WalkingPDF

guide

Why redactions fail, and how to test a file before you serve it

A redaction is not a black rectangle. It is the removal of the underlying text, and most failed redactions are rectangles drawn over text that is still in the file. The content survives in four ways: the text object is still present under the shape, the text is in the invisible layer, the metadata or annotations carry it, or an earlier revision of the page still holds it. Each is detectable by reading the file.

The rectangle is a drawing, not a deletion

A PDF page is a list of drawing operations: place this glyph here, draw this rectangle there. Covering a line of text with a filled rectangle adds one more operation. The glyphs are still in the content stream, in the same positions, in the same order. Selecting the text, copying it, or extracting it returns what is underneath.

Four ways redacted content survives

  • The text object remains beneath the shape, so text extraction returns it even though the page looks black.
  • The text was rendered white, or with a zero-size font, so it is invisible on screen but present in the layer.
  • The content was removed from the page but left in an annotation, a comment, an embedded file, or the document metadata.
  • An earlier revision of the page is still stored in the file, and the reader can reach it.

What a check can see

Reading the file shows whether text exists in the region covered by a filled shape, whether the page carries text in an invisible render mode, whether annotations or embedded files contain text, and whether the document holds incremental revisions that still expose earlier content.

It cannot tell you whether a visible passage should have been redacted. That is a privilege question, and it belongs to the lawyer who knows the case. The check answers the narrower and more mechanical question: is the content that was supposed to be gone actually gone.

What to do when a check finds a leak

The fix is a re-production from the source with true redaction applied - text removed, not covered - and then a re-check of the output, because the failure mode is silent. A file that has been re-exported after a rectangle was drawn is not automatically clean: flattening removes some leaks and preserves others, depending on how it was done.

Frequently asked questions

Is drawing a black box over text enough?

No. The text remains extractable. Redaction means removing the underlying characters so that nothing is left to extract.

Does flattening the PDF fix a redaction?

Sometimes, and it is not a guarantee. Flattening converts the page to an image in some tools and merges layers in others; only the first reliably removes the text. The way to know is to test the output rather than assume it.

What about scanned pages?

A scanned page has no text layer, so text extraction returns nothing and a leak by extraction is less likely - but the image itself may still show the content if the redaction was applied to a different layer. The visual check still matters there.

Keep reading

guide How to QC an incoming production A production you receive has to load, read and reconcile before anyone can review it. These twelve checks run in the order that surfaces the expensive failures first: does the load file import, does every row have an image, does the numbering hold, and is every page searchable. Run them before the review clock starts, not after. guide How pages go missing from a production Pages disappear from a production in three ways: a document was never imaged, a document was imaged but dropped from the set, or the set was renumbered after the drop so the sequence looks continuous. The first two leave a gap in the Bates range. The third leaves no gap at all, which is why the range alone is not enough - the page count per document and the load file have to agree with it. hub Watermark glossary A short glossary of the words used around PDF watermarking, each defined in one or two sentences before any detail: what a watermark is, the difference between text and image marks, what diagonal, background, tiled and confidential mean, and why a dynamic watermark is not the same as a fixed one. Every term links back to the page that shows it in use. guide How to watermark a PDF To watermark a PDF, upload the file, choose a watermark style, type your wording, set the opacity to somewhere between 25% and 35%, apply it to every page and download the marked copy. The whole procedure takes about a minute in a browser, needs no software, and the original file is not changed - you keep the marked copy as a new file. guide What is a PDF watermark? A PDF watermark is text or an image placed on the pages of a document to say something about it - who it belongs to, what state it is in, or who may read it. It becomes part of the page rather than a note attached to the file, so it appears in prints, screenshots and single-page extracts, and it carries no technical protection: anyone can still open, copy and forward the document. guide Batch watermarking PDF files Batch watermarking means applying the same mark to several PDFs so they look like one set. With a browser tool you do it by settling the settings once - style, wording, opacity, colour - then running the tool once per file with those same settings, which takes under a minute each and produces consistent output without installing anything or creating an account.