If you regularly edit images then occasionally you might find yourself with multiple copies of the same picture, and wonder how (or even if) they might be different. And that can be a surprisingly difficult question to answer.

Checking the file size might be a good place to start, of course (if they’re the same format, anyway). And visually comparing the pictures could help. But if you need to know for sure then you may want to try DiffImg, a simple tool which can highlight the pixel differences between any two images of the same size.

Launch the program and it’ll ask you to select the pictures you’d like to compare. DiffImg can open a good range of formats (BMP, GIF, ICO, JPG, MNG, PBM, PGM, PNG, PPM, TGA, TIFF, XBM, and XPM files), and in theory at least you can add support for others via Qt plugins (see “Infos” in the settings dialog).

Once you’ve specified your images, DiffImg will spend a few seconds analysing them, before displaying a pixel mask which highlights any differences. Interpreting this can be tricky as there are several viewing options, but if you select “Difference Image” and “Display Difference Image” on the toolbar (the 6th and 8th icon – they have tooltips if you’re unsure) then it should be fairly easy to understand what’s happening.

A pixel mask highlights image differences

If DiffImg simply displays a blank grey pane then that means it’s found no differences, so there’s nothing to display – your two images are the same.

When an image has been edited in some specific areas then you’ll see those modified pixels only. If you’ve used a “remove red-eye” option, say, you’ll see a group of pixels around the eye area. (If you’re not sure where those pixels are, toggle between “Display Modified Image” and “Display Difference Image” to see their position on the full picture.)

But if you’ve also saved the second image using lossy compression, like JPEG, then this will affect a much wider area, and you’ll see red pixels representing an error, yellow when the error is greater than average.

Some of this can appear quite complex. There are lots of statistics describing any image differences, for instance. You can apparently set an “acceptable” level of difference which won’t be highlighted, and even choose a new method for calculating any differences (options are by channel, by channel mean, or by image lightness). And understanding the impact of all of this isn’t helped by the lack of any Help file.

You don’t have to get into that level of detail, though. And once you’ve figured out the basics then DiffImg works well, making it easy to compare images and spot any differences between them.