Getting the most out of PDFs

Compress a PDF or cut its page count — which do I need?

If the file will not attach to an email you need compression; if the document is tedious to read or expensive to print you need fewer pages — compression never removes a page and combining pages barely changes the file size.

3 min read

"Make this PDF smaller" means two unrelated things, and picking the wrong one wastes an afternoon.

The ten-second test

Ask what actually went wrong.

  • "It will not attach." — megabytes. You need Compress PDF.
  • "It is ninety pages and I have to print it." — sheets. You need SmartShrink.
  • "It takes forever to scroll through." — sheets.
  • "It is slow to open." — usually megabytes, occasionally a very large number of images.

Neither tool helps with the other's problem. Compression will not remove a page. Combining pages will not meaningfully shrink a file, because the content was never the size problem.

What compression actually does

A PDF's weight is almost always its images. Compression re-encodes them at a lower resolution or a higher JPEG loss, subsets the embedded fonts to the characters actually used, and drops redundant objects.

What this means in practice:

  • A scanned document compresses enormously, because it is entirely images. Ten megabytes to one is ordinary.
  • A text document barely compresses, because there is nothing heavy in it. If a 3 MB text-only PDF only reaches 2.6 MB, that is not a failure — there was nothing else to remove.
  • Compression is lossy for images. Push it hard and photographs go soft and scanned text gets mushy. The levels exist so you can choose where to stop.

What reducing page count actually does

It changes the layout, not the encoding. Pages that are mostly empty are stacked onto single sheets at their original size, so the text is untouched and the document simply becomes shorter.

The file size afterwards is roughly what it was. Occasionally a little smaller, because there are fewer page objects; never dramatically. If the number of megabytes is what you are watching, this is the wrong tool.

When you need both

A long scanned report is heavy *and* long. Do it in this order:

  1. Reduce the page count first. Fewer pages means fewer images to process afterwards.
  2. Compress second, so the compression runs on the final document.

Doing it the other way round works too, but you compress pages you are about to throw away.

The honest limits

  • Compressing an already-compressed PDF gains almost nothing. Most tools that claim otherwise are re-encoding images that were already re-encoded, which loses quality without saving space.
  • Reducing page count on a dense document gains almost nothing either, and a tool that produces a big number on a dense document did it by shrinking the text.

Both of those are worth knowing before you start, because both look like the tool failing and neither is.

Common questions

Why did compression barely change my file?

Because it was mostly text. Compression works on images, fonts and redundant objects; a text-only PDF has very little of any of those to remove.

Does reducing the page count make the file smaller?

Only slightly. The same content is still there, just arranged onto fewer sheets — the saving is a handful of page objects, not the content itself.

Which should I do first?

Page count first, then compression, so the compression runs on the finished document rather than on pages you are about to combine.