How to Compress a PDF Without Losing Quality
Reduce PDF file size by 40-70% without visible quality loss. Learn what makes PDFs large, how compression works, and when not to compress.
Key Takeaways
- Embedded high-resolution images are the single biggest driver of PDF file size. A scanned 10-page document can exceed 50 MB before compression.
- Downsampling images to 150 DPI for screen viewing or 300 DPI for print is lossless to the human eye at normal zoom, but cuts file size by 40-70%.
- Gmail caps attachments at 25 MB and Outlook at 20 MB. Most web upload forms cap at 10 MB. Oversized PDFs hit these limits constantly.
- Stream compression and font subsetting are fully lossless. Image downsampling introduces some data loss, but it's invisible at typical viewing distances.
- Do not compress legal contracts, archival copies, or print-ready files. Compression is for sharing, not for your master copy.
Compress Your PDF
Drop your PDF below. The tool downsamples embedded images, subsets fonts, and applies stream compression entirely in your browser.
Why Does PDF File Size Matter?
Gmail’s attachment limit is 25 MB, and Outlook caps attachments at 20 MB (Microsoft Support, 2025). Most government and legal portals drop to 10 MB or lower. A single scanned document with high-res images routinely blows past all three limits.
Beyond email, storage adds up fast. A 200-employee company sharing even one 5 MB PDF per day creates a gigabyte of accumulated attachments in four months. Page load speed matters too. PDFs embedded in websites or linked from landing pages load measurably faster when compressed. Google’s Core Web Vitals documentation (2025) classifies download time as a user experience signal, and PDFs are no exception.
Compression solves the problem. A 20 MB scanned report can often reach 3-5 MB with no visible difference at normal screen viewing sizes.
What Makes a PDF Large?
The Adobe PDF specification, maintained by ISO as ISO 32000 (2022 revision), defines PDF as a container format. It can hold images, vector graphics, fonts, metadata, digital signatures, form fields, JavaScript, and embedded files all in one document.
That flexibility is exactly why PDFs vary so wildly in size.
Embedded Images
Images are almost always the dominant factor. A PDF created from a scanned document at 600 DPI stores the scan as a raw or lightly compressed image. A standard A4 page at 600 DPI in colour produces roughly 25 MB of raw image data before any PDF overhead. Ten such pages: 250 MB.
Even PDFs created from word processors can carry surprise image weight. A Word document with a single high-res product photograph or an unoptimised logo can produce a PDF three to five times larger than necessary.
Embedded Fonts
PDF files embed fonts to guarantee that the document looks identical on any device. Full font embedding carries every glyph in the font, including characters your document never uses. A single professional typeface can weigh 500 KB to 2 MB (Microsoft Typography documentation, 2024). A document using three custom fonts might embed 4-6 MB of font data.
Font subsetting solves this. It strips the font down to only the glyphs present in the document. A report using 350 unique characters from a 2 MB font file might subset to 80 KB. That’s a 96% reduction in font weight with zero visible change.
Uncompressed or Redundant Streams
PDF files are built from “streams”: binary blobs of content data. Image pixels, font programs, page content operators, form field data - each lives in a stream. Older PDFs and those produced by poorly-configured export pipelines often contain uncompressed streams, or streams compressed with older algorithms like LZW instead of the more efficient Flate (DEFLATE/zlib) compression.
Recompressing all streams to Flate is lossless and often cuts 10-25% from documents that haven’t been optimised before.
Metadata and Overhead
Document properties (author, title, creation date), embedded colour profiles, thumbnail previews, version history, and edit annotations all add bytes. These are usually small relative to image weight, but stripping unnecessary metadata is another safe, lossless reduction.
Citation capsule: PDF files can contain images, fonts, streams, and metadata in a single container format defined by ISO 32000 (2022). Embedded high-resolution images dominate file size, with a single A4 page scanned at 600 DPI generating roughly 25 MB of raw colour image data before compression is applied.
How Does PDF Compression Actually Work?
PDF compression combines three distinct techniques, each targeting a different type of content. According to Adobe’s documentation on PDF optimisation (2024), the most substantial size reductions come from image downsampling, followed by stream recompression and font subsetting.
Image Downsampling
Downsampling reduces the pixel density of embedded images. A photo scanned at 600 DPI for printing contains far more pixels than a screen can display. For a document read on screen, 72-96 DPI is often sufficient. For printing, 150-300 DPI covers almost every use case.
The process resamples the image using interpolation algorithms (bicubic for quality-preserving downsample, nearest-neighbour for speed). Pixels are discarded. This is a lossy operation. The discarded pixel data cannot be recovered. However, the visual result at normal viewing distances is identical. You’d need to zoom above 200% on a high-DPI monitor to see the difference between a 300 DPI and a 600 DPI embedded image.
