filter, so flat areas such as skin, sky and walls become clean while edges and outlines stay sharp. A detail slider brings back fine texture, and presets give a quick start.
- Files stay on your device
- No sign-up
- Free to use
Noise reduction always trades a little fine texture for smoothness. Start low and raise it until the grain is gone.
Noise reduction
Edge-preserving smoothing (guided filter) on brightness and colour separately: flat areas get smooth while edges stay sharp. Zoom into the preview and hold Compare to judge it.
Output
How to use Image Noise Reducer
- Drop one or more photos onto the page.
- Choose a preset or set luminance and colour noise reduction.
- Use Keep detail and hold Compare to judge the result.
- Download the cleaned photos.
Image Noise Reducer features
Edge-preserving
Guided filter keeps outlines sharp.
Two kinds of noise
Grain and colour blotches handled separately.
Detail control
Restores fine texture after smoothing.
Presets
Light, night photo and phone settings.
Batch
Apply the same settings to up to 50 images.
Private
Processed in your browser; nothing is uploaded.
When to use Image Noise Reducer
- Cleaning up photos from concerts, parties and restaurants.
- Improving night and astro shots.
- Preparing phone photos for print.
- Reducing noise before upscaling or sharpening.
Image Noise Reducer FAQ
What causes image noise?
In low light, the camera amplifies a weak signal from the sensor (high ISO). Random variations in that signal appear as grain in brightness and as coloured speckles. Small sensors in phones show it earlier than large camera sensors.
Why two separate sliders?
Colour noise can be removed strongly without visible loss, because our eyes see colour detail at much lower resolution than brightness detail. Brightness noise has to be treated more gently to keep texture.
Will the photo look plastic?
Only with high luminance settings. Keep detail restores part of the fine structure; for natural results, leave a little grain.
Is this AI noise reduction?
No. It is a classic, transparent filter that runs entirely in your browser. AI tools can reconstruct more detail but may also invent it.
When should I use the Image Denoiser instead?
For isolated specks, dust and salt-and-pepper noise. This tool is for even, fine grain across the whole photo.
Are my photos uploaded?
No. The images are processed in your browser and never leave your device.
How edge-preserving noise reduction works
The simplest way to remove noise is to blur the image, but blurring also destroys edges and texture. Good noise reduction distinguishes between random variation, which should be smoothed, and real structure, which should be kept. The tool uses a guided filter, a well-known edge-preserving technique from image processing research.
The guided filter looks at each small neighbourhood of pixels and measures how much the values vary. Where the variation is small – a patch of sky with some grain – it replaces the pixels with a smooth local average. Where the variation is large – the edge of a building or an eye – it keeps the original values. The strength sliders set the boundary between “noise” and “structure”.
Before filtering, the photo is split into brightness (Y) and two colour channels (Cb and Cr), as in JPEG compression. Colour noise is filtered with a larger radius and more strength, because blotchy colour speckles are the most distracting part of high-ISO noise and colour detail is rarely important at pixel level. Brightness is filtered more carefully.
Keep detail mixes back a share of the fine brightness variations that the filter removed. This restores the impression of texture in skin, fabric and foliage and avoids the waxy look of over-processed photos.
For the best result, reduce noise before sharpening or upscaling, since both would amplify the grain. Your photos stay on your device during the whole process.