Why Scanned PDFs Are Enormous — and How to Fix It

Three sheets of A4 come out of the scanner as a 48 MB PDF. The same three sheets, scanned by a colleague, are 900 KB. Neither person changed anything deliberately. The difference is entirely in settings that most scanning software hides behind an Advanced button.

The arithmetic behind a huge scan

A scan is a grid of pixels, and the grid grows with the square of the resolution. An A4 page is 8.27 × 11.69 inches, so:

Doubling the resolution quadruples the data. Now add the second multiplier: colour depth. A colour pixel needs three bytes, greyscale needs one, and pure black-and-white needs one bit. Scanning a black-ink-on-white-paper document in full colour therefore costs roughly 24 times more than it needs to — and colour scans of white paper compress badly too, because the sensor records thousands of slightly different off-whites where a greyscale scan would record one flat tone.

Settings that actually matter

300 DPI is the honest default for documents. It is the threshold above which OCR accuracy stops improving meaningfully, it prints cleanly, and it survives a moderate zoom on screen. 600 DPI is for reproducing images, not for reading words.

Two settings that quietly ruin files

"Scan as image" on a text document. Many multifunction printers default to producing a PDF that is one photograph per page with no text layer at all. The document looks fine and is completely unsearchable. If your scanner offers searchable PDF or OCR, turn it on: you get a text layer behind the image, at almost no size cost, and the file becomes findable and quotable.

Automatic colour detection. A single coloured logo in the letterhead can flip an entire multi-page scan into colour mode. If the document is essentially monochrome, set greyscale explicitly rather than trusting the detector.

Fixing a scan you have already been sent

Often the file arrives from someone else and rescanning is not an option. In that case the order is:

1. Try lossless compression first. It usually recovers only a few percent on a scan, but it costs nothing and occasionally surprises you when the scanner produced a badly structured file.
2. Re-render to a target size. Shrinking is unusually effective on scans, precisely because the original is over-resolved. Cutting a 600 DPI colour scan to a 1600-pixel page at quality 75 typically removes 90% of the file and looks unchanged on screen.
3. Check the small things afterwards. Stamps, handwritten initials, and figures in tables are what suffer first.

Be aware that re-rendering discards any OCR text layer. If the scan was searchable, it will not be afterwards. When searchability matters more than size, keep the original and split the document instead.

Photographing instead of scanning

A phone camera is a 12-megapixel sensor pointed at a sheet of paper — comparable to scanning at roughly 400 DPI, with worse lighting and a perspective problem. Phone scanning apps compensate by detecting page edges, correcting the angle, and flattening the background to white, which both improves legibility and shrinks the file dramatically. Using the raw camera photos instead produces files several times larger for a visibly worse result.

If you do end up with a folder of photographs, merging them into one PDF in the right order is usually the last step. Our guide on turning phone photos into a clean PDF covers the details.