Chip in

Shrink Without Losing Quality

Finds the smallest file you cannot tell apart from the original, by measuring instead of guessing.

Finds the smallest file you cannot tell apart from the original, by measuring instead of guessing.

What Shrink Without Losing Quality is for

Every other converter asks you to pick a quality number, which is a strange question: nobody wants "82", they want the smallest file they cannot tell apart from the original. This measures instead of guessing, encoding the picture several times and comparing each against the original until it finds the point where the difference stops being visible.

How to use it

  1. Add your files. Drop them on the page or choose them. They stay on your device.
  2. Set the output format and quality. The defaults suit most files.
  3. Press convert and save the results, individually or as one ZIP.

Questions

How does it know what looks the same?

It compares each attempt against your original using SSIM, a measure built to track what people actually notice rather than raw pixel difference. A file scoring above 0.995 is indistinguishable outside a pixel-by-pixel comparison, and that is the bar the safe setting aims for.

Why not just use a fixed quality setting?

Because the right number depends entirely on the picture. A smooth gradient can survive quality 80 and come out 98% smaller. A photograph of dense foliage needs 94 to reach the same fidelity, and forcing 80 on it would visibly wreck it. One number cannot be right for both, which is why picking one is a poor way to ask the question.

Is it slow, encoding the image that many times?

It takes six or seven attempts rather than the fifty a plain sweep would, because it halves the range each time. On a typical photograph that is under a second. It is worth saying why nobody else offers this: seven encodes per image is a running cost on a server and free on the machine already holding the file.

What if my image cannot reach the target?

Some pictures cannot. Fine text or sharp diagrams pushed into JPEG will never look identical at any quality, because the format is the limit rather than the setting. It says so instead of quietly handing back the least bad attempt, and WebP or PNG is usually the better answer for that kind of image.

Does it ever make the file bigger?

It can, if the original was already compressed harder than your chosen fidelity allows. The result is reported with the size either way, so you can keep the original if the trade is not worth it.