You scanned ten pages on the office scanner and got a 48 MB PDF. A five-page contract ballooned past 20 MB. A book you scanned for archiving became an 800 MB file that won't open on your phone. These aren't mistakes — they're what happens when scanner settings are left on defaults.
A scanned PDF is fundamentally different from one exported from Word, which is why the usual fixes don't work on it. This guide explains what's inside a scanned page, the methods that actually shrink it, and how to configure your scanner so the problem never starts.
- A scanned page is one big image; its size depends on resolution, colour mode and JPEG quality.
- Light compression barely helps. Use strong compression or target size in Compress PDF.
- Converting to grayscale first gives clearer text at the same size.
- For the future: scan at 150–200 DPI for sharing, 300 DPI for archiving and OCR, grayscale for text.
- Need searchable text afterwards? Use OCR.
Why are scanned PDFs so big?
A PDF exported from Word stores letters as letters: the word "contract" takes a few bytes and the font draws it on screen. A scanner doesn't know there are words on the page — it captures every dot as a picture and wraps that picture in a PDF.
Three factors decide the size of that picture:
1. Resolution (DPI)
Dots per inch. An A4 page is about 8.3 × 11.7 inches (US Letter is 8.5 × 11), so:
| DPI | A4 pixels | Total pixels |
|---|---|---|
| 600 | 4960 × 7016 | ~34.8 million |
| 300 | 2480 × 3508 | ~8.7 million |
| 200 | 1654 × 2339 | ~3.9 million |
| 150 | 1240 × 1754 | ~2.2 million |
Doubling resolution quadruples the pixels, because it grows in both directions. A 600 DPI scan holds about 16 times the data of a 150 DPI scan before compression.
2. Colour mode
- Colour (RGB): three values per pixel.
- Grayscale: one value per pixel — roughly a third of the raw data.
- Black & white (bitonal): one bit per pixel, compressed with specialised methods such as CCITT Group 4 and JBIG2 that are extremely efficient for text.
Many scanners default to colour, even for a pure black-and-white contract.
3. JPEG quality
Scanners usually store page images as JPEG; the higher the quality, the bigger the file. Some default to very high quality or lossless storage, producing huge files.
Why light compression doesn't help
Light compression rewrites and cleans the PDF structure but doesn't touch images. In a scanned PDF, images are nearly everything — 99% images, 1% structure. Even if light compression cleaned the structure perfectly, the size would barely move. That's not a bug; it simply doesn't process the part that makes the file big.
Method 1: Strong compression or target size
This is the core method for scans:
- Open Compress PDF. The file stays on your device.
- Choose strong for a big reduction at reading quality, or target size if you have a specific number (a portal or email limit).
- Compress, open the result and zoom to 100% to check numbers and dates are clear.
Strong compression re-renders each page at lower resolution and JPEG quality, often shrinking a scan to a small fraction of its size. Target size automatically searches for the best quality under your number.
Method 2: Convert to grayscale first
If the file was scanned in colour and doesn't need it, convert with PDF to grayscale before compressing. Two benefits:
- Useless colour data disappears — yellowed paper, coloured shadows.
- Compression then needs less quality reduction to hit the same size, so text stays sharper.
Keep colour for: IDs and passports, documents where a coloured stamp is the only proof of authenticity, colour maps and charts, and anything the recipient explicitly wants in colour.
Method 3: Remove unnecessary pages
Blank pages are everywhere in scans, especially duplex scans: every blank back becomes a white page — and it isn't truly empty, since the scanner captures paper texture and specks.
Remove them with Delete PDF pages. In a duplex scan of 20 sheets you might find 15 blank sides — an instant one-third reduction.
Method 4: Split very large files
A 400-page scanned book may stay big even after compression and strain devices when opened. Split it into chapters or parts with Split PDF. Smaller files are easier to open, share and search, and each can be compressed without exhausting memory.
Method 5: Rescan with the right settings
If the paper original is available, rescanning correctly beats any later compression, because you start clean instead of degrading a huge image.
| Purpose | DPI | Mode | Note |
|---|---|---|---|
| Portal uploads and email | 150 | Grayscale | Smallest size, good clarity |
| Comfortable reading and normal printing | 200 | Grayscale | Excellent balance |
| OCR / text extraction | 300 | Grayscale or B&W | Commonly recommended for OCR |
| Long-term archiving | 300 | As original | Keep an uncompressed master |
| Important colour documents | 300 | Colour | Don't compress hard |
| Very small print (tables, transcripts) | 300 | Grayscale | Lower DPI confuses digits |
Also look in your scanner settings for:
- Skip blank pages.
- Document or Text mode, which boosts contrast and whitens the background.
- File quality / size: choose medium rather than highest.
No scanner? A phone is enough — see How to scan documents to PDF with your phone.
Method 6: True black & white for pure text
For purely textual documents — printed contracts, letters without coloured stamps — scanning in black & white (1-bit) gives the smallest files of all, because each pixel is one bit and specialised compressors like CCITT Group 4 and JBIG2 shrink it dramatically.
