Black boxes that are semi-transparent (partial opacity annotation) or that are drawn as annotations rather than as page content can sometimes be seen through. More importantly, even opaque black boxes leave the underlying text in the PDF's content stream — it is invisible but extractable. Select all and copy (Ctrl+A, Ctrl+C) on a page with black-box "redaction" and paste into a text editor — you will often see the supposedly hidden text appear in the pasted output. True redaction requires removing the text from the content stream itself.

Whiteout (painting over content with a white rectangle or applying a white box annotation) hides content visually but leaves it in the PDF — identical in vulnerability to black-box covering. Redaction removes the underlying content from the page stream permanently. For any document requiring genuine privacy protection, only true redaction is appropriate. For cosmetic correction of mistakes visible only on screen, whiteout may be acceptable — but never for sensitive content.

Text in a scanned PDF is stored as pixels in a raster image — there is no separate text layer to delete. To redact scanned PDFs: (1) use OCR to identify the text locations if needed; (2) overwrite the image pixels in the redaction region with solid fill at the image resolution. PyMuPDF and Acrobat Pro both support pixel-level image redaction. However, be careful with high-resolution scans — zooming in on a redacted area can sometimes reveal ink impressions from the original text if the redaction fill is not a complete overwrite of the pixel region.

PDF form fields can store data independently of the page content stream — in the field value dictionary. A redaction that only modifies the page content stream will not remove the data stored in the field's value. To fully redact form field data: (1) apply page content redaction to remove the rendered field value from the page display; (2) flatten the form (convert fields to static content) to eliminate the field dictionary; (3) sanitize to remove any remaining field data from the AcroForm dictionary. Acrobat's Sanitize Document handles this as part of the full sanitisation pass.

No — by design, applied PDF redaction is irreversible. Once you apply redactions and save the file, the content is permanently removed from the page stream. This is the point of redaction. Always keep an unredacted copy of the original document in a secure location before applying redactions, and work on a copy rather than the original. If you accidentally redacted the wrong content, restore from the backup copy and redo the redaction process.