Image Compression Explained

Two kinds of smaller

An image gets smaller in one of two ways. Lossless compression finds patterns that repeat, like a run of identical sky pixels, stores them more briefly, and rebuilds the exact original afterward. Lossy compression throws information away, deciding which details the eye is least likely to notice, so the result is close but not identical. JPG is lossy; PNG's standard mode is lossless.

What the eye actually forgives

Lossy formats exploit the limits of human vision: we see brightness finely and color coarsely, so encoders store color at lower precision and drop fine high-frequency detail first. A face survives moderate settings, while text on a flat background grows ringing shadows around every letter, and smooth gradients like skies show banding first.

Three dials sharing one budget

For a target size, resolution is strongest: pixels multiply in each direction, so 2000 to 1000 wide removes three quarters of the data. Format matters next, since newer codecs model natural images more efficiently. Quality is tuned per image, not set once and forgotten.

1. Open the Compress image tool and choose your photo.
2. Decide the real display size and resize to it first.
3. Choose a modern lossy codec for photographs, lossless for graphics and text.
4. Lower quality stepwise, previewing after each step.
5. Stop when you can no longer tell the copy from the original at viewing size.

Review and troubleshooting

Learn to name the artifacts: banding in gradients, ringing along high-contrast edges, blocked flat regions, color shifts in fine texture. Each means raise quality, change format or accept a larger file. Compression is a one-way street, so keep an original and start every future edit from uncompressed data.

Common questions