BYTETOOLS

Auto White Balance

Remove an orange, blue or green colour cast from a photo automatically using grey world or white patch estimation, with adjustable correction strength.

Drop a photo with a colour castProcessed locally — never uploaded

What is the Auto White Balance?

A colour cast is what you get when the light in a room was not neutral: tungsten bulbs push everything amber, open shade pushes it blue, and cheap fluorescents add green.

  • Grey world and white patch illuminant estimation
  • Correction strength slider from 0% to 100%
  • Optional 0.5% percentile levels stretch after balancing
  • Estimated illuminant shown as a before and after swatch
  • Exact per-channel gains reported for every correction
  • 100% local processing with full-resolution PNG export

How to use the Auto White Balance

  1. 1

    Drop a photo with a colour cast into the upload area.

  2. 2

    Choose an estimation method: grey world for mixed scenes, white patch when there is a genuine white object.

  3. 3

    Check the estimated illuminant swatch and the red, green and blue gains.

  4. 4

    Lower the correction strength if the result overshoots into the opposite cast.

  5. 5

    Tick the levels stretch for extra contrast, then click Download PNG.

About the Auto White Balance

A colour cast is what you get when the light in a room was not neutral: tungsten bulbs push everything amber, open shade pushes it blue, and cheap fluorescents add green. The ByteTools auto white balance tool estimates that illuminant from the pixels themselves and scales each channel to cancel it, with no eyedropper or grey card required.

Two classic estimators are available. Grey world assumes the average colour of a scene should be neutral and scales each channel toward the overall mean; white patch assumes the brightest 1% of pixels ought to be white. A strength slider dials the correction back when it overshoots, and an optional levels stretch adds contrast at the 0.5% percentiles.

The estimated illuminant is shown as a swatch next to neutral white, with the exact red, green and blue gains printed above it. Everything is measured and applied in your browser — processing is 100% local, your photo is never uploaded, and the corrected full-resolution file downloads as a PNG.

Frequently asked questions

What is white balance in a photo?

White balance is the adjustment that makes neutral objects in a scene actually look neutral. Cameras guess it when the picture is taken, and when the guess is wrong the whole image picks up an orange, blue or green cast that this tool measures and cancels.

How does automatic white balance work without a grey card?

It infers the illuminant from statistics of the image itself. Grey world assumes the average of a varied scene should be neutral grey, while white patch assumes the brightest pixels represent a white surface. Each channel is then multiplied by whatever gain makes that assumption true.

Which method should I choose, grey world or white patch?

Grey world works best on busy, colourful scenes where no single hue dominates, such as street or landscape photos. White patch works better when the picture genuinely contains a white or very bright neutral object, like a shirt, a wall or a highlight on glass.

Why did grey world make my sunset look grey?

Grey world assumes the average colour of the scene is neutral, which is false for images dominated by one hue such as a sunset, a forest or a red brick wall. Lower the correction strength to keep some of the original mood, or switch to white patch instead.

Is my image sent anywhere for analysis?

No. The illuminant estimate and the correction both run on canvas pixel data inside your browser, and no file is uploaded to any server. The tool keeps working offline once the page has loaded.

Related tools