How are already-rotated PDFs handled?

Current rotation + applied angle. A page at 90° + 90 = 180. + 270 = 0 (back to original).

What about unselected pages?

Untouched. Content, fonts, and images of non-rotated pages remain identical to the source.

No. Only the PDF's rotation metadata changes — the text layer and selectability are preserved.

Example

File: scan.pdf (10 pages, scanned landscape)
Angle: 90°, Scope: All

scan-rotated-90.pdf (10 pages)
→ All rotated 90° clockwise

Rotation accumulates — a page already at 90° + 90 → 180. The original file is left untouched.

Usage / FAQ

Q.How are already-rotated PDFs handled?: A.Current rotation + applied angle. A page at 90° + 90 = 180. + 270 = 0 (back to original).
Q.What about unselected pages?: A.Untouched. Content, fonts, and images of non-rotated pages remain identical to the source.
Q.Does rotation break text search / selection?: A.No. Only the PDF's rotation metadata changes — the text layer and selectability are preserved.

Fun facts

PDF page rotation just toggles the page object's /Rotate key — only 0/90/180/270 are valid. Actual content coordinates stay; the viewer applies rotation at display time. That's why rotation is virtually free.
ISO 32000-1 §14.8 (Pages — Rotate)
Landscape-printed PDFs are surprisingly common — sheet feeders that pass paper sideways. The receiver shouldn't rotate the monitor; rotate the PDF instead.
Wikipedia — Image scanner
OCR results and selectable text survive rotation because PDF text is stored as vector coordinates, independent of rotation metadata. The viewer only applies the rotation at render time.
pdf-lib