How to compress a PDF without losing quality (and fit within system size limits)

You put together the final document, the time came to upload it to the system — and the message appears: "File exceeds the maximum limit of 5 MB." Or 10 MB. Or whatever number the system arbitrarily chose. And your PDF is 47 MB.

This is one of the most frustrating problems for anyone who works with documents: PDF compression seems simple, but done wrong it produces illegible text, pixelated images, or — worst of all — breaks the digital signature that was on the file. This guide explains what actually takes up space in a PDF, how to reduce the size intelligently, and what to do when compression isn't enough.

Why PDFs get heavy — the anatomy of a large file

Before compressing, it helps to understand what's inside a PDF. A PDF file is technically a container: it packages text, fonts, images, metadata, page structure and optionally scripts, forms and layers. The final size depends directly on which of these elements are present and how they were generated.

Images are the main culprit in most cases. A photo inserted in a report at camera resolution (6000×4000 pixels, TIFF, 24 MB) continues to occupy the same space when Word exports it to PDF — unless there is a compression step during export. Technical reports, plant drawings, dossiers with photos and reports with screenshots are the most affected.

Scanned PDFs are essentially packages of images: each page is a photograph of the paper, with no text layer. A scanned 100-page case file can easily exceed 200 MB depending on the scanner resolution.

Embedded fonts also contribute. By default, text editors embed in the PDF the fonts used — sometimes complete fonts with hundreds of glyphs, even if the document uses only a subset of them. This ensures the text appears correctly on any computer, but inflates the file.

Metadata, layers and residual objects are smaller, but present: intermediate versions, deleted objects not removed from the stream, forms with data — all of this occupies bytes that in some cases add up to more than you'd imagine.

Lossy vs. lossless compression: the distinction that matters

There is a fundamental technical distinction that every professional who works with documents should know:

Lossless compression reorganizes the data without discarding any information. The decompressed file is bit-for-bit identical to the original. For text and vector graphics, this is the only acceptable option: any change in the text data would be an alteration of the content. Algorithms such as DEFLATE (used internally in PDF) and JBIG2 (for monochrome images, such as black-and-white scanned text) operate losslessly.

Lossy compression discards data that the human eye hardly notices. For photographic images (JPEG), the visible result can be identical, but the bytes are far fewer. The right question is not "did I lose quality?", but "is the remaining quality sufficient for the intended use?". A technical report with a photo for electronic filing does not need print-quality resolution.

In practice: for most legal, technical and administrative documents, a good compression tool applies DEFLATE to text, downsamples and recompresses images in JPEG (to screen-appropriate resolution, typically 150 dpi), and removes unnecessary metadata and objects. The result is readable, verifiable and substantially smaller.

How much you can reduce — realistic expectations

The compression ratio depends on the content:

• Report with many high-resolution photos: 60% to 90% reduction is common and the result is perfectly readable on screen and in print;
• Contract generated in Word and exported as PDF (predominantly text): compression has already been applied during export; additional gain is lower, typically 10% to 30%;
• Black-and-white scanned PDF (text, without grayscale): monochrome-specific algorithms (JBIG2, CCITT G4) are very efficient; 70% to 85% reduction while preserving text quality;
• PDF with tables and vector graphics: vectors do not compress like images; gains here come mainly from removing metadata and redundant fonts, generally 10% to 20%.

If the file is 47 MB and is made up mostly of high-resolution photographs, reaching under 10 MB with no visible impact is perfectly feasible.

What to do when the file is still too large

There are situations where compression is not enough — when the system limit is very aggressive (2 MB, 3 MB) and the document simply has too many pages with images. In these cases, the complementary solution is to split the PDF into parts: the original document is sliced into volumes by page range, and each part is sent separately.

This practice is standard in electronic filing systems: the main petition goes in one file, each numbered exhibit in another. The system maintains traceability, and the professional sends the entire case within the size limits.

