How to Reduce the Size of a Scanned PDF
Scanned pages are photographs of paper, so the file is heavy before you touch it. Shrink it afterwards, or, if you can rescan, scan smaller to begin with.
Why scans are big
A scanned page is a photograph of paper. Its size depends on the resolution it was scanned at, whether it is colour or grayscale, and the compression the scanner or app applied. A phone camera or a scanner set to high resolution and full colour produces pictures that are far more detailed than reading a page needs.
Fix it at the source
If you can rescan:
- Use a moderate resolution. A page of ordinary text is commonly scanned at around 200 to 300 DPI. Much higher adds size, not readability.
- Choose grayscale for black-and-white documents, unless colour matters (a stamp, a highlighted field).
- Keep the page straight and flat. Skewed, shadowed pages compress worse and look worse.
- Don’t scan blank pages.
Shrink an existing scan
Compress PDF re-compresses the scanned pictures in place and leaves the page layout alone. It never goes below a readability floor of 120 DPI and a JPEG quality of 50. In real testing at that floor, 6.5 to 9 point scanned text stayed legible but soft; text smaller than about 5 point can become hard to read.
Real results from our tests (Safari’s engine, sizes in KB, Balanced mode / Smaller mode):
| Scan | Original | Balanced | Smaller |
|---|---|---|---|
| Certificate (JPEG 2000, huge page) | 18,830 | 5,820 | 2,780 |
| Scanned report with text layer | 1,790 | 1,190 | 850 |
| Scanned A4 certificate | 1,110 | 681 | 302 |
| Scanned poster | 1,360 | 366 | 285 |
Page count, page sizes and any hidden text layer (the invisible text an OCR program added) were kept, and the tool re-reads the output to confirm.
Where it won’t help much
- A scan that is already compact. A 299-page scan of already-compact JPEGs shrank only about 17%.
- Black-and-white fax-style scans (JBIG2 or CCITT pictures) are left as they are, because JPEG would be larger.
- Heavy JPEG 2000 scans, common in archive downloads, can take 10 to 30 seconds to decode. On a phone, a page of around 120 megapixels may run out of memory.
- Password-protected files can’t be opened.
If it’s still too big
Rebuild the PDF from photos with Image to PDF, which can keep the result under a size limit, or read PDF too large to upload for the other options.