Methodology
How we test sizes and file targets
Every number on ResizeAny should be something you can check. This page explains where the sizes come from, how the tool reaches a file-size target, and which figures on the site are estimates.
Where preset sizes come from
Each preset is a size someone actually needs: an official document photo, an image size published by a platform, a print format or a common upload limit. For official documents we link the government or embassy page in our guides. Platforms change their recommended sizes from time to time, so we compare presets against their help pages when we update them.
Some requirements differ between sources, for example Indian passport photos taken abroad. In those cases we say so and point you to the official page instead of picking one number.
How a file-size target is reached
When you ask for 50 KB, ResizeAny first checks whether the image fits at its full pixel size with good JPG or WebP quality (about 78% or higher). If it does, it raises the quality as far as the limit allows. If it does not, it reduces the pixel size instead of pushing the quality down until the photo looks blocky, then fine-tunes the quality to land just under the target. PNG has no quality setting, so PNG files are only scaled down. The final size is shown before you download.
ResizeAny counts 1 KB as 1,024 bytes, so 50 KB means 51,200 bytes. Some upload forms count 1 KB as 1,000 bytes. If a form rejects a file that should fit, set the target a little lower, for example 48 KB for a 50 KB limit.
Pixels, print sizes and DPI
Print sizes are converted with pixels = millimetres ÷ 25.4 × DPI. A 35 x 45 mm photo at 300 DPI is 413 x 531 pixels. The DPI is written into the file (JFIF data in JPG, the pHYs chunk in PNG), so print software uses the size you chose instead of guessing.
Which numbers are estimates
Preset pages show how many pixels typically fit in a file-size budget. Those figures are rules of thumb for photos saved as JPG at web quality, roughly 0.08 to 0.3 bytes per pixel depending on detail. A photo of a plain wall compresses much better than one of a crowd, so your result can differ. The tool always works from your actual image.
Reporting a mistake
If a size is out of date or a calculation looks wrong, email [email protected] with the page link and, if you have it, the official source. We correct confirmed errors and note the change on the page.