Recommended flow:

1. Compress the full PDF first — compress PDF;
2. If it still exceeds the limit, split by page range and number the volumes sequentially;
3. Register the SHA-256 hash of the original file in the integrity checker before splitting — you will have evidence that the parts came from an intact document.

For documents that need to be combined before sending — separate attachments that must go in a single file — the PDF merger provides the reverse path, then compress the set.

The critical mistake: compressing after signing

This is the most important warning in this article and the most common mistake made by people who discover PDF compression too late:

Any compression of a digitally signed PDF invalidates the signature.

The logic is mathematical and unavoidable: the digital signature is technically the cryptographic hash of the document encrypted with the signer's private key. When you compress the file, bytes are reorganized and eliminated — the file changes, the hash changes, and the signature no longer matches. Anyone verifying the signature will see a "document modified after signing" warning.

The rule is simple and absolute: compress before signing. The correct flow is:

1. Merge all attachments and parts of the document;
2. Remove pages that should not be in the final submission;
3. Compress the set to the appropriate size;
4. Only then collect digital signatures.

Want a deeper understanding of the relationship between PDF manipulation and digital signature validity? Read our complete guide on document cryptography, hashes and digital signatures.

How to compress PDF online without sending the file

Most online PDF compression tools work with upload: the file goes to a server, is processed there, and you download the result. For confidential documents — contracts, medical reports, balance sheets, petitions — this is an unnecessary risk and, depending on the context, may be a violation of professional confidentiality obligations or data protection law.

RoseLab's PDF compressor works differently: compression happens inside your own browser, using your computer's processing power. The file does not leave your machine at any point — this is not a privacy policy promise, it is the technical operation of the tool.

To verify this yourself: open the compressor in Chrome or Edge, press F12, go to the Network tab, load your PDF and run the compression — you will see that no request carrying the content of your file is made to external servers. Read more about privacy in PDF tools: comparing PDFs without upload — why it matters.

Step by step: compressing PDF in RoseLab

1. Open the PDF compressor — nothing to install, no account needed;
2. Drag the PDF file or click to select it;
3. Choose the compression level based on the estimated size shown — from lightest (best quality) to most aggressive (smallest file);
4. Click Compress and wait for processing;
5. Download the compressed file — the final size is shown before download so you can assess whether it is within the destination system's limit.

If the result still exceeds the limit, use the PDF splitter to slice it into smaller parts.

Complementary tools in the document preparation flow

Compression rarely comes alone. The complete preparation flow for a document typically involves:

• Merge PDFs — combine petition and exhibits in a single file before compressing;
• Split PDF — slice the result into parts within the system's size limit;
• Remove pages — delete blank pages, duplicates and unnecessary cover pages before compression;
• Convert images to PDF — turn photos and images into a document before merging everything;
• PDF to Word — extract text content when the destination is editing, not filing;
• Verify integrity — register the SHA-256 hash of both the original and compressed file, creating a documented chain of custody.

For a full view of how these steps chain together in a professional workflow, read our complete guide to comparing documents and proving what changed.

Frequently asked questions

Does compression change the text content? No — compression affects images and the internal structure of the file, not the text data. The text in a compressed PDF is identical to the original and can be copied, searched and selected normally. What may change is the resolution of embedded images, within the configured parameters.

Does the file hash change after compression? Yes — compression rewrites the file, altering the bytes, and as a result the SHA-256 hash changes. If you registered the hash before compressing, the compressed file will have a different hash — which is expected and normal. Register the final (compressed) version's hash separately for a complete chain of custody.

Can I compress a password-protected PDF? Files with an opening password require the password to be entered before any processing. Files with an editing password (but free opening) may have restrictions depending on the tool's implementation.

Does it work on Mac, Linux and mobile? Yes — since everything in RoseLab runs in the browser, it works on any operating system and device with a modern browser.

What is the difference between compressing and converting to PDF/A? PDF/A is a long-term archiving format that paradoxically tends to be larger than a regular PDF, because it embeds everything needed for faithful future reproduction (including color profiles, complete fonts). Do not use PDF/A when the goal is to reduce size for filing.