Font Subsetting
The compressor scans every character in the document, builds a list of required glyphs, and writes a new font file containing only those glyphs. This is lossless. The output looks identical because the same glyphs render with the same outlines. Only unused characters are removed.
Stream Recompression (Flate/ZIP)
All content streams in the PDF are decompressed and then recompressed using Flate, which is the same algorithm as ZIP. No data is discarded. This is fully lossless and particularly effective on streams that were previously uncompressed or used older algorithms.
Removing Metadata and Overhead
Creation timestamps, edit history, embedded preview thumbnails, redundant cross-reference tables, and named destinations that point to nothing can all be stripped safely. The PDF renders identically without them.
What 'lossless' means for PDFs
When people ask how to compress a PDF without losing quality, they’re usually asking whether text will still be crisp and readable. The answer is yes: text in PDFs is vector data, not pixels. It renders at any zoom level without degradation regardless of compression. Image quality inside the PDF is the variable, and that depends entirely on what DPI target you choose for downsampling.
Citation capsule: PDF compression combines three techniques: image downsampling (lossy, discards pixels above the target DPI), font subsetting (lossless, removes unused glyphs), and stream recompression using Flate/ZIP (lossless). Adobe’s PDF optimisation documentation (2024) identifies image downsampling as the technique responsible for the largest file size reductions.
Lossless vs Lossy: What Actually Changes in Your PDF?
This distinction matters because “compression” means different things depending on the technique. Understanding it prevents surprises.
What’s Always Lossless
Font subsetting, stream recompression, and metadata removal produce bit-identical output for all text content. Every character, every vector shape, every annotation renders exactly as before. These operations are safe for any document.
What’s Technically Lossy
Image downsampling permanently discards pixel data. If your PDF contains photographs of a product, scans of handwritten notes, or charts with fine detail, aggressive downsampling will degrade those images.
The key word is “aggressively.” Downsampling from 600 DPI to 300 DPI is invisible in normal use. Downsampling from 600 DPI to 72 DPI will visibly blur fine print, maps, and detailed photographs.
| Compression Level | DPI Target | Typical Size Reduction | Quality Impact |
|---|---|---|---|
| Light (screen) | 150 DPI images | 30-50% | Invisible at normal screen viewing |
| Medium (balanced) | 96 DPI images | 50-65% | Minor loss in zoomed fine detail |
| Aggressive (smallest) | 72 DPI images | 65-75% | Visible blur on photos and dense text images |
For most use cases (email attachments, web uploads, review copies), the “light” or “medium” settings produce files that look perfect and email without issue.
Do not compress your only copy
Compression is a one-way operation. The downsampled pixels don’t come back. Always keep your original uncompressed PDF or source file (Word, InDesign, Illustrator) and treat the compressed version as a distribution copy. Do not overwrite the original.
Citation capsule: PDF compression is lossless for text, vectors, fonts, and metadata. It becomes lossy only when downsampling embedded images. Downsampling from 600 DPI to 300 DPI is visually identical at normal zoom. Downsampling to 72 DPI produces visible blur in photographs and fine-detail graphics. Text in PDFs is vector data and is never degraded by compression.
Step-by-Step: How to Compress a PDF
Compressing with the PDF Compressor takes under a minute.
Step 1: Upload Your PDF
Click the upload area or drag your file into the tool. The PDF loads entirely in your browser; it doesn’t upload to a server.
Step 2: Choose a Compression Level
Three settings cover most use cases:
- Screen (light): 150 DPI for images, full font subsetting and stream recompression. Best for documents you’ll read on screen and email.
- Balanced (medium): 96 DPI for images. Smaller files, still readable at 1:1 zoom.
- Maximum (aggressive): 72 DPI for images. Smallest possible output. Use when you must hit a strict file size limit and visual quality of embedded images is not critical.
If your PDF is text-only (no scanned images, no photographs), all three settings produce nearly identical output. The DPI setting only affects embedded raster images.
Step 3: Download the Compressed File
The tool processes the document, shows you the before-and-after file sizes, and lets you download the compressed PDF. Check the output against your size target before closing the browser tab.
Check size before emailing
Gmail bounces attachments over 25 MB and Outlook over 20 MB. If your compressed PDF is still over 10 MB, it’ll fail on most government portals and many intake forms. Run a second pass at “maximum” compression, or consider splitting the document using the Merge/Split PDF tool.
Platform-Specific Tips
How to Reduce PDF Size on macOS (Preview)
macOS Preview has a built-in “Reduce File Size” Quartz filter under File > Export as PDF > Quartz Filter. It works, but it’s aggressive: the filter drops image resolution to roughly 75 DPI, which causes visible blur on scanned documents and photographs.
For most macOS users, the online compressor gives better results because you control the DPI target. Use Preview’s filter only when you need a quick rough reduction and quality doesn’t matter.
How to Reduce PDF Size on Windows
Windows has no native PDF compression tool. Options include:
- Online tools (like this one): Best for occasional use, no installation.
