PDF Studio › Guides › Why some PDFs will not get smaller
Why some PDFs will not get smaller
You run a PDF through a compressor and it comes back barely smaller, or not smaller at all. That is not the tool failing: it is a sign that the size is coming from something other than pictures. Here is how to find out what, and what actually helps.
Where the size in a PDF comes from
Almost everything in a PDF falls into one of four groups:
| Part | Typical size | Can a compressor shrink it? |
|---|---|---|
| Pictures: photos, scanned pages, screenshots | Hundreds of KB to several MB per page | Yes, a lot. This is what compressors are built for |
| Fonts embedded so the text looks right everywhere | A few KB to several MB per font | Usually not without the original document |
| Text and vector drawings: letters, lines, charts, maps | Tiny for text; large for detailed drawings | Barely; it is already compact |
| Leftovers: attachments, old revisions, form data, unused objects | Anything | Only by rebuilding the file |
A typed page is small. Our own 5-page test report with a table weighs 5.2 KB in total, about 1 KB per page. So when a text document is several megabytes, the text itself is almost never the reason.
Step 1: check what kind of PDF you have
- Try to select a wordIf you can select and copy text, the page is real text. If the whole page highlights as one block, or nothing selects at all, it is a scan (a picture of paper).
- Divide the size by the pagesMore than 300 KB per page almost always means pictures. A few KB per page is already about as small as it gets.
- Run it through Compress PDF once on BalancedCompress PDF tells you how many images it found and optimised, for example "3 of 3 images optimized", or "No images found — only structure was optimized". That message is the fastest diagnosis there is.
Case 1: it is mostly pictures, but not small enough yet
This is the easy case. Choose a stronger setting or a size target. If you need a specific limit, pick the largest limit the form allows: the tool stops at the first setting that fits, so a looser target gives a clearer result. If even the strongest setting is not enough, the file simply has too many pictures for the limit. Split it into parts and upload them separately if the form allows, or remove pages that are not needed.
Case 2: large embedded fonts
To look identical on every device, a PDF carries its fonts. Good PDF makers include only the letters actually used (a "subset"), which costs a few kilobytes. Some programs embed the entire font instead. Fonts for Hindi, Chinese, Japanese and other large scripts can be several megabytes each, so a one-page letter can end up at 8 MB.
What helps: go back to the original document and export again with a smaller setting. In Microsoft Word, Save As > PDF with Minimum size (publishing online) selected. In LibreOffice the export dialog has options for embedding fonts. Printing the document to PDF from the original app often produces subset fonts too. A compressor working on the finished PDF cannot safely remove font data without risking missing letters.
Case 3: detailed vector drawings
Maps, architectural drawings, engineering plans and charts with thousands of data points are made of lines, not pictures. They stay perfectly sharp at any zoom, and they can be large. There are no images to re-encode, so a compressor finds nothing to do.
What helps: if the file is only for viewing, you can turn pages into pictures and then compress those: PDF to JPG at 150 dpi, then JPG to PDF, then Compress PDF. The trade-off is real: text is no longer selectable or searchable, and fine lines lose their perfect sharpness. Keep the original for printing.
Case 4: leftovers from editing
Some programs save changes by adding them to the end of the file, so the old version stays inside. Files can also carry attachments, embedded form data or objects nothing refers to any more. None of this shows on the page.
What helps: rebuild the file from its pages. In Split PDF, type 1- as the range and tick Put all ranges into one PDF. The tool creates a fresh document and copies only the pages and what they use, so unused leftovers stay behind. Note that attachments, bookmarks and fillable form fields are not carried over either, so keep the original.
Case 5: pictures stored in a way that is risky to touch
A few kinds of images are deliberately left alone by the compressor: very small images such as logos, images used as masks for transparency, and images stored with certain lossless encodings. Changing those can produce visible damage, so the tool skips them. If they make up most of the file, the result will say it is the smallest the file can go.
A quick decision guide
| What you see | Most likely cause | Try this |
|---|---|---|
| Scanned pages, big file | Pictures | Compress PDF with a size target |
| Short text document, several MB | Full fonts embedded | Export again from the original app at minimum size |
| Plans, maps or charts, sharp at any zoom | Vector drawings | Convert to images only if a smaller viewing copy is acceptable |
| File grew after editing | Leftovers and old revisions | Rebuild with Split, range 1-, one PDF |
| Many pages, each already small | Just a long document | Split into parts for upload |
Questions people ask
Does compressing twice make a file smaller?
Rarely by much. The second run re-encodes pictures that are already compressed, which mostly lowers quality. Go back to the original and choose a stronger setting instead.
Why did my file get slightly bigger?
It happens with files that were already tightly compressed. The tool always keeps your original, so just use that one.
Is it safe to compress official documents?
Text and layout are not changed. Still, check the preview for small print, such as registration numbers, before you upload.