Toolsyy’s Image Compressor does one thing mechanically, and almost every behaviour people find surprising falls out of it. The image is decoded into pixels, painted onto a drawing canvas, and encoded again from that canvas at the quality you choose. Understanding that single pipeline explains the privacy benefit, the missing PNG slider, the black background nobody expects, and the file that somehow got bigger.
Everything that is not a pixel is left behind
A photograph off a phone is not only an image. Wrapped around the pixels is a block of EXIF metadata recording, quite often:
- GPS coordinates — frequently to within a few metres
- Camera make, model and sometimes the body’s serial number
- Lens, aperture, shutter speed and ISO
- The exact date and time, down to the second
Because compression here decodes to pixels and re-encodes from pixels, none of that survives. It is not a feature someone added and could later remove; there is simply no path for it to travel. The canvas holds colour values and nothing else.
That makes this the most reliable way to strip location data from a photo before posting it. Set the quality high — 90 percent or more — and the visible change is negligible while the GPS coordinates are gone. Worth knowing that many social platforms strip EXIF on upload, but plenty of places do not: a file emailed as an attachment, dropped in a chat, or uploaded to a forum or a marketplace listing usually carries everything.
Lossy and lossless are different jobs
JPEG and WebP are lossy. They discard information the eye is poor at noticing — fine gradations of colour, subtle high-frequency detail — and the quality setting decides how aggressively. The size curve is steep at the top: dropping from 100 to 80 percent often halves the file for a difference you cannot see at normal viewing size.
PNG is lossless. Every pixel is reproduced exactly, which is why the quality slider disappears when you select it. That is not the interface being coy — the browser’s encoder ignores a quality argument for PNG entirely, so a slider there would be a control that visibly moves and does nothing at all.
The practical rule follows from the content, not the file you started with:
- Photographs — JPG or WebP. Continuous tone is exactly what lossy compression is built for.
- Screenshots, logos, diagrams, line art — PNG. Lossy compression puts visible fuzz around sharp edges and text.
- Anything needing transparency — PNG or WebP. Not JPG, for the reason below.
The black background nobody expects
Convert a logo with a transparent background to JPG and the transparency comes back black. This catches people out constantly, and it is not a defect in the tool.
JPEG has no alpha channel. None. There is no way to represent “transparent” in the format, so on the way out the alpha has to go somewhere, and what remains is whatever colour sits underneath. A drawing canvas begins its life as transparent black — the pixels are genuinely black with an alpha of zero — so when the alpha is discarded, black is what you are left with.
It surprises people because they expect white, since white is what a transparent PNG looks like against a page. But the page was supplying that white, and the canvas is not a page.
If the image has transparency, keep it as PNG, or choose WebP, which supports an alpha channel and still compresses far better than PNG. If you specifically need a JPG with a white background, place the logo on a white rectangle in an image editor first.
Why a file sometimes gets bigger
Compressing is not guaranteed to shrink anything, and there are two common ways it goes the other way.
The wrong output format. Choosing PNG for a photograph stores every one of those millions of subtly different pixels exactly, and the result can be several times the size of the JPEG you began with. The format, not the setting, is doing the damage.
Re-encoding something already well compressed. A JPEG that has already been optimised has had its smooth detail removed and replaced with compression artefacts. Feed it back in at high quality and the encoder dutifully spends bits preserving those artefacts, because it cannot tell them apart from real detail. You pay for the noise of the last pass.
Which leads to the rule that matters most: always compress from the original. Each lossy pass compounds the damage of the one before — this is generation loss, the same effect as photocopying a photocopy. Three rounds at 80 percent look considerably worse than one round at 60, despite the higher number each time. Keep the original file and re-derive from it rather than overwriting it.
What this tool deliberately does not do
It re-encodes at the original pixel dimensions. A 4,000-pixel-wide photo comes out 4,000 pixels wide, just with a smaller file.
That is worth stating plainly, because for web use resizing usually beats compressing. If an image will be displayed in a 800-pixel-wide column, scaling it to 1,600 pixels for high-density screens and then compressing will produce a dramatically smaller file than quality alone can — you are removing three quarters of the pixels before the encoder even starts. Compress here for the file size and the metadata stripping; resize in an editor first when the image is headed for a web page.
Everything runs in the page. The image is never uploaded, which is the whole point when the file you are stripping GPS data out of is a photo of your house.
FAQ
Does compressing an image remove EXIF and GPS data?
Yes, and it is unavoidable rather than an added feature. The image is decoded to raw pixels and encoded again from those pixels, so anything that was not a pixel is simply not present in the new file. That covers GPS coordinates, camera make and model, lens and serial number, and the original timestamp. If stripping location data before sharing a photo is your actual goal, compressing at a high quality setting achieves it with almost no visible change.
Why is there no quality slider for PNG?
Because PNG compression is lossless, so there is no quality to trade away and the setting would silently do nothing. PNG reproduces every pixel exactly, which is the right choice for screenshots, logos, diagrams and anything with sharp edges or transparency. For a photograph it produces a very large file, and switching the output format to JPG or WebP is what actually reduces the size.
Why did my transparent PNG come out with a black background as a JPG?
Because the JPEG format has no transparency at all. A drawing canvas starts out as transparent black, and when the alpha channel is discarded on the way into a JPEG the black underneath is what remains. It is not a rendering error, it is the only thing that can happen when transparency is removed. Keep the image as a PNG, or choose WebP, which supports transparency and still compresses well.
What quality setting should I use?
For photographs, somewhere between 70 and 80 percent is the usual sweet spot: the file is often less than half the size while the difference is invisible at normal viewing size. Go lower for thumbnails and higher for anything that will be enlarged or printed. Since the before and after sit side by side, the honest method is to lower the setting until you can see the difference, then step back up once.
Why did my compressed file come out larger than the original?
Usually because the output format was wrong for the content. Saving a photograph as PNG stores every pixel exactly and can easily produce a file several times bigger than the JPEG you started with. It also happens when re-encoding an already well-optimised JPEG at a high quality setting, since the encoder has to spend bits describing the compression artefacts of the previous pass as though they were real detail.
Related guides
- HEIC, WebP and JPG explained — which format to convert into, and why your phone stopped producing JPEGs.
- Turning images into a PDF — for when the compressed photos need to become a document.
- How QR codes encode a link — another case where the output format decides how much detail survives.
Try it now: Free Image Compressor