How to Compress a PDF Without Losing Quality

A PDF that won't attach to an email, won't upload to a portal, or takes thirty seconds to render on a phone is a PDF that has failed its purpose. The usual reflex is to grab the first compressor online, drop the file in, and hope for the best. That reflex produces two predictable problems: the file either gets compressed into something unreadable, or barely shrinks at all. This guide explains how PDF compression actually works, how to pick the right setting for your situation, and what to do when the file simply won't get any smaller.

What compression actually does to a PDF

A PDF is a container: text, vector graphics, raster images, fonts, and metadata all live inside it. Compression works in two layers. The first layer is image compression — raster images (photos, scans, screenshots) are the biggest contributor to file size in most PDFs, and they can be downsampled to a lower resolution and re-encoded with a more efficient codec. The second layer is structural compression — unused objects are removed, font subsets are deduplicated, cross-reference tables are rebuilt, and streams are re-encoded with more aggressive deflate settings. Together these two layers are what every serious PDF compressor runs under the hood.

The layer that matters most is image downsampling. A scanned page stored at 600 DPI (appropriate for prepress) is about sixteen times larger than the same page stored at 150 DPI (appropriate for email). Dropping that DPI is the single biggest lever compression has. The risk is obvious: if you downsample too aggressively, text inside scans becomes unreadable and photos look blotchy.

Pick the preset that matches the destination

• Screen (72 DPI): smallest file, fine for web viewing on a laptop or phone. Not usable for print — text inside scans will look fuzzy on paper.
• Ebook (150 DPI): the best default for email attachments and mobile reading. Keeps scans legible and photos acceptable.
• Printer (300 DPI): office-quality print. Cuts file size meaningfully versus uncompressed but keeps every page printable at standard sizes.
• Prepress (600 DPI): professional print shops. This preset barely shrinks anything because it is designed to preserve maximum fidelity.

A good rule of thumb: if the recipient will only read the PDF on a screen, pick Ebook. If you do not know how the recipient will use it, pick Ebook. Start there and only go smaller (Screen) if the attachment still doesn't fit. Go larger (Printer) only when the file will genuinely be printed at a proper printer.

The three categories of PDF and what to expect

Compression results depend enormously on what kind of PDF you are starting with. There are three common categories, and each responds to compression differently.

• Born-digital text-only PDFs (exports from Word, Google Docs, or a report generator). These files are already mostly text and small structural overhead. Compression typically saves 10–25% — not dramatic, but meaningful.
• Image-heavy reports (slides, brochures, annual reports with photos and diagrams). These often shrink to 30–50% of their original size on the Ebook preset because their images are the bulk of the weight.
• Scanned PDFs (pages that are basically photos of paper). These shrink the most — sometimes to a quarter of their original size — because the DPI drop affects every page. This is also where aggressive compression is most visible, because scan text is an image, not real text.

What to do when compression barely helps

If your file has already been through one compression pass, or if it is a pure-text born-digital PDF, you may see only a tiny reduction. Before assuming the tool is broken, check three things. First, is the file full of embedded fonts? A document embedding five different font families, each with multiple weights, carries substantial weight. Merging to a single font or using only the base fonts can trim megabytes. Second, are there hidden pages or layers? PDFs can contain optional content groups (layers) that are invisible but still ship inside the file. A structural optimizer removes them. Third, are there duplicated images? A table of contents with the same logo on every page should not re-encode that logo 40 times.

When every option has been exhausted and the file is still too big, split the PDF into smaller parts and send them separately, or link to the file from a cloud share rather than attaching. Compression is not magic — past a certain point, the file is already at its minimum practical size.

Preserving searchability and accessibility

A critical warning: some cheap compressors cheat by rasterizing every page into a single flat image. The file gets dramatically smaller, but the PDF is now broken in ways you may not notice until someone complains. The text layer is gone — the file is no longer searchable. Copy-paste pulls out nothing. Screen readers can't read it. Government and publisher submission portals may reject it. And if the file later needs to be edited, there's nothing to edit.

Before you ship a compressed PDF, do a quick three-check: run Ctrl-F and search for a word you know is on page 3; right-click and try to copy a paragraph; open the bookmarks panel and confirm the navigation is still there. If any of these fail, throw away the compressed version and use a tool that preserves structure.

Workflow: a realistic compression session

1. Check the original file size. Note it — you'll compare against it.
2. Decide the destination (email, web, print) and pick the matching preset.
3. Run the compression. Note the new file size.
4. Open the output and do the three-check (search, copy, bookmarks).
5. If the file is still too big, drop one preset (Ebook → Screen) and try again.
6. If the file is now unreadable, move up one preset (Screen → Ebook) and try again.
7. Once you land on a version that fits and remains usable, archive a copy under a clear name so you don't re-compress it later.

Compress scanned documents the smart way

Scanned PDFs deserve their own note. Because every page is an image, they respond dramatically to compression, but they are also the most fragile. Before compressing a scan, run it through OCR (optical character recognition). OCR adds an invisible text layer under each page image, so the PDF becomes searchable and copy-able. Once the OCR layer is in place, even aggressive image compression on the visual layer doesn't kill searchability — the text layer is still intact. Compression without OCR on scans gives you a smaller file that still can't be searched. OCR first, then compress.

Privacy during compression

Compression tools see the full contents of your uploaded file. Pick a tool that runs compression on isolated workers, deletes the file shortly after download, and does not retain document contents for training or analytics. OkFarsi processes every compression job in an isolated worker, strips uploads after your download, and never reads or retains document contents. If the PDF you are compressing contains confidential material — contracts, financial statements, legal documents — those retention rules matter as much as the compression ratio.

Ready to compress?

Open the Compress PDF tool, pick the preset that matches your destination, and run it. If you are unsure, Ebook is almost always the right first choice. And if the result is smaller than you needed but still readable, you're done — don't keep compressing for sport.