Quick answer
The short version
For an existing scan, remove blank or accidental pages and try balanced compression before a stronger setting. Inspect handwriting, stamps and fine print in the output. If you can rescan the paper, grayscale at a sensible document resolution often produces a better small file than aggressively compressing an oversized colour scan.
Compress PDFStep-by-step
A reliable workflow
- 01
Keep the original scan and note whether colour carries important information.
- 02
Remove blank backs, accidental desk images and duplicate pages.
- 03
Run balanced compression and compare the output size with the original.
- 04
Inspect small type, handwriting, seals and faint pencil marks.
- 05
Retry from the original with high compression only if the first result is still too large.
Why scans grow quickly
A scanned page is normally stored as an image covering the entire sheet. Colour channels, high resolution, camera noise and a visible background all increase the data needed for each page. Ten oversized colour scans can outweigh hundreds of text-native pages.
- Grayscale is suitable only when colour is not meaningful.
- Cropping empty borders can reduce pixels without blurring text.
- Straight, evenly lit scans compress more predictably than phone photos with shadows.
Compression and OCR solve different problems
Compression makes the file smaller; OCR adds searchable text derived from the page image. OCR can improve findability and accessibility, but it does not guarantee a smaller file or perfectly recognised text.
If searchability matters, compress a readable copy and then use OCR. Check names, numbers and tables against the page image before relying on extracted text.
Before you continue
Limitations
- Strong compression may erase faint handwriting or make small characters ambiguous.
- Compression does not correct missing pages, poor focus or severe shadows.
- OCR accuracy varies and recognised text requires review.
Questions
Troubleshooting and common questions
Should I convert a scan to black and white?
Only when colour and grayscale details are not needed. Thresholded black and white can make faint marks disappear.
Will OCR reduce the scan size?
Not necessarily. OCR adds a text layer and is intended for search and selection, not primarily for compression.