When Photo Compression Helps, and When It Does Not
Compression looks like free space — shrink a photo, get back megabytes, no visible loss. The reality is more nuanced. Some photos shrink cleanly to a fraction of their size with no difference you can see; others lose meaningful detail the moment you compress them. Knowing which is which is the difference between cleaning up your library and quietly degrading it.
What compression actually does
Compression shrinks a file by describing the same pixels with less information. Photos with smooth areas — sky, walls, blurred backgrounds — compress dramatically because most of the image is uniform. Photos with fine detail — leaves, fur, textured fabric, distant crowds — compress less cleanly because every pixel carries information that cannot be summarised away.
Great candidates for compression
Screenshots are the obvious win. They are usually large files (high-res display captures) full of flat areas (UI, solid colours, text) that compress to a small fraction of their original size with zero visible loss. Documents photographed once and kept for reference, panoramas shot for memory rather than reprinting, and any photo you keep but rarely zoom into also compress well.
Photos to keep at original size
Anything you intend to print, crop aggressively later, or zoom into for detail belongs at its original size. Portrait shots where the catch-light in the eye matters, landscape photos you will look at on a large display, and product shots that need to remain sharp are not the right candidates. Compression is a tradeoff, and these are the photos where the tradeoff goes the wrong way.
Pick a target before you start
The most useful compression screen lets you choose the target rather than guessing. A balanced setting for "keep, do not enlarge" use is usually enough for everyday photos; a smaller setting is right for screenshots and reference shots; a near-original setting is appropriate for the few photos that need to stay as they are. Without a target, you tend to over-compress the keepers and under-compress the candidates.
Recompress only when it helps
A photo that is already small should usually be left alone. Compressing a 400 KB photo to 250 KB saves little space and may shave visible detail. The compression screen's job is to surface the savings before you commit — when the projected file is meaningfully smaller and the photo is the right kind, compress; when it is not, skip.
Local processing, original kept until you confirm
Compression runs on your device using local image processing. The new file is written to your library, and the original is only replaced after you confirm. Until then, both versions live in the library, so there is no risk of losing the original by accident.