Compare two images

Did compression damage the photo? What exactly changed between two versions of a design? Are these two files really the same picture? Looking at two images in separate windows rarely answers these questions. Putting them on top of each other, at the same size and position, makes even small differences obvious.

Both images are compared in your browser; neither file is uploaded.

How to use it

1. Load image A, usually the original.
2. Load image B, for example a compressed, edited or older version.
3. Choose a view: slider, side by side or difference.
4. In the slider view, drag the divider to reveal A on one side and B on the other. In the difference view, adjust the sensitivity slider and read the percentage of differing pixels.
5. Check the info table for the dimensions, file size and format of both images.

If the two images have different dimensions, B is scaled to the size of A for the comparison, and the tool notes this.

For web images, the image optimization guide explains which quality settings are usually a sensible starting point.

Judging compression with the slider

The slider is the quickest way to decide how far you can compress a photo. Lossy formats like JPG and WebP show damage in predictable places, so drag the divider over these areas first:

• Smooth gradients such as skies and studio backgrounds, where steps or bands appear.
• Sharp edges and text, which pick up faint halos and smudges.
• Fine textures like hair, grass and fabric, which turn soft or blocky.

A practical routine: compress a copy with Compress image or Compress to a target size, compare it here at normal viewing size, and lower the quality until you can just see a difference. Then go back one step. If you cannot tell the two apart at the size people will see the image, the smaller file is the right choice.

Reading the difference view

The difference view compares the images pixel by pixel. The sensitivity slider decides how big a change must be before a pixel counts as different.

At high sensitivity, even invisible changes are flagged. Saving a JPG again, for example, changes the values of many pixels slightly, so the whole image can light up although it looks identical. At lower sensitivity, tiny shifts are ignored and only real changes stand out: an edited object, moved text, a removed blemish.

The percentage gives you a single number to compare. A result of 0% means the pixels match within the chosen sensitivity. The files themselves can still differ in metadata, format or compression, which the info table helps you see.

When B had to be scaled to match A, scaling alone creates small differences everywhere. Lower the sensitivity in that case, or prepare both images at the same dimensions first.