Compress JPG Images

What JPG compression actually removes

JPG is a lossy format, so saving one discards information the encoder judges a viewer would not notice. It does this most aggressively on high frequency detail, which is why fine texture, foliage, and hair soften first, and least on flat areas, which is why a clear sky survives. Understanding that is what makes the quality slider predictable: at a high setting little is discarded, at a low setting the file gets small because detail is being thrown away, and the artefacts appear as blocking, ringing around edges, and banding in smooth gradients.

NoteDetail and edges go first, flat colour survives longest. That is why banding appears in skies before it appears in a wall.

Resizing usually beats compressing hard

A large photograph has a large pixel count, and at a given byte budget each pixel receives very little data, so the encoder has to discard a great deal and the result is soft. Resizing the image to roughly the size it will actually be displayed first means each remaining pixel gets a larger share of the budget, and the result looks noticeably sharper at exactly the same file size. Neither operation is better in general; which is right depends entirely on how large the image will be shown.

• Displayed small: resize first, then compress. Sharper at the same byte count.
• Needs full resolution: compress only, and accept a softer result.
• Already small enough: the file is usually already fine, leave it.
• Keep the original. Compressing a compressed file is the worst case.

NotePreparing the dimensions first is usually the better trade at a given file size.

Why the file size varies so much for the same image

File size at a given quality depends on the content, not just the dimensions. A detailed landscape produces a much larger file than a flat graphic of identical size, because there is more unique information to encode. Two photos from the same camera at the same quality can differ several times over. This is why a target size cannot be guaranteed by a quality setting alone, and why a tool that searches for the best quality that fits is more reliable than one that applies a fixed setting and hopes.

• More detail means a larger file at the same quality.
• Identical dimensions can produce very different file sizes.
• A fixed quality setting cannot guarantee a target size.
• Searching for the highest quality that fits is more reliable.

Format choice when JPG is not required

JPG is the safest format and the right default for email, uploads, and anywhere a destination might reject a modern format. But it is not the most efficient. WebP and AVIF reach the same apparent quality at a smaller file, and both support transparency, which JPG cannot do at all. For a website where you control delivery, a modern format is a straightforward win. For a form that validates extensions, JPG is what works.

• Email, print, and any destination with a strict format: JPG.
• A website: WebP or AVIF for a smaller file at the same appearance.
• Transparency required: JPG cannot hold it. Use PNG or WebP.
• Check the destination accepts a modern format before batch converting.

MDN: image file type and format guide