Cropping images — why it costs no quality, and when it still hurts

Cropping discards what you do not want, so the pixels that remain are untouched — the crop itself costs no quality. If the result looks soft, it is because what remained was smaller than the space it had to fill and got stretched. That is an enlargement problem, not a cropping one.

You crop to the part of a photo you actually want, use it, and it looks soft. Plenty of people conclude that cropping costs quality. That is half right. More precisely: **cropping does not degrade anything; what happens after the crop does.**

Cropping and resizing are different jobs

	Crop	Resize
What it does	Throws away the outside	Redraws the whole image smaller
Remaining pixels	Exactly as in the original	New pixels merged from several
Field of view	Narrower	Unchanged
Quality	No loss in what remains	Downscaling keeps it sharp

Both look like 'making it smaller', but cropping reduces the field of view while resizing reduces the density.

Take the middle 1000x1000 out of a 4000x3000 photo and those 1000x1000 pixels are the originals. Nothing was synthesised, so nothing was lost.

So why does it look soft?

Because the crop then goes **into the slot you had planned for the full image**. If a blog header is 1200 pixels wide and your crop came out at 1000, the browser stretches it by 1.2x to fill the space. Stretching means inventing pixels that were never captured, and that is where the softness comes from.

So before cropping, know the **pixel dimensions of the destination**. A YouTube thumbnail is 1280x720; an OG image is 1200x630. Your crop must not come out smaller than that. If it is heading that way, crop less tightly or find a larger source.

Do not trust the preview's size

A cropping screen scales the whole photo down to fit the window. A region that looks about 200 pixels across on screen may be 1,600 pixels in the original. Judging by eye — 'that looks too small' — is usually wrong. Check the actual pixel dimensions of the result instead.

Switching's crop tool works **against the original pixels regardless of how small the preview is**. Drag roughly on screen and the cut still happens at full resolution.

Fixing the aspect ratio first saves work

If the destination has a fixed ratio — square, 16:9, 3:4 — crop to that ratio from the start. Crop to some other shape and whatever you hand it to will crop again or pad it, and you no longer control which part survives.

ID photos come with the ratio decided: 3:4 for a résumé, 7:9 for a passport.
Social avatars are square but displayed as circles, so keep anything important away from the edges.
A 16:9 YouTube thumbnail can lose a little from the sides in mobile listings.

The save format matters too

The crop is lossless, but saving the result as JPG runs it through lossy compression once. A single pass is invisible; cropping and saving the same photo repeatedly accumulates. If more editing is coming, keep intermediates as PNG and export to JPG once at the end.

Everything runs inside your browser, so the photo is never uploaded — which matters when you are cropping a face or an ID document you would rather not send anywhere.

Frequently asked questions

Does cropping reduce file size?

Yes — fewer pixels generally means a proportionally smaller file. But if hitting a size limit is the goal, compression or resizing addresses it directly.

Can I undo a crop?

The discarded area is gone from the file, so not from the result alone. Keep the original and crop a copy.

Can I crop several photos to the same region?

If they share dimensions and framing, cropping them to the same ratio keeps the set consistent. But when the subject sits differently in each shot, one fixed region tends to look worse, not better. Outside of fixed-spec work like ID photos, it is usually worth looking at them one at a time.

What happens to EXIF after cropping?

Most of it disappears during re-encoding. If removing location data is the actual goal, use the EXIF tool so you can see what was there and confirm it is gone.