How to fix a scanned PDF where the pages are sideways or out of order

You scan a stack of paper and open the result. Some pages are upside down. The order runs backwards. Every other page is blank. One page is sideways because it was landscape and the feeder did not care.

None of this needs a rescan. All of it is fixable in the file, and knowing why each happens tells you how to avoid it next time.

Why scans come out wrong

A document feeder takes whatever orientation the paper was in. Pages that were stapled and then separated frequently end up rotated relative to each other, and nothing in the scanner detects this.

Reverse order comes from feeders that pull from the bottom of the stack, or from placing the pile face-up when the scanner expected face-down. The scan is perfectly correct — it simply started from the other end.

Interleaved blanks come from duplex scanning a single-sided document. The scanner dutifully captures the back of every sheet, and the backs are empty.

Sideways pages are usually correct in the file and wrong in your reader: a landscape page scanned in portrait orientation is stored as a portrait image containing sideways content.

Fix it in the right order

1. Delete the blank pages first, so the numbering you work with afterwards is the numbering you will end up with.
2. Fix rotation next, page by page, so you can actually read the pages while reordering.
3. Reorder last, once every page is legible and the count is final.
4. Read through the finished document from start to end before sending it anywhere.

Doing it in that order matters more than it sounds. Reordering pages you cannot read, then deleting blanks, then rotating, means renumbering everything twice and losing track of where you were.

Rotation that sticks

There is a subtlety worth knowing. Rotating a page can either set a rotation flag that readers honour, or actually rewrite the page content. The flag approach is instant and lossless, and it is what most tools do.

It occasionally causes surprises: a page that looks correct on screen can print or convert in the original orientation if some other piece of software ignores the flag. If a rotated page keeps reverting in another application, that is what has happened, and flattening the document fixes it permanently by baking the orientation into the page.

Scanning better next time

• Square the stack and check every sheet is the same way up before loading it
• Turn off duplex scanning for single-sided documents rather than deleting blanks afterwards
• Scan landscape pages as a separate batch with the orientation set correctly
• Scan a two-page test before committing a two-hundred-page stack to the feeder
• Remove staples properly, since a torn corner is what makes a feeder skew a page

The two-page test is the single most valuable habit. It catches the wrong side, the wrong orientation and the wrong resolution in about thirty seconds, before those mistakes have been repeated two hundred times.

Before you fix, decide what you are fixing for

If the document is going to be read by a person, correct rotation and order are the whole job. If it is going to have text recognition run on it, rotation matters even more — recognition performs badly on sideways text and often fails outright on upside-down pages, so straightening comes before OCR rather than after.