Shrinking and enlarging are not opposites

Resizing is safe in one direction only. Shrinking throws away information that exists; enlarging invents information that does not. Both are one-way, but they are one-way in different ways.

What actually happens to the pixels

What each direction does to the pixels
Action	What happens	Going back
Shrinking	Several pixels are averaged into one	What was discarded does not return
Enlarging	Missing pixels are guessed from their neighbours	The guesses stay

Shrink and then enlarge again and you do not get the original back — you get a soft copy of it. Keeping the original is the only way around that.

This tool says so when a target is larger than the source. It does not refuse: enlarging deliberately, before printing or before an upscaler, is a real thing people do.

What unlocking the ratio means

With the aspect ratio kept, one number is enough and the other follows. Unlocked, both numbers are used exactly as typed — necessary when fitting a fixed frame, and otherwise the way people end up stretched sideways.

So keeping the ratio is the default and unlocking it means reaching for a switch. It is not somewhere you land by accident.

Pixel size and print size are different questions

"Make it 10cm wide" may be a question about density rather than about pixels. To keep every pixel and change only how large it prints, the DPI converter does that. The print DPI guide has the difference in full.

What this tool holds to

The range and the rules
Property	Value
Smallest edge	1px
Largest edge	8192px
Output format	The same as the source
Original metadata	Gone — the image is encoded again

Keeping the format is deliberate. Changing size and format on one screen makes it impossible to tell which of them spoiled the result. If format is the goal, the format converter is the tool.

Back to PrismMesh