- Adobe Acrobat Pro ($19.99/month per Adobe’s pricing page, 2025): Full control, batch processing, best output quality.
- LibreOffice (free): Can re-export PDFs with compression, though output quality varies.
For most Windows users who compress PDFs occasionally, an online tool is the practical choice.
How to Reduce PDF Size on Mobile (iOS / Android)
Neither iOS nor Android has a built-in PDF compressor. The options are third-party apps (many require subscriptions) or mobile browser access to online tools. The PDF Compressor works on mobile browsers. Upload, compress, and download directly to your device without installing anything.
Citation capsule: macOS Preview’s built-in Quartz filter reduces images to roughly 75 DPI, causing visible blur on scanned content. Windows has no native PDF compression tool. Adobe Acrobat Pro costs $19.99/month per Adobe’s pricing page (2025). Online compression tools are the practical choice for users who compress PDFs occasionally across all platforms.
When Should You NOT Compress a PDF?
Compression is for sharing, not archiving. Three situations call for keeping the original uncompressed file.
Legal and Compliance Documents
Contracts, deeds, court filings, and government forms often have authenticity requirements. Some legal workflows treat file size or metadata as integrity signals. A recompressed PDF might still be legally valid, but check with your legal team before compressing anything that has regulatory implications.
Archival Copies
If a document needs to be readable and fully recoverable 10, 20, or 50 years from now, keep the highest-fidelity version. Compressed copies are distribution copies. ISO 19005 (the PDF/A standard for archival), as defined by the ISO organisation, explicitly requires embedded full fonts and prohibits certain compression types to ensure long-term reliability.
Print-Ready Files
Print-ready PDFs sent to professional printers typically need 300-600 DPI embedded images, CMYK colour profiles, and bleed marks intact. Compressing a print-ready PDF to 72 DPI will ruin the print quality. Press-ready files should never be the file you send to the compressor.
Password-protected PDFs
Compression tools generally cannot modify password-protected PDFs without the password. If your PDF is encrypted, you’ll need to remove the password (using the password and a PDF editor) before compressing. The compressed output does not inherit the password. Re-apply encryption after compression if needed.
Merge PDF
Combine multiple PDF files into one document — drag, reorder, and merge with no upload and no watermark.
Frequently Asked Questions
Why is my PDF still large after compression?
The most common cause is embedded high-resolution images that weren’t aggressively downsampled. If a 30 MB PDF compresses to 22 MB, it likely contains photos or scans at 300-600 DPI. Try the “maximum” compression setting to target 72 DPI, or check if the PDF contains embedded fonts from large professional typeface packages. Compression on a text-only PDF rarely produces more than 15-20% reduction because there are no images to downsample.
Can I compress a PDF multiple times?
Yes, but with diminishing returns. Each subsequent compression run starts from an already-optimised file. The first pass produces the largest reduction. A second pass on the same settings typically saves another 1-3%. There’s no quality compound-loss for text and vector content across multiple runs. For image content, multiple aggressive downsample passes will progressively degrade quality, so keep the original and compress from scratch if you need a different output size.
Does compressing a PDF remove the password?
No. Compression tools that respect PDF encryption cannot modify the file content without the password. If you supply the password to an unlocking tool first and then compress, the output will be an unencrypted compressed file. Re-apply a password using a PDF editor if needed.
What’s the difference between compressing and optimising a PDF?
They describe the same process. “Optimising” is the term Adobe Acrobat uses; “compressing” is what most users search for. Both refer to reducing file size through image downsampling, font subsetting, stream recompression, and metadata removal. The output is the same.
How much can I realistically reduce a PDF?
It depends entirely on the content. A scanned document at 600 DPI can shrink 70-80% with medium compression settings and remain perfectly readable on screen. A PDF exported from a text editor or spreadsheet with no embedded images might only reduce by 10-20% because there are no images to downsample. According to Foxit’s PDF compression research (2024), image-heavy PDFs see average reductions of 40-70% with standard compression settings.
Conclusion
PDF file size is dominated by embedded images. Compressing a PDF means targeting those images with downsampling, while font subsetting and stream recompression handle everything else losslessly. For screen-readable documents, 150 DPI is the sweet spot: files shrink 40-60% and look identical at normal zoom.
Three rules keep you out of trouble. Always compress a copy, not your original. Don’t compress print-ready files or legal archives. Check the output size before you try to email it.
The PDF Compressor processes everything in your browser. Your document doesn’t leave your device. Upload, choose a compression level, download, and send.
Related articles
Image Compression Explained: Lossy vs Lossless
Learn how lossy and lossless image compression work. JPEG quality 80 cuts file size 75% with no visible loss. Real size comparisons and format guide.
File Size Reference: How Big Is a Megabyte, Really?
Visual file size guide: 1 MB holds 500 text pages or 1 compressed photo. Byte to petabyte explained with real-world comparisons and storage planning tips.