What quality setting should you use — why 0.8 is the default

After size, quality is the next number you meet when shrinking photos. It is a single value between 0 and 1, and it rarely comes with any explanation of what it decides. The short answer: leave it near 0.8 and move on. What follows is why, and the cases where you should depart from it.

What the quality value actually does

JPEG and WebP are lossy formats. They do not store the original exactly; they throw some of it away to make the file smaller. And they do not discard at random — they discard what the human eye notices least, first.

Our eyes are sensitive to changes in brightness and much less so to changes in colour. We also track gradual shifts across a broad area well, while missing abrupt detail in small ones. Compression exploits exactly that: colour information goes first, then fine detail gets smoothed away.

So the quality value is not "how sharp" but "how much to throw away." Lower it and more goes; the point where the losses become visible shifts a little with every image and every viewer.

Why 0.8 is the default

What you discard and what you can see do not scale together. Roughly:

Going from 0.8 to 0.95 nearly doubles the file while almost nobody can tell the two apart on screen. Drop below 0.7 and it degrades fast. The returns bend at around 0.8, which is why that is the default.

Where different images diverge

• Ordinary photos — 0.75 to 0.85. Landscapes, food and product shots look essentially identical across that range.
• Portraits — lean towards 0.85. Skin is a broad, gentle gradient, which is the first place compression artefacts show.
• Photos with large skies or flat backgrounds — also towards 0.85, for the same reason: banding appears there first.
• Noisy photos — raising quality only grows the file without improving anything. To the compressor, noise is detail that cannot be discarded, so it is expensive. 0.75 is fine.
• Screenshots, logos, diagrams, text — switch to PNG instead of adjusting quality. Lossy compression on lines and flat colour leaves fuzzy edges around letters and often produces a larger file.

PNG has no quality setting

PNG is lossless, so nothing is discarded and moving the slider changes nothing. This tool says so on screen when you pick PNG, precisely so you do not spend time adjusting a control that has no effect.

To make a PNG smaller you have to reduce its dimensions or its number of colours, not its quality. Saving a photograph as PNG multiplies its size, so photographs belong in WebP or JPEG from the start.

Do not compress twice

Saving an existing JPEG as JPEG again discards on top of what was already discarded. This is generation loss. Pushing quality to 1.0 does not bring back information that is already gone; it only makes the file bigger.

A photo that arrived through a messaging app has already been compressed once. Shrinking it again makes twice. Work from the original in one pass where you can. When there is no original, set quality a little higher, around 0.85, to be safe.

Where to look when checking

The Bulk Image Resizer shows the before and after sizes and the percentage saved once it finishes. More useful than the numbers is looking, and there are specific places to look.

• Broad skies or plain walls — check for banding
• Skin — check whether smooth areas have picked up a grainy texture
• Edges of text and lines — check for smearing or shadowing

If those three are clean, the rest will be. To try a different setting you do not need to re-add the files: change the value, press convert again, and the same images are reprocessed.

Size or quality first?

Size does far more work. Shrinking to 1600 pixels at quality 0.8 gives a much smaller file that looks better on screen than keeping 4000 pixels at 0.95, because only 1600 pixels are ever drawn.

So the order is: get the dimensions right, then think about quality. With the dimensions right, the default quality rarely needs touching. Which dimensions to use is covered in image dimensions by use.

Open the Bulk Image Resizer