Image Quality Settings Explained: What Does 80% Actually Mean?
What the quality slider really does, why 80 is usually the sweet spot, and why AVIF 55 looks like JPEG 85.
The number is a knob, not a percentage
Set the quality slider to 80 and it is tempting to read that as "keep 80 percent of the image." That is not what happens. The number is an instruction to the encoder about how aggressively it may discard detail — roughly, how coarsely it rounds off fine information in each block of pixels. There is no universal definition of what a given value means, and no codec measures the result against your original to verify a percentage of anything.
In JPEG, for instance, the quality value scales a set of quantization tables. Lower values mean coarser rounding: fine texture gets smoothed away, subtle color transitions merge, and hard edges pick up faint ringing. Higher values mean gentler rounding and bigger files. The slider is a tradeoff dial between file size and fidelity, and the useful question is not what 80 means but where the tradeoff stops being worth it.
Lossy vs. lossless: two different kinds of compression
Lossy formats — JPEG, standard WebP, AVIF — shrink files by permanently throwing away information your eye is least likely to miss. Human vision is poor at noticing small changes in fine texture and much better at noticing edges and faces, so encoders spend their bits accordingly. The discarded detail is gone for good; no later step can recover it.
Lossless formats — PNG, and WebP in its lossless mode — work more like a zip archive. Every pixel comes out exactly as it went in, which is why converting to PNG never asks for a quality number: there is nothing to trade away. The cost is size. For photographs, a lossless file is often several times larger than a lossy version most people could not tell apart from it.
Why 80 is usually the sweet spot
The size-versus-quality curve is steep at both ends. Pushing quality from 80 up toward 95 or 100 can inflate a photo's file size dramatically while changing almost nothing you can see at normal viewing distance — you pay a lot of bytes to preserve noise and micro-texture. Pulling quality down from 80 toward 50 saves less than you might hope, and the visible artifacts arrive fast: blocky patches in smooth areas, banding in skies, mushy detail in hair and foliage.
That is why defaults across the industry tend to land in the 75-85 range for photographic content. It is the flat part of the curve: most of the savings with little or none of the visible damage. If you have no specific reason to deviate, 80 is a sensible place to leave the slider.
Why 80 in JPEG is not 80 in WebP or AVIF
Quality scales are codec-specific. Each format's encoder interprets the number through its own internal math, so the same value produces different results in different formats. AVIF, the newest of the three, compresses far more efficiently than JPEG — a useful rule of thumb is that AVIF at around 55 looks comparable to JPEG at around 85, at a substantially smaller file size. WebP sits between the two; at similar visual quality it typically comes out 25-35 percent smaller than JPEG.
The practical consequence: do not carry a number across formats. If your JPEG habit is 85, the mid-50s in AVIF or the mid-70s in WebP will usually land you in similar visual territory. And when it matters, trust your eyes over the number — convert, open the result next to the original at 100 percent zoom, and look at skies, skin, and fine texture.
Generation loss: re-encoding always costs something
Every lossy encode is a fresh round of discarding, even at high settings. Save a JPEG at quality 80, open it, and save it again at 80, and the second file is slightly worse than the first — the encoder re-rounds detail that was already rounded once, and the errors compound. Repeat that across a chain of edits or format hops and the degradation becomes obvious: muddier texture, stronger halos around edges.
Two habits prevent most of it. First, keep your original and always convert from it, never from an earlier compressed copy. Second, understand that going lossy-to-lossless does not undo anything: converting a JPEG to PNG gives you a pixel-perfect copy of the already-damaged image in a larger file. That conversion is still useful when a tool demands PNG input, but it does not restore quality.
When lossless is the right call
Lossy compression is tuned for photographs. It struggles with the opposite kind of image: screenshots, user interfaces, text, logos, charts, and line art. Sharp edges and flat areas of color are exactly where JPEG artifacts show most — fuzz around letters, smudges along clean lines. Those same images happen to compress extremely well losslessly, so PNG or lossless WebP is both cleaner and often not much bigger.
Lossless also earns its size when an image is a working master — something you will crop, retouch, or re-export repeatedly — or when you are archiving originals. Edit in a lossless format, then export a lossy copy once, at the end, for delivery. For a finished photograph headed to the web, though, lossless is usually wasted bytes.
Practical recommendations
For photos going on the web or into messages: AVIF around 50-65 or WebP around 70-80, or JPEG at 75-85 when you need maximum compatibility with older software. For screenshots, graphics with text, and logos: PNG or lossless WebP. For working files you will edit again: stay lossless until the final export. When in doubt, leave the slider at 80 and move on.
Because Convertmaxxing runs entirely in your browser — files are converted on your device and never uploaded — experimenting is cheap. Convert the same image at two or three settings, compare them side by side at full zoom, and keep the smallest one that looks right to you. It takes a minute, it is free, and it teaches you more about your own images than any general rule can.