Skip to main content
FixMyTech

How to Compress a Scanned PDF

By

Published

8 min read

Share

Short answer

Scanned pages are images, so the only thing that shrinks them is re-encoding those images. Downsample to 150 DPI and convert colour scans of black-and-white documents to greyscale — together these typically cut 80–95% while staying perfectly readable. If you still have the paper, re-scanning at 200 DPI greyscale beats fixing it afterwards.

On this page

Ordinary PDF compression advice fails completely on scanned documents, and the reason is worth understanding before you waste an afternoon on it.

Lossless compression works by removing redundancy — duplicated fonts, unused objects, an inefficient cross-reference table. A scanned page has none of that. Every page is a photograph, and a photograph of text is just as much data as a photograph of a landscape. There is nothing to deduplicate.

The only thing that shrinks a scan is re-encoding the images. That is lossy by definition, and the whole job is choosing settings where the loss is invisible.

Confirm it is actually a scan

Open the PDF and try to select text with your cursor.

  • Text highlights — it is a real text document. Compressing it is a different job; see how to reduce PDF file size.
  • Nothing highlights, or a whole page highlights as one block — it is a scan.

The per-page arithmetic confirms it. Divide the file size by the page count: anything over 1MB per page is a scan stored at high resolution. Why PDFs get large covers the full diagnostic table.

The settings that matter

Three levers, in order of how much they save.

Resolution

This is the big one, because data scales with the square of the DPI. Halving the resolution removes three quarters of the pixels.

DPI A4 page pixels Suitable for
600 4960 × 7016 Archival, fine line art — rarely needed
300 2480 × 3508 Printing, small print, OCR of poor originals
200 1654 × 2339 OCR, general documents
150 1240 × 1754 Reading on screen, emailing, office printing
96 794 × 1123 Screen only, large text

150 DPI is the working default. It reads cleanly on any screen, prints acceptably on an office printer, and it is roughly a sixteenth of the data of a 600 DPI scan.

Go to 200 DPI if the document has small print, footnotes or handwriting. Do not go below 150 for anything anyone needs to read carefully.

Colour

Scanners default to colour. Most documents are black text on white paper.

Storing three colour channels for single-channel content wastes half to two-thirds of the file. Converting to greyscale on a typical text scan is a large, free saving — free because there was no colour information worth keeping.

Keep colour only when it carries meaning: highlighting, coloured charts, stamps, signatures in blue ink where that matters evidentially.

JPEG quality

The fine adjustment. 70–80% is indistinguishable on a document; below about 60% you start seeing artefacts around letter edges, which is exactly where it hurts readability.

Do it in the browser

Our PDF compressor handles this directly, and it runs entirely on your device — which matters here more than for most files, because scanned documents are usually contracts, payslips, medical letters and forms with signatures on them.

  1. Drop the PDF in.
  2. Choose Downsample pages to images.
  3. Set the resolution to 150 DPI.
  4. Tick Convert to greyscale if the document is black and white.
  5. Leave quality at 75%.

The result shows the before and after size. If it is still too large, drop to 96 DPI and check readability; if the text has softened too much, go back up to 200 DPI.

Note the trade this makes explicitly: re-rendering the pages means any existing text layer is discarded. That is covered below.

The alternative: re-scan properly

If you still have the paper, this beats compressing after the fact. You avoid creating the excess data rather than throwing it away, and the result is visibly cleaner than a downsampled 600 DPI scan.

In your scanner software, set:

  • Resolution: 200 DPI (300 if you plan to OCR a poor original)
  • Colour mode: Greyscale, or Black & White for clean printed text
  • Format: PDF with JPEG or, for pure black-and-white text, CCITT G4
  • Quality: Medium

Black & White mode deserves a specific mention. It stores one bit per pixel rather than eight, which is dramatically smaller — a clean printed page can come out under 50KB. It looks poor on anything with photographs or shading, but for typed text on white paper it is both the smallest and the sharpest option.

Scanning with a phone

Phone camera photos make terrible PDFs — they are full-resolution colour images with uneven lighting and perspective distortion.

Use a document scanner mode instead, which detects the page edges, corrects the perspective and saves in a document-appropriate mode:

  • iPhone: Notes app → new note → camera icon → Scan Documents. Or the Files app → Scan Documents.
  • Android: Google Drive → +Scan. Many phones also have this in the camera app.

The difference is not subtle — typically 200KB per page instead of 4MB.

Keeping the text searchable

Downsampling replaces pages with images, so any existing text layer goes. If you need the document to stay searchable, the order matters.

Compress first, then OCR. Running OCR on the compressed file adds a fresh text layer sized to the new images, and the text layer itself is tiny — a few kilobytes per page.

Where to do it:

  • Windows: Microsoft Lens, or the OCR built into OneNote.
  • macOS: Preview has text recognition built in on recent versions.
  • Google Drive: upload a PDF, right-click → Open with → Google Docs. It performs OCR and gives you the text, though it loses the layout.
  • Adobe Acrobat Pro: Scan & OCR → Recognise Text, which is the best result if you have it.

OCR needs at least 200 DPI to be reliable. If searchability matters, compress to 200 rather than 150.

What to expect

Real numbers from a typical case — a 12-page contract scanned at 600 DPI in colour:

Stage Size
Original scan 48MB
Downsampled to 150 DPI, colour 4.2MB
Downsampled to 150 DPI, greyscale 1.6MB
Re-scanned at 200 DPI greyscale 1.1MB

That is a 97% reduction, and at 150 DPI greyscale the document is perfectly readable on screen and prints cleanly.

The pattern holds generally: the first downsample does most of the work, and greyscale roughly halves what remains.

What not to do

  • Do not compress the same scan twice. Each pass re-encodes already-degraded images, and the artefacts compound visibly. Go back to the original and compress once with the right settings.
  • Do not go below 150 DPI on anything anyone needs to read properly. The saving from 150 to 96 is far smaller than the saving from 600 to 150, and the readability cost is much higher.
  • Do not upload sensitive scans to free online compressors. This is the category most likely to contain personal data and most likely to be uploaded carelessly.
  • Do not use Black & White mode on anything with photographs. It renders greyscale content as harsh dithering.
  • Do not delete the original until you have checked the compressed version at 100% zoom on the pages with the smallest text.

Frequently asked questions

Why does normal PDF compression not shrink my scan?
Because lossless compression removes redundancy — duplicate fonts, unused objects, inefficient structure. A scanned page has none of that; it is a photograph, and the only way to make a photograph smaller is to re-encode it at lower resolution or quality.
What DPI should a scanned document be?
200–300 DPI for scanning, 150 DPI once compressed for sharing. Below 150 DPI small print starts to break up, and below 100 DPI text becomes difficult to read. OCR needs at least 200 DPI to work reliably.
Does converting a scan to greyscale really help?
Substantially, when the original is a black-and-white document scanned in colour — which is the scanner default. It stores three colour channels for content that only needs one, so greyscale typically removes half to two-thirds of the remaining size.
Will compressing a scan break the OCR text layer?
Re-rendering the pages discards it, so run OCR after compressing rather than before. If the document already has a text layer you need to keep, use a tool that downsamples images in place rather than rasterising pages.
Is it better to re-scan or to compress?
Re-scan if you still have the paper and the scan settings were wrong — 200 DPI greyscale produces a better result than compressing a 600 DPI colour scan, because you never create the excess data in the first place.

All PDF guides