Why duplex scans are full of blank pages
Scan a stack of single-sided pages on a duplex feeder and the reverse of every sheet gets captured too — which is blank paper, scanned. A fifty-page document becomes a hundred-page file, half of it nothing. It happens with photocopiers left on two-sided mode, with phone scanning apps that capture a page before the next one is placed, and with any workflow where a machine scans both sides without being told which ones actually have content. This tool finds those pages and removes them.
Blank is a judgement, not a fact
No scanned page is ever pure white. The glass carries dust, the sensor has noise, the paper has a slight texture that shows up as faint grey. What separates a genuinely blank page from one with content is how much of it is covered in something darker than that background noise — which is a threshold, not a binary answer, and different scanners produce different amounts of noise.
This tool renders each page small, measures what fraction of it is meaningfully darker than white, and compares that fraction against a sensitivity setting you control. Raise it if genuinely blank pages are being kept because your scanner is noisy; lower it if a page with only a small stamp, a signature or a faint watermark is being removed by mistake.
What it will and will not catch
| Page content | Removed at default sensitivity? |
|---|---|
| Genuinely blank paper, scanned | Yes |
| A page header or footer only, rest empty | Usually not — headers carry enough ink to register |
| A single small stamp or signature | Usually not |
| A faint watermark across the whole page | Sometimes — raise sensitivity if it is being kept |
| A page that is mostly a photograph or chart | No |
| A page rotated sideways with content | No — rotation does not affect ink coverage |
The tool lists exactly which page numbers it removed, so the result is checkable before you trust it — which matters, because a page with only a signature is easy to lose at an aggressive setting.
Getting the sensitivity right
- Start at the default and look at what got removed. If a page you expected to lose survived, lower the sensitivity slightly and run again.
- If the whole document looks blank at your setting, the sensitivity is almost certainly too high — the safeguard that keeps at least one page is there specifically for this case.
- Noisy or old scanners need a higher threshold, since their background noise reads as ink at a low one.
- Documents with faint stamps that must be kept need a lower threshold, at the cost of possibly keeping a couple of genuinely blank pages too.
Where this fits with other tools
- If the blank pages are entirely predictable — every even page, say — Delete PDF Pages with an explicit range is faster and needs no judgement call at all.
- If pages came out upside down as well as blank, run Rotate PDF first — rotation does not affect whether a page reads as blank, but it does affect whether the final document is usable.
- Once blanks are gone, Compress PDF shrinks what is left, since removing a hundred blank scanned pages was probably most of the file's weight anyway.