Smart vs extreme PDF compression: which to choose

Why there are two approaches at all

It helps to picture what a PDF is made of. Most PDFs mix two very different kinds of content: vector text and graphics, which are described mathematically and weigh almost nothing, and images, which are stored as pixels and can be enormous. Any compressor has to decide how to treat each. That decision is the whole story, and it splits into two fundamentally different strategies with very different consequences.

Smart compression: keep the structure

Smart compression leaves the PDF’s structure intact and only re-encodes the images inside it at a lower resolution or quality. The vector text is not touched at all, so it stays perfectly sharp, fully selectable and searchable. You end up with the same document, structurally identical, just lighter — because the part that was actually heavy (the images) has been squeezed while the part that was already tiny (the text) is left alone.

This is the right default for almost everything. It gives you a smaller file with no loss of function: you can still select text, copy quotes, search the document and have it reflow correctly. The only limitation is that it cannot promise a specific final size, because how much it saves depends entirely on how much image data the file contained.

Extreme compression: flatten everything

Extreme compression takes a blunter approach: it renders each page to a single flat image and rebuilds the PDF from those images. Because it throws away all the structure, it can reach dramatically smaller sizes — small enough to meet strict limits that smart mode cannot. But there is a real price. The text is now part of a picture, so it can no longer be selected, copied or searched, and heavy compression can soften it.

This mode exists for one job: hitting a hard size limit when readability is all that is required. A portal that demands a file under 100 KB does not care whether the text is selectable — it just needs a legible document that fits. For that, extreme mode is the only option that can guarantee the target.

How target-size presets work

When you must land under an exact size — 100 KB, 500 KB, 1 MB — guessing settings by hand is tedious. Target-size presets automate it. They compress in extreme mode, check the resulting size, and if it is still too big they step down to a lower resolution and quality, repeating until the file fits and keeping the smallest result that meets the target. The larger the target, the gentler the compression they can use, which is why a 1 MB result looks noticeably better than a 100 KB one of the same document.

A simple way to choose

Put the two side by side and the decision is usually easy:

• Want a smaller file but need to keep selectable, searchable text? Use smart mode.
• Must fit under a strict size limit and only readability matters? Use extreme mode or a target-size preset.
• Compressing a scanned document? Either works well, since scans are already images — smart mode alone often shrinks them a lot.
• A pure-text PDF that barely shrinks in smart mode? That is expected; only rasterizing will make it much smaller, at the cost of selectable text.