You have a PDF ready to send, and the form says "maximum file size 2 MB". Your file is 14 MB. This happens all the time with scanned certificates, signed agreements, quotations full of product photos, and assignments put together from phone pictures.
Compressing a PDF is easy. Compressing it without making it unreadable, or quietly breaking something the recipient needs, takes a little more thought. This guide explains what actually makes a PDF big, which kind of compression suits which file, and how to check the result before you hit send.
Why is my PDF so big?
A PDF is a container. The things inside it decide its size:
- Images. This is the usual culprit. A scanned page is really a photo of a page, and a phone camera photo can easily be several megabytes on its own. Ten scanned pages can add up to tens of megabytes.
- Embedded fonts. Fonts are often packed into the file so it looks the same everywhere. That adds some size, but rarely tens of megabytes.
- Structure and leftovers. Some PDFs carry unused objects, duplicated data or inefficient internal layout from the program that made them.
Text itself is tiny. A 50-page report that is pure text is often well under a megabyte. So the first question is always: is this PDF mostly text, or mostly images? A quick test: try to select a word with your cursor. If you can highlight individual words, the page has real text. If the whole page selects as one block, or nothing selects at all, it's an image, which usually means a scan.
Two kinds of compression
Broadly, PDF compression works in one of two ways.
Lossless optimisation rewrites the file's internal structure more efficiently and leaves the content alone. Text stays selectable and searchable, and nothing looks different. The catch is that the savings are often modest, especially if the PDF was already saved efficiently.
Image-based (lossy) compression turns each page into a compressed image at a lower resolution. On scanned documents this can shrink the file dramatically, because the original scans were often far more detailed than needed. The trade-off is that text becomes part of a picture: it can't be selected, copied or searched, and very small print can get soft.
Neither is "better". They suit different files.
Step by step: compressing a PDF
Here's the process using the free PKD Ventures Compress PDF tool, which offers both approaches as separate modes. The same thinking applies to any decent compressor.
1. Keep a copy of the original. Always compress a copy. Image-based compression can't be undone, and you may need the full-quality version later for printing or records.
2. Add your PDF. The tool accepts one PDF at a time, up to 50 MB.
3. Pick a mode based on what's inside the file.
- Optimise (lossless) for text-based PDFs: reports, invoices exported from accounting software, documents saved from Word. Text stays selectable. Expect small savings.
- Strong (image-based) for scans and image-heavy PDFs. It converts pages to compressed images, so text becomes non-selectable. Strong mode handles up to 100 pages.
The page may also show a third Ghostscript option when that server feature is available. It recompresses the images inside the PDF while keeping the text, which is a good middle ground for mixed documents with both text and photos.
4. If you chose Strong, pick a strength. There are three levels:
- Better quality (150 DPI) for documents with small print, stamps or signatures that must stay legible.
- Balanced (120 DPI) as a sensible default for most scans.
- Smallest (96 DPI) when the upload limit is tight and the document has large, clear text.
DPI (dots per inch) is how much detail is kept per inch of the page. Lower DPI means smaller files but softer text.
5. Compress and compare. The tool shows the before and after size, so you know straight away whether it worked. If the result isn't smaller, it tells you, and suggests trying the other mode.
6. Check the output before sending. Open the compressed file and zoom to 100% or more on the smallest text: the fine print, the registration number on a certificate, the signature. If anything important is hard to read, go back and use a higher-quality setting.
Picking the right mode: quick examples
| Your file | Try first | Why |
|---|---|---|
| Report or invoice exported from software | Optimise | Mostly text, so keep it selectable |
| Scanned certificates or ID documents | Strong, Balanced | Pages are already images |
| Assignment made from phone photos | Strong, Balanced or Smallest | Phone photos are far larger than needed |
| Contract with signatures and fine print | Strong, Better quality | Legibility matters more than size |
| Brochure with text and product photos | Ghostscript if offered, else Optimise | Shrinks images while keeping text |
When the file still won't shrink enough
Sometimes one pass isn't enough. A few options:
- Fix the source instead. If the PDF was built from photos, resize the photos before making the PDF. For example, shrink them with an image compressor, then combine them with JPG to PDF. Starting from right-sized images usually beats squeezing a bloated PDF afterwards.
- Send fewer pages. If the portal only needs the first page of a certificate, don't upload all eight.
- Split the submission if the form allows several attachments.
- Combine at the end. If you need one file from several documents, compress the heavy ones first, then join everything with Merge PDF. Merging copies pages as they are, so it won't undo your compression.
Common mistakes
Using image-based compression on a text PDF. Turning crisp text into pictures can make the file bigger, and it loses searchability. If your PDF is mostly text, start with lossless optimisation.
Compressing the only copy. Lossy compression throws detail away for good. Keep the original.
Not checking the fine print. A file that fits the limit but has an unreadable registration number will just get rejected. Zoom in and check.
Compressing again and again. Running a lossy compression on an already-compressed file mostly adds blur. If one pass at Balanced isn't enough, go back to the original and try Smallest instead.
Expecting searchable text after Strong mode. If the recipient needs to copy text or search the document, image-based compression isn't the right choice. Use lossless or Ghostscript-style compression, or send the text-based original.
Forgetting passwords. Password-protected PDFs often can't be processed by compression tools. Remove the protection first (if you're allowed to), then compress.
For developers: doing it on the command line
If you process PDFs in bulk, Ghostscript is the standard open-source tool for this, and its presets map neatly onto the choices above:
# Balanced: good for most documents
gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.5 -dPDFSETTINGS=/ebook \
-dNOPAUSE -dQUIET -dBATCH -sOutputFile=output.pdf input.pdf
# Swap /ebook for /screen (smallest) or /printer (higher quality)
Check the output size with ls -lh output.pdf, and always open a sample to confirm it's still legible.
A simple routine
- Copy the original.
- Check whether the PDF is text or scanned images.
- Text: use lossless. Scans: use image-based at Balanced.
- Compare sizes, then zoom in on the smallest important text.
- Still too big? Go back to the original and step down one quality level, or fix the source images.
Closing note
This guide comes from PKD Ventures, a Kanpur-based team that builds websites, business apps and custom software. We also keep a set of free tools for everyday PDF, image and document jobs at pkdventures.in. If your business handles a steady stream of uploaded documents, it's often worth building compression and size checks into your own forms, so customers never hit the "file too large" wall in the first place.
Top comments (0)