Compress Images

By quality gives you a slider. Higher keeps more detail, lower makes a smaller file, and the same setting applies to every image in the batch. This is the mode to use when you care how the result looks.

By target size works backwards from a number. You say 200 KB and the tool finds the highest quality that still fits. Use it when something else imposes the limit — an upload form that rejects anything over 2 MB, an email attachment cap, a job application that specifies a maximum.

For photographs, the honest answer is that 80 is right almost always. The range worth knowing:

• 90 to 100 — visually identical to the original, but the file barely shrinks. Mostly wasted.
• 75 to 85 — the useful range. Typically 60 to 80 percent smaller with no difference anyone notices on a screen.
• 60 to 75 — visible if you look closely at edges and skies. Fine for thumbnails and background images.
• Below 60 — obvious blockiness and smearing. Only for previews, or when a hard size limit leaves no choice.

Compression works by discarding detail the eye is bad at noticing, so how much a file shrinks depends entirely on what is in it. A photograph of a blank wall compresses enormously; a photograph of gravel or foliage barely compresses at all, because every pixel genuinely differs from its neighbours.

This is why the target size mode has to encode the image several times and measure the result. There is no formula to shortcut it — the tool narrows the range with each pass until it lands on the best quality that fits.

The single biggest saving usually is not compression at all. A photo straight from a phone is around 4000 pixels wide. Displayed in a 600-pixel column, more than 90 percent of those pixels are thrown away by the browser after being downloaded.

Run the image through the resizer first, down to roughly twice the size it will actually be displayed at, and then compress. The combination routinely turns a 5 MB photo into 150 KB with no visible difference, which no amount of compression alone would achieve.

JPG and WebP are lossy: each pass permanently discards detail, and the damage accumulates. Compressing an already-compressed file twice does not halve it again — it just degrades it. Keep your originals and always start from them.

PNG is the exception, being lossless. That also means a PNG photograph will never shrink much no matter what you do, because the format is obliged to reproduce every pixel exactly. If it is a photograph, convert it to JPG or WebP.

Upload forms routinely name a number and refuse anything above it, and the numbers are small: 100 KB for a photograph, 50 KB or even 20 KB for a signature, 200 KB for a document scan. Exam boards, visa applications and government portals are the usual source.

That is what the target size mode is for. Give it the number the form wants and it searches for the quality that lands just underneath, rather than leaving you to guess at a percentage and re-export five times.

When it cannot reach the target it says so instead of handing back something too large. The fix is nearly always the same: the image has far more pixels than the form needs. A 12-megapixel photograph will not compress to 50 KB while it stays 12 megapixels, and no quality setting changes that. Resize it to something near the size it will actually be displayed at, then compress.

Where a form gives both a size and a dimension — “maximum 200 KB, 600 × 600 pixels” is typical — resize to the dimension first with the Image Resizer, then bring the file under the size here. Doing it in that order almost always works on the first attempt.

Your images are decoded, redrawn and re-encoded by your own browser. No file is transmitted, no copy is stored, and there is no queue or daily limit to hit — the work happens on your processor, which is also why a large batch runs at the speed of your machine rather than your connection.