The Complete Guide to PDF Compression

PDF compression shrinks a file by re-encoding the images inside it and stripping redundant data. Text stays sharp because it isn't stored as pixels. Photo-heavy files and scans can shrink dramatically; text-only files barely change. This guide explains how PDFs store content, lossless versus lossy compression, what results to realistically expect, and how to compress step by step.

Why compress a PDF at all?

Almost every compression job starts with a hard limit somewhere else: an email provider that rejects large attachments, a job portal or government form with a strict upload cap, a slow connection, or simply a folder of scans eating storage. Compression exists to get the same document under that limit with as little visible change as possible. Keeping that goal in mind matters, because it defines success properly: the question is never "how small can this file get?" but "does it fit where it needs to go, and does it still look right when it gets there?"

How does a PDF store your content?

To understand compression, it helps to know that a PDF is not one blob of data. It's a container holding several very different kinds of content, and they behave very differently when you try to shrink them:

• Text is stored as instructions, not pictures.A PDF doesn't save an image of the letter "A"; it saves "draw an A in this font at this position." That's why text stays crisp at any zoom level, and why it takes up remarkably little space. Hundreds of pages of pure text can fit in a couple of megabytes.
• Vector graphics work the same way.Lines, tables, charts, and logos drawn as shapes are stored as compact drawing commands. Like text, they're resolution-independent and already lean.
• Embedded images are grids of pixels. Photos, scans, screenshots, and any picture pasted into the document are stored pixel by pixel. This is where the megabytes live: one full-page scan can outweigh an entire chapter of text.
• Fonts, metadata, and structure round out the file: embedded typefaces, bookmarks, form fields, and sometimes leftover edit history. Usually minor, occasionally bloated.

The practical takeaway: a PDF's size is almost always determined by its images. Everything a compressor does flows from that fact.

What does compression actually do?

A PDF compressor works through the file looking for weight it can remove without changing what you see on the page. In rough order of impact:

1. Downsampling images.A photo embedded at very high resolution doesn't need all those pixels to look good on a screen. Reducing an image's resolution to match how it will actually be viewed is the single biggest lever.
2. Re-encoding images. Images saved in inefficient formats, or at higher quality settings than the eye can appreciate, get re-saved more efficiently.
3. Removing redundant data. Duplicate embedded objects, unused fonts, orphaned edit history, and bloated metadata are stripped out.
4. Repacking the file structureso the PDF's internal streams are stored as tightly as possible.

Notice what isn't touched: your text and vector content pass through unchanged. A well-compressed PDF stays fully searchable, selectable, and sharp, just with lighter images.

Lossless vs lossy: what's the difference?

Every compression technique falls into one of two camps, and the distinction explains most of the tradeoffs in this guide.

Lossless compression stores the exact same data more efficiently, like vacuum-packing a sweater. Unpack it and you get every thread back. Stripping duplicate objects, repacking streams, and trimming metadata are all lossless: the document is byte-for-byte equivalent where it counts. The catch is that lossless techniques have limited headroom, because well-made PDFs already store text and structure efficiently.

Lossy compressionpermanently discards information the eye is unlikely to miss: fine grain in a photo, resolution beyond what a screen can display. This is where the big savings come from, and it's a one-way door: once detail is discarded, no amount of re-processing brings it back. That's why you should always keep your original file, and why compressing an already-compressed PDF a second time yields little and risks visible artifacts.

Practical PDF compression combines both: lossless cleanup everywhere, plus carefully judged lossy re-encoding of images. Text is never treated lossily; it stays exact.

Why do some PDFs refuse to shrink?

If you've run a file through a compressor and watched it drop by two percent, one of these is almost certainly why:

• It's mostly text. A contract or essay has nothing heavy to remove. That tiny reduction is correct behavior, not a broken tool.
• It was already optimized on export.Word, Google Docs, and most "save as PDF" dialogs compress images as they save. The headroom was used before you ever saw the file.
• Its images are already low quality. Screenshots and web images have typically been squeezed once; squeezing them again buys single-digit percentages.
• It's vector-heavy.CAD exports and detailed maps can be large because of millions of drawing instructions rather than images, so image re-encoding can't help them.

