Guide

A practical guide to image compression

Images are the heaviest thing on most web pages, and the single easiest thing to make lighter. Here is what actually happens when you compress one, what each common size limit buys you, and how to choose settings you won't regret.

Compressing to a specific file size

Most people arrive here because something imposed a limit: an application form, a portal, an e-mail gateway. Type the number in Maximum file size on the compressor — or tap one of the presets — and the tool searches for the highest quality that stays under it, reducing dimensions only when quality alone cannot get there. Here is what each common ceiling realistically buys you.

How to compress an image to 100 KB

100 KB is the tightest limit still in common use, typically on older government portals and forum avatars. A normal photograph will not keep its full resolution here: expect to land somewhere around 900–1,200 pixels on the long edge. Because the budget is so small, cropping first pays off more than anything else — 100 KB spent on a tightly framed face or document looks far better than the same 100 KB spread across a wide scene full of background detail.

How to compress an image to 200 KB

This is by far the most common ceiling on government, university and job application forms, which is why it is the tool's default. 200 KB is comfortable: a 12-megapixel photo usually keeps roughly 1,600 pixels on the long edge at a quality that still looks clean, and a scanned document stays readable at full page width. If a form rejects your file without saying why, 200 KB is the first limit worth trying.

How to compress an image to 500 KB

At 500 KB you are usually no longer trading away pixels. Most photographs reach this target by lowering quality alone, keeping their original dimensions intact — the tool will tell you, since each result reports the dimensions it ended on. This is a sensible ceiling for e-mail attachments, content management uploads and anywhere you want the file smaller without visibly touching it.

How to compress an image to 1 MB

1 MB is rarely a hard requirement; more often it is a house rule for keeping a media library or an inbox manageable. Almost any single photograph fits at full resolution and high quality, so treat it as a safety net rather than a constraint. If you are preparing images for a web page rather than a form, ignore the megabyte and aim far lower — see why page weight matters below.

Lossy and lossless are two different jobs

Lossless compression — PNG, and WebP in lossless mode — rewrites the same pixels in a more compact form. Reverse the process and you get the original back, byte for byte. It works brilliantly on flat colour: logos, screenshots, diagrams, anything with large uniform areas and hard edges. It works poorly on photographs, because photographic noise leaves almost no repetition to exploit.

Lossy compression — JPEG, WebP, AVIF — throws information away and never gives it back. The trick is that it throws away the information your eyes are worst at noticing. The encoder splits the image into small blocks, converts each one from pixels into frequencies, and then rounds off the high-frequency detail most aggressively. It also stores colour at lower resolution than brightness, because human vision resolves luminance far more sharply than hue. This is why a photograph can lose 90% of its bytes and still look untouched, while the same treatment applied to a screenshot of text produces visible fringing around every letter.

Why file size is worth your attention

Page weight is a ranking and revenue problem, not just a tidiness one. Largest Contentful Paint — one of Google's Core Web Vitals — is usually a hero image, so its file size directly sets how fast your page is judged to load. On a mid-range phone over a mobile network, shaving 800 KB off a hero image can move that metric by more than a second. Visitors on metered data pay for every kilobyte you send them, and conversion rates fall measurably as load times climb.

There are blunter reasons too. Government portals, job applications, insurance claims and university forms routinely cap uploads at 200 KB, 500 KB or 2 MB, and reject anything larger without explaining how to fix it. Email providers cap attachments around 25 MB. Content management systems slow to a crawl when a media library fills with 8-megapixel originals nobody ever needed at full resolution.

Choosing a format

• WebP — the sensible default. Roughly 25–35% smaller than JPEG at matching perceived quality, supports transparency and animation, and is supported by every current browser.
• JPEG — still the safest choice for photographs when something else has to read the file: older software, print shops, or a form that names it explicitly. No transparency.
• PNG — keep it for flat graphics, screenshots with text, and anything that must stay pixel-exact. Avoid it for photographs, where it is often larger than the original camera file.
• AVIF — the smallest of the four, and excellent for large photographic hero images, but slow to encode and not yet universal in older software.

If the format is the only thing you need to change — a PNG that has to become a JPG for an upload form, a WebP an older program refuses to open — the image converter does that in one re-encode and leaves the dimensions alone. Converting first and compressing afterwards puts the picture through two lossy passes where one would have done.

Getting a good result

Resize before you reach for quality. A 4000-pixel-wide photo displayed in a 800-pixel column is carrying twenty-five times more pixels than it will ever show; halving the dimensions cuts the file to roughly a quarter before the encoder does any work at all, with no visible loss whatsoever. Only once the dimensions are right should you lower quality, and quality between 75 and 85 is the range where most photographs look identical to the original at a fraction of the weight. Below about 60 you start to see blocking in skies and smooth gradients.

Finally, always compress from the original. Every lossy pass discards detail permanently, and a file compressed three times carries the scars of all three. Keep your masters somewhere safe and treat compressed copies as disposable output.

Converting between image formats

A conversion changes the container, not the picture. That single sentence explains most of the surprises people run into: why a converted file can come out larger than the one they started with, why a blurry JPEG stays blurry when it becomes a PNG, and why a logo loses its transparent background on the way to JPG. Here is what each direction actually does.

PNG to JPG

By far the most requested conversion, and usually the right one. A photograph saved as PNG is stored losslessly, which means every speck of sensor noise is preserved at full cost — a 3 MB PNG routinely becomes a 300 KB JPEG that nobody can tell apart. The one thing you give up is transparency: JPEG has no alpha channel, so anything transparent is filled with white before encoding. That is fine for photographs and wrong for logos.

JPG to PNG

This is the conversion that most often disappoints. PNG is lossless, but lossless means "preserves exactly what it is given" — and what it is given is a JPEG that has already thrown detail away. You get a bigger file containing the same compression artefacts, not a restored image. There is one good reason to do it anyway: you need transparency, or a tool down the line refuses to accept JPEG. Editing rather than archiving is the honest use case.

WebP to JPG

Compatibility, almost always. WebP is smaller and better, and it is also the format that an older photo printer, a corporate document system or an application form written in 2011 will reject without explaining why. Converting back to JPEG costs you a re-encode, so do it from the best source you have rather than from a copy that has already been through the loop twice.

JPG or PNG to WebP

The conversion worth doing if the destination is a web page. WebP is roughly 25–35% smaller than JPEG at matching perceived quality and it keeps transparency, so it replaces both of the older formats at once. Every current browser supports it. Avoid it only where the file leaves the web and lands somewhere you do not control — an email attachment, a print shop, a government upload form.

Why a converted file can get bigger

Two causes, and both are the format doing its job. Converting a photograph to PNG asks a lossless format to store millions of slightly different pixels, which it cannot do cheaply. Converting a flat graphic — a screenshot, a chart, a logo — to JPEG asks a photographic format to encode hard edges, which it does badly and expensively, adding a halo of fringing around every line of text into the bargain. Match the format to the picture: photographs to JPEG or WebP, flat graphics to PNG or WebP.

Converting is not compressing

The converter re-encodes once, at high quality, and leaves the dimensions alone. It answers "I need this as a PNG", not "I need this under 200 KB". If a form is rejecting your file for its weight rather than its extension, the compressor is the tool you want.