iformat.io Logo iformat.io

JPEG Quality Settings — What Is the Sweet Spot Between Size and Quality?

P
Mar 13, 2026
6 min read
Every image editor and compression tool shows you a quality slider from 0 to 100. You know that 100 means best quality and 0 means garbage, but what do the numbers in between actually mean? Is 90% significantly better than 80%? Is 60% acceptable? The answer depends entirely on what you're using the image for, and most people either leave quality at 100% (wasting enormous amounts of bandwidth) or pick a random number and hope for the best. Let's demystify what's actually happening at each quality level.

What the Quality Percentage Actually Controls

The JPEG quality setting controls the quantization step in compression — how aggressively fine details are discarded. At 100%, virtually no detail is discarded, and the image is as close to the original as JPEG allows (though it's still not truly lossless — use PNG for that). At lower percentages, the algorithm is more aggressive about rounding off fine detail, discarding subtle colour variations, and simplifying texture information.
Importantly, the quality percentage is not standardized across different software. Quality 80% in Photoshop looks different from quality 80% in GIMP or in an online compressor. The numbers are relative to each tool's internal scale. What matters is the visual result and the file size, not the number itself. Always preview the output rather than trusting the number blindly.

The Diminishing Returns Curve

Here's the key insight that most people miss: the relationship between quality and file size follows a steep curve, not a straight line. Consider a typical 1920x1080 photograph. At 100% quality, it might be 1.5 MB. Drop to 95%, and it's around 900 KB — a 40% size reduction with absolutely no visible difference to the human eye. At 85%, it's around 450 KB — a 70% reduction with differences only visible if you zoom to 200% and compare side by side. At 75%, it's around 300 KB. At 60%, around 180 KB. At 30%, around 80 KB but with clearly visible artefacts.
The practical takeaway: going from 100% to 85% saves you more than two-thirds of the file size with essentially no visible degradation. Going from 85% to 60% saves another 60%, but you start to notice degradation in certain areas (smooth gradients, fine text, subtle textures). Below 50%, the savings are small but the quality loss is significant. This is why 75-85% is universally recommended as the sweet spot for most purposes.

Recommended Settings by Use Case

Quality Settings Cheat Sheet

Print (high quality): 90-100%. File size matters less than visual fidelity. Use this for anything that will be professionally printed. Web — hero images and portfolios: 80-90%. The primary visual content should look its best while still loading reasonably fast. Web — general blog images: 75-85%. The sweet spot for most web content. Looks great, loads fast. Email attachments: 60-75%. Recipients are viewing on screen, often on mobile. Smaller files mean faster delivery and less inbox storage used. Thumbnails and previews: 50-65%. Small display size masks compression artefacts. File size is more important than detail at thumbnail resolution. Social media: 75-85%. Platforms recompress your images anyway, so uploading at very high quality is pointless — the platform will add its own compression on top.

What You Lose at Each Quality Level

At 100-90%, you lose almost nothing. Pixel-level differences exist but are imperceptible in normal viewing. At 90-80%, very subtle detail in smooth gradients and fine textures starts to soften. You might notice it at 200% zoom if you're comparing side by side. In normal viewing, the image looks identical to the original. At 80-70%, compression artefacts become visible if you look for them — slight blockiness in smooth sky areas, minor smearing around sharp text, some loss of fine texture detail. Most viewers won't notice unless the image is displayed very large.
At 70-50%, artefacts are clearly visible on close inspection. Smooth areas show obvious banding, edges get halos, and fine details are lost. The image still looks acceptable at small sizes or with a quick glance, but it won't hold up to scrutiny. Below 50%, quality degrades rapidly. Blockiness is obvious, colour accuracy drops, and the image looks "crunchy." This range is only useful for tiny thumbnails or situations where you're desperately trying to meet a size limit.

EXIF Data and Metadata Stripping

Every photo from a digital camera or phone contains EXIF metadata — camera model, lens settings, GPS coordinates, date and time, colour profile, and sometimes a thumbnail preview. This metadata typically adds 10-100 KB to the file. For a 5 MB photo, that's negligible. For a compressed 200 KB web image, the metadata can be 10-30% of the total file size. Stripping EXIF data is a free file size reduction with no visual impact whatsoever.
There's also a privacy consideration: if your photos contain GPS data and you're posting them publicly, you're sharing your exact location. Most image compression tools strip EXIF data during compression, which reduces file size and removes potentially sensitive location data in one step.

Progressive JPEG vs Baseline JPEG

Standard (baseline) JPEGs load from top to bottom — the browser draws the image line by line as data arrives. Progressive JPEGs load in multiple passes — first a blurry version of the complete image appears, then it sharpens in successive passes. For web use, progressive JPEGs provide a better perceived loading experience because the visitor sees a rough version of the full image almost immediately rather than watching it load from the top.
Progressive JPEGs are also slightly smaller than baseline JPEGs for images over 10 KB (the multiple-pass encoding actually compresses more efficiently). For images under 10 KB, baseline can be slightly smaller. Most modern image conversion tools default to progressive encoding for web output, which is the right choice for almost all use cases. If you want the best possible combination of quality, size, and loading experience, compress at 80% quality with progressive encoding and EXIF data stripped. That's the recipe for an efficient, fast-loading web image.
Browse All Posts