Most advice about shrinking PDFs starts with a tool and works backward. That is the wrong order. A PDF is a container, and until you know what is filling the container, you cannot make an informed decision about what to squeeze. Sometimes a 40MB file can drop to 2MB with no visible change; sometimes it genuinely cannot shrink at all, and the honest answer is to split it or trim pages instead. This guide covers both cases.

Why PDFs Get Huge

In roughly nine out of ten oversized PDFs, the culprit is images. Everything else is a rounding error by comparison. Here is the usual lineup, in order of how much space each one typically eats:

• Uncompressed or high-DPI images. A photo taken on a modern phone is often 12+ megapixels. Drop it into a document destined for on-screen reading and you are embedding four to eight times more pixel data than any display will ever show. Scanned pages are worse: a letter-size page scanned at 600 DPI in full color is enormous before any compression is applied.
• Fully embedded fonts. To guarantee a document renders identically everywhere, PDF generators embed font files. A well-behaved generator embeds a subset — only the glyphs actually used. A lazy one embeds the entire typeface, sometimes several weights of it, adding hundreds of kilobytes per font for characters that never appear on any page.
• Duplicated resources. Poorly written export pipelines embed the same logo, header image, or font once per page instead of once per document. A 200-page report with a letterhead graphic repeated 200 times balloons for no reason at all.
• Revision history and incremental saves. The PDF format allows edits to be appended to the end of the file rather than rewriting it. Convenient for fast saves, but after many editing sessions the file carries every previous version of itself as dead weight. A full rewrite (sometimes called "save as" or optimization) discards it.
• Attachments. PDFs can carry embedded files — spreadsheets, other PDFs, even video. People forget these exist because most viewers hide them in a side panel.

Techniques, Ranked by Impact

1. Downsample images to a purposeful resolution

This is the big one. The right resolution depends entirely on the document's destination: around 150 DPI is plenty for on-screen reading and email, while print work generally wants 300 DPI. Downsampling a 600 DPI scan to 150 DPI cuts the pixel data by a factor of sixteen before any encoding tricks even enter the picture. The key word is purposeful: pick the target based on where the file is going, not on a vague fear of quality loss.

2. Choose the right encoding: JPEG vs. lossless

Photographs and scans compress dramatically better with JPEG (lossy) encoding, and at reasonable quality settings the difference is invisible in normal use. But JPEG is the wrong choice for screenshots, line art, and diagrams with sharp edges — it produces visible ringing artifacts around text and hard lines. Those belong in lossless formats like Flate (the ZIP-style compression PDFs use internally). Good compression tools make this call per image; bad ones JPEG everything and hope.

3. Subset embedded fonts

If a document only uses 60 glyphs of a typeface, there is no reason to ship all 2,000. Font subsetting strips the unused glyphs while keeping rendering pixel-identical. The savings are modest next to image work — typically tens to hundreds of kilobytes per font — but on text-heavy documents with several embedded families it adds up, and there is no quality trade-off whatsoever.

4. Strip unused objects and flatten

A structural rewrite removes orphaned objects, discards the incremental-save history mentioned earlier, deduplicates repeated resources, and can flatten form fields and annotations into plain page content. On a file that has been through many editing rounds, this alone sometimes halves the size — and it is exactly the kind of cleanup a dedicated PDF compression tool performs alongside image re-encoding.

5. When the file must stay pristine: split or trim instead

Some documents cannot tolerate lossy compression — signed contracts, print-ready artwork, archival scans. In those cases, stop trying to compress and change the problem. If an upload portal caps file size, split the PDF into parts and send them separately, each at full fidelity. And before compressing anything, ask whether every page is even needed: a 60-page export where the recipient needs pages 3–9 is better served by deleting the unneeded pages than by degrading all 60.

Matching the Workflow to the Destination

The right target size is set by wherever the file is going, so work backward from that:

Students juggling portal uploads and shared coursework hit these limits constantly — our roundup of PDF tools for students covers the broader toolkit beyond compression.

What Not to Do

Two popular "fixes" deserve a warning, because they destroy something you cannot get back:

• Screenshotting pages and rebuilding the PDF. This converts crisp, infinitely-zoomable vector text into low-resolution pixels. The result is often larger than the original, is blurry when zoomed, and — critically — is no longer searchable or selectable. Copy-paste, screen readers, and text search all stop working.
• Print-to-PDF round trips. "Printing" a PDF back to PDF through a virtual printer sometimes shrinks it, but many print pipelines rasterize text and vector graphics along the way. You lose text quality, accessibility, and internal links, and you frequently lose bookmarks and metadata too.

The pattern behind both mistakes: they convert structured content into pictures of content. Real compression keeps the structure and shrinks only the heavy parts.

One note on tooling: compression requires parsing and rewriting your entire document, so where that happens matters. Andromeda's compressor does all of its work in your browser — the file never leaves your machine, which is worth caring about when the oversized PDF is a contract, a medical record, or anything else you would not forward to a stranger.

FAQ

Will compressing a PDF make the text blurry?

Not if the text is real text. Proper compression re-encodes images and leaves vector text untouched, so it stays razor sharp at any zoom level. The exception is scanned documents, where the "text" is itself an image — there, aggressive downsampling will soften it, which is why around 150–200 DPI is the sensible floor for scans you still need to read.

How small can I expect my PDF to get?

It depends entirely on what is inside. Image-heavy files with high-DPI photos routinely shrink by 80–90%. Text-only documents from a well-behaved generator might shrink 10% or not at all — they were never bloated to begin with. If a compressor promises a fixed ratio regardless of content, be skeptical.

Is it safe to compress a signed or certified PDF?

Compression rewrites the file, which invalidates digital signatures — the signature is a fingerprint of the exact bytes. If the signature matters, do not compress. Split the document or send it through a file-transfer service at full size instead.

Why did my PDF get bigger after editing it?

Almost certainly incremental saves: your editor appended the changes to the end of the file instead of rewriting it, so the file now contains its own previous version. A compression pass performs a full rewrite and discards that history, usually recovering all of the growth and then some.