How to black out a PDF properly

1 Four ways it goes wrong

• A box drawn over text. A filled rectangle, from any editor that can draw one. The letters are still in the page, under it.
• A black highlight. A name marked in black in a word processor or a web page, then saved as PDF: black text on a black box.
• A box over a scan. The page is a picture; the box covers part of it on screen. Take the picture out of the file and the name is on it.
• Redaction marked, never applied. Editors that redact properly first mark the areas, then apply them. Marked only, everything is still there.

Open the sample blacked-out letter: hexscope reads out the name, the phone number and the amount under its boxes, and a name under a mark never applied.

2 Black it out

1. Open the PDF in hexscope. It says straight away if boxes on it already hide something.
2. Under Black out text yourself, type a name, a number or an address — or pick Email addresses, Phone numbers, Card and bank numbers or Dates. Every place it appears is listed with a tick; untick any to keep.
3. For words that are part of a picture, a scan or a screenshot, choose Draw a box on a page: the page is shown, scan and all, and you drag a box over what should go.
4. Look at the pages as the copy will show them, then save the copy.

The copy takes the letters out of the page and draws a black box where they were, leaving the rest of each line in place. Where a box lies on a picture, the picture's own pixels there are made black. It also drops what the file says about who made it, and any earlier versions it kept.

3 Check the copy

Open the copy in hexscope — the button is right there — and it should find nothing hidden. Then try it the way someone else would: open it in a viewer, search for the name, select the line and copy it. Nothing should come back.

A copy that still shows a warning has not taken everything out: do not send it. Tell us what kind of file it was — that is a bug.

4 What it cannot do

Scans kept as JBIG2, CCITT (fax) or JPEG 2000 pictures cannot be edited here, so no copy is made rather than one that keeps them: black those out in the program that made the scan. Search finds text, not the words in a picture — draw a box over those. And it cannot take anything back from a file you have already sent.

On a computer, the command line does the same for many files: hexscope redact --text "Olena Kovalenko" letter.pdf.