We cover this failure mode in depth in Why Won't My PDF Compress?, but the short version is that a PDF that won't compress is usually a PDF that already was.

What compression results should you expect?

Honest ranges by document type. Your exact result depends on your file's mix of text and images, so treat these as expectations, not promises:

How do I compress a PDF with Universal PDF?

1. Open Compress PDF and drop in your file (anything up to 100MB works).
2. Click Compress. Your file uploads over an encrypted connection; images are downsampled and re-encoded while text and vector content pass through untouched. Most files finish in seconds; large scans can take a few minutes.
3. Check the before/after size and download the result.

One habit worth building: keep the original.Compression's lossy step is irreversible, and storage is cheap. If you ever need to print at high quality, you'll want the uncompressed copy.

Should you compress images before creating the PDF?

Yes, when you control the source. It's the single most effective move in this whole guide. Compressing images before they go into a document beats compressing the finished PDF afterwards, for a simple reason: you can judge each image individually at full size, instead of asking one global setting to suit every picture in the file.

Run your photos through Compress Images first (it works entirely in your browser, so the images never leave your device), then insert the slimmed-down versions into your document and export. A report built from pre-optimized images often needs no PDF compression at all. The same logic applies to scanning: 150–200 DPI in grayscale is plenty for readable documents, and a fraction of the size of 600 DPI color.

Print or screen: how much quality do you actually need?

The right compression level depends entirely on where the document ends its life, because screens and printers make very different demands on images:

• Screen reading and email. Displays show far less detail than print, so images can be downsampled aggressively before anyone notices. If the PDF will only ever be read on screens, compress with confidence. This is the sweet spot where files shrink dramatically with no visible cost.
• Office printing.Standard printouts of reports and handouts are forgiving. Moderate compression survives an office printer fine; body text is unaffected either way, since it's not an image.
• Photography and professional print. Here the lossy tradeoff bites. Fine detail discarded for screen viewing becomes visible on a high-quality print or a deep zoom. Portfolios, print-shop jobs, and archival masters should stay uncompressed. Make a compressed copy for emailing and keep the original for output.

Can you undo compression, or compress twice?

Two questions that come up constantly, with the same underlying answer:

No, compression can't be undone.The lossy step discards image data permanently; there's no "decompress" that restores detail that was thrown away. If you only have the compressed copy, that's now your best copy. This is the entire argument for keeping originals.

And no, compressing twice doesn't help.The first pass spends the available headroom; a second pass finds images that are already at the target resolution and quality, and either changes nothing or degrades them visibly for a few more percent. If a compressed file is still too large, the answer isn't another pass but a different strategy: split the document, extract only the pages you need, or go back to the source and rebuild it with lighter images.

What about scanned PDFs?

Scans deserve a special mention because they're both the best and the trickiest case. Because every page is one big image, they compress the most, but they're also just pictures of text: nothing is selectable or searchable, before or after compression. Running your scan through OCR PDF adds a recognized text layer, so the document becomes searchable while keeping its scanned appearance. Compression and OCR solve different problems; heavily used scans often deserve both.

And sometimes the honest answer is that you don't need a PDF at all. If someone only needs to see a single page, exporting it as an image with PDF to JPG, which also runs entirely in your browser, can be smaller and simpler than compressing and sending the whole document.

The short version

PDF size lives in the images; text is practically free. Compression re-encodes those images (losslessly where possible, lossily where it pays), so scans and photo-heavy files shrink a lot and text-only files barely move. Compress images before building the document when you can, match quality to the destination (screen tolerates far more than print), keep your originals, and let Compress PDF handle the rest.