Compress a PDF to a target size

How to use it#

1. Drop the PDF onto the box above, or click it to choose one.
2. Type the size limit in KB or MB, such as 150 KB, or pick a common one: 50, 100, 200, 300, or 500 KB, or 1, 2, 5, or 10 MB.
3. Keep pictures in colour, or choose Greyscale to turn photos and scans grey, which leaves more room under the limit.
4. Check the new size. Changing the limit or the colour compresses the file again straight away.
5. Download the smaller copy. The original is left as it was.

The clearest setting that fits#

A size limit usually comes from somewhere that will not bend: a job application portal, a court filing system, a visa form, or a mail server that bounces anything over 10 MB. The file has to be under it, and any smaller than that only costs picture quality.

So it works through nine settings. The first is the Lossless level of the PDF compressor, which keeps every picture exactly as it is, so a file that is over the limit only because of data stored without compression fits with nothing lost. The next keeps pictures at up to 600 dots per inch at 95% JPEG quality, which few people could tell from the original, and two more step down, through the compressor’s Minimal level at 300, to its Light level at 220. Then come its other levels, down to Strong, which keeps pictures at 100, with one setting between each pair. It compresses the whole document at each one it tries and measures the real result, because a PDF’s size after compression cannot be worked out in advance. It stops at the clearest setting that comes in under the limit, which takes five tries at most, so a 5 MB limit gets the sharpest pictures that fit in 5 MB rather than a file far smaller than it needed to be.

Only the pictures lose detail. Text, fonts, links, and drawings are copied across as they were, so every word can still be selected and searched. Two savings that cost nothing are made at every setting as well: parts of the file stored without compression are compressed, and data left behind by earlier edits, which nothing in the file uses, is dropped.

Greyscale for scans and small limits#

A colour picture stores three values for every pixel and a grey one stores one, so the same scan takes less room in grey at the same sharpness. With Greyscale chosen, every setting the search tries writes the pictures in grey, starting at the clearest one after Lossless, since Lossless changes no picture. The room that saves lets it stop at a lighter setting than it could in colour. It makes the biggest difference to colour photos, which often come out a third smaller or more, and to a scanned ID or certificate printed in colour. A scan of black text on white paper has little colour to take out, so it gains only a few percent.

Only pictures turn grey. Text, lines, and shapes the PDF draws itself keep their colours, so a typed document with a blue logo drawn as shapes keeps its blue. A scanned document is made of nothing but pictures, one per page, so a scan comes out grey all over. Pictures in print colours (CMYK) and JPEG 2000 pictures stay in colour, and the result says how many turned grey.

With Greyscale on, a file already under the limit is still compressed, at 95% quality, so that its pictures turn grey as asked. With colour kept, it is handed back untouched.

When it cannot get under the limit#

Some files will not go under, and the page says so. The saving comes from photographs and scans, so a long document that is mostly text, or one whose pictures are already small, may stay over the limit however hard the pictures are squeezed. The result then shows how close the strongest setting got, and that copy can still be downloaded.

It does not go past Strong. Below 100 dots per inch the small print inside a scanned page starts to break up, and a file that fits the form but cannot be read is no use to whoever receives it.

If its pictures are in colour, choose Greyscale first. After that, make the file shorter. Take out pages nobody needs with the page remover, or split it into parts that each fit, since many forms accept more than one file. For a scan of text, scanning again in black and white at 150 to 200 dots per inch makes a far smaller file than any colour or grey scan, and it is kept as it is here.

How the megabytes are counted#

The limit is counted in thousands: 1 MB here is 1,000,000 bytes, the smaller of the two ways a megabyte is counted. A form or a computer that counts in 1024s allows a little more, 1,048,576 bytes to the megabyte, so a file under the limit here is under it there too, whichever way the form counts.