Why crooked scans are a problem
When paper is fed at an angle or photographed by hand, every line of text slopes. The document looks careless, is harder to read, prints badly, and text recognition (OCR) makes more mistakes on tilted lines.
How it works
Choose “Estimate automatically”, and each page is checked for the angle of its text lines; tilts of up to ±10° are corrected. Or choose your own correction angle in degrees (clockwise) when all pages are tilted the same way. Each page is rebuilt as an image, slightly scaled so that no corner is cut off. Up to 100 pages are processed.
What changes in the file
Because pages become images, any selectable text, links and form fields in the original are not kept. For scanned documents that had no text anyway, nothing is lost — and you can run OCR PDF afterwards to add searchable text to the straightened pages.
Common problems and how to fix them
A page is still slightly tilted: automatic estimation works best with clear printed text lines; use a manual angle of 1–3 degrees. A page turned the wrong way: you entered the angle in the wrong direction — try the negative value. Pages with mostly pictures are not straightened well: correct them manually.
Good habits
Straighten before running OCR, keep the original scan, and next time align the paper with the scanner guide or hold the phone parallel to the page.