Understanding image quality and compression settings

'Quality 80' means something different in every tool and every format. Here is what the number actually controls, and how to use it with confidence.

Published Updated 8 min read

Quality 95, 80, 60 and 30 on the same photo — the difference is small until it suddenly isn't.

"Quality" sliders in image tools all look similar — a number from 0 to 100 — but they don't mean the same thing across formats, and the relationship between the number and the visible result is not a straight line. Understanding what's actually happening under the slider makes it much easier to pick a sensible value on the first try.

Lossy vs lossless, briefly

Lossless compression (PNG, and the lossless mode of WEBP) rearranges data more efficiently without discarding any of it — decompress it and you get back exactly the original pixels. Lossy compression (JPG, and normal WEBP) deliberately throws away information the human eye is least likely to notice, and that loss is permanent. Only lossy formats have a meaningful quality slider, because only they have something to trade against.

What the JPG/WEBP quality number actually controls

Both formats break the image into small blocks and represent each one as a mix of frequency components — smooth gradual changes plus increasingly fine detail. The quality setting controls how aggressively the fine, high-frequency detail gets rounded off. At high quality almost nothing is rounded away; at low quality only the coarse shape of each block survives, which is what produces visible blocky artefacts and colour smearing.

Quality rangeWhat you'll seeGood for
90–100No visible loss, largest filePrint, archival, close-up product detail
75–89No visible loss at normal viewing distanceMost web and email images
50–74Slight softening in fine texture, fine for busy photosThumbnails, background imagery, long feeds
Below 50Visible blocking and smeared colour in flat areasRarely worth it outside extreme size constraints

Quality numbers don't compare across formats

WEBP quality 80 is not the same amount of visual loss as JPG quality 80 — WEBP's encoder is generally more efficient, so it typically looks better at the same number, or equal quality at a smaller file size. Never assume a number carries across formats; judge by the actual output image and its byte count, not the setting alone.

The same setting behaves differently on different images

Compression exploits redundancy — similarity between neighbouring pixels. A plain sky or a smooth studio background compresses extremely well even at high quality; dense foliage, fabric texture or film grain resists compression and needs a higher quality setting to avoid visible artefacts at the same perceptual result. There's no universal 'right' number — there's a right number for a given image, found by looking at the actual result.

How to pick a setting without guessing repeatedly

  • Start at quality 80 for JPG or WEBP — it's a reasonable default for the vast majority of photographic content.
  • View the result at 100% zoom, not in a shrunk preview, where artefacts are invisible.
  • If it looks fine, try dropping 10 points and check again — you may have headroom.
  • If you see blocking or smearing, go back up in steps of 5 rather than jumping straight to 95.
  • Always check the reported file size alongside the visual check — the goal is the smallest file that still looks right, not the smallest file possible.

Tools for this