The downside: gradients disappear, so thin signatures, pale stamps and photos can vanish or distort. Use it for text only, and test on one page first.
Method 7: OCR afterwards to restore search
A scanned PDF, compressed or not, contains no searchable text. If you need to search or copy:
- Compress the file to make it manageable, keeping enough resolution for machine reading.
- Run OCR with the right language.
- Keep the extracted text alongside the PDF.
OCR needs decent resolution, so heavy compression beforehand lowers accuracy. Best practice: OCR a good-quality copy, then compress a separate copy for sending. See Extract text from images and scanned PDFs.
Methods compared
| Method | Expected reduction | Quality impact | When to use |
|---|---|---|---|
| Strong / target size | Very large | Moderate, fine for reading | Almost always for scans |
| Grayscale | Moderate, improves compression | None on text | Non-colour documents |
| Remove blank pages | Depends on count | None | Duplex scans |
| Split | No total reduction | None | Huge files |
| Correct rescan | Very large | Best result | When the original is available |
| 1-bit B&W | Largest | Loses gradients | Pure text |
What happens, in numbers
Take one A4 page scanned in colour at 300 DPI:
- Raw: 2480 × 3508 colour pixels — about 26 MB of raw data. The scanner saves it as high-quality JPEG, often 1–3 MB.
- Grayscale: raw data drops to about a third, and JPEG compresses it more efficiently.
- 150 DPI: a quarter of the pixels — 1240 × 1754.
- Medium JPEG quality: discards detail the eye won't miss on a text page.
The result is often tens to a few hundred KB per page — a tenfold reduction or more, with text that reads clearly on screen.
Long scans: books and archives
- Keep settings consistent across all pages — same DPI and mode.
- Scan in numbered batches:
Book-001-050.pdf,Book-051-100.pdf. - Keep an archive master at 300 DPI without heavy compression, plus a compressed reading copy.
- Add page numbers if the original lacks clear ones, with Add page numbers.
- Add a contents page at the front for files you'll use often.
- Copyright: scanning books for personal use is one thing; distributing them is another.
When to keep full quality
Not every scan should be compressed. Keep the high-quality original for:
- Original legal documents that may be needed as evidence.
- Old family photos scanned for preservation.
- Historical documents and institutional archives.
- Documents you'll print at high quality later.
- Files headed for OCR.
Rule: masters for keeping, compressed copies for sharing. Storage is cheap; rescanning may be impossible if the paper is lost.
Common mistakes
Compressing an already-compressed file. Every lossy JPEG pass adds artefacts. Always go back to the original.
Deleting the original. Your 300 DPI scan is your archive. Keep it and compress a copy.
Scanning at 600 DPI "just in case". Rarely needed except for fine photos or artwork; for text it wastes space and time.
Ignoring skewed pages. Skewed pages look worse after compression and hurt OCR. Fix with Rotate PDF or rescan.
Judging on a small screen. Open the compressed file on a large screen and zoom before accepting it, especially if it'll be printed.
Worked example: an HR employee file
HR scanned a 30-sheet employee file in duplex colour at 300 DPI: 60 pages, 95 MB, and it has to go by email.
- Remove blanks: 22 pages were blank backs — 38 remain.
- Grayscale: no important coloured stamps, so convert.
- Compress: target size 9 MB to pass any mail server.
- Check: open pages with ID numbers and salaries and zoom in — digits are clear.
- Archive: keep the original in the department archive and send the compressed copy.
Result: a file about ten times smaller without losing a single piece of information.
Decision table
| Situation | Action |
|---|---|
| Scan to send by email | Grayscale → target size under 10 MB |
| Scan for a 200 KB portal | Remove unneeded pages → grayscale → target 190 |
| Company archive | Keep 300 DPI master → compressed sharing copy |
| Book to read on a phone | Split into chapters → strong compression |
| Document you need to search | OCR a good copy → compress a separate copy |
FAQ
Why didn't my scan shrink with light compression?
Light compression doesn't process images, and every page of a scan is an image. Use strong or target size.
Can a scan be compressed without any quality loss?
Not meaningfully. Any noticeable reduction means lowering image resolution or quality; the goal is loss that doesn't affect reading.
What's the best scanning resolution for documents?
150–200 DPI for sending and uploading; 300 DPI for archiving, OCR and small print.
Does grayscale really reduce size?
Yes, usually noticeably — grayscale images carry less data and JPEG compresses them more efficiently.
Does compression affect OCR?
Heavy compression reduces OCR accuracy. OCR a good-quality copy, then compress a copy for sending.
Are my scans safe with WPPR?
Yes. Compression happens in your browser; files are never sent to a server. See our security page.
Bottom line
Scanned PDFs are big because they're stacks of images whose size depends on resolution, colour and JPEG quality. Light compression won't help; use strong compression or target size, go grayscale when colour isn't needed, and remove blank pages. For the future, scan at 150–200 DPI grayscale for sharing and 300 DPI for archives.
Shrink your scan now with Compress PDF — free, and nothing is uploaded.