Black Box vs Redaction: What Actually Hides Text
Short answer
A black box draws a rectangle over text and the text stays in the file — selectable, searchable, copyable, recoverable by anyone who deletes the box. Redaction deletes the content itself, sweeps the hidden data around it, and verifies the result. Cropping is closer to the box than to deletion: what leaves the visible page stays inside the file. If the document will be seen by anyone you do not fully trust, only deletion counts.
Three ways to make text disappear from view
“Make it look gone” and “make it gone” are different jobs, and PDF tools rarely say which one they do. Here is what each approach actually leaves behind in the file.
The black box (an overlay)
A filled rectangle — or an annotation, or a text box with a colored background — sits on top of the words. Visually the line is unreadable. In the file, nothing changed: the original characters are still text objects at their original coordinates. Delete the rectangle in almost any editor and the words come back. Copy-paste, search, screen readers and text extraction all still find them. This is what most quick editors do, because it requires no font work and never fails.
Cropping
Cropping shrinks the visible page — the crop box — so content outside it is not shown. The content is still in the file, byte for byte; “cropping out” a name is concealment at viewing size. PDFMason’s crop tool says this twice: before you apply, and after. Printing often includes only the visible area, which makes cropping feel safe; printing is not deletion, and the file you send is the one that still has everything.
Redaction (deletion, then verification)
Real redaction removes the marked content from the file: text objects and image pixels are deleted, document metadata, embedded attachments and unreachable leftovers from earlier saves are swept in the same pass, and the finished copy is re-opened and checked before you get it. PDFMason for Android runs the whole job on the phone and keeps your original file untouched — the verified result is a new copy.
The table version, for the three questions people actually ask:
| Question | Black box | Cropped | Redacted | |
|---|---|---|---|---|
| Is the text still in the file? | Yes | Yes | No | |
| Does search still find it? | Yes | Yes | No | |
| Recoverable by a third party? | Trivially | Yes | Verified against |
How to verify, and what the checks cannot prove
Whichever tool you used, check the saved file yourself. These checks take a minute and catch the common failure — the box that was only paint.
Search for the hidden wording
Open the saved file in a plain reader and search for an exact phrase from the “removed” text. A hit means the words are alive in the file. This is the single fastest check, and it defeats nearly every fake redaction.
Select and copy across the area
Drag a selection over and past the covered region, copy, and paste into a plain text editor. Whatever comes out is what the file still holds. Screen readers read the same thing.
Extract the text with a different tool
Run the file through a different PDF library or an online text extractor — something that did not produce the file. Tools are kindest to their own output; a stranger’s extractor is the honest witness.
Check the metadata and attachments
Look at the document properties. The same sentence you redacted from the page often sits in the title, subject or keywords — a real redaction pass sweeps these; a cosmetic one leaves them smiling at you.
A passed check is evidence, not a guarantee. Search proves a specific string is gone, not that every copy, every image region and every earlier revision is. That is why redaction tools verify programmatically — under every mark, in the metadata, across file revisions — instead of relying on spot checks.
Verification has one blind spot worth knowing: a scan redacted by drawing over the image can pass every text check while the words remain readable pixels. If the source is an image, either the pixels themselves must change (what PDFMason checks after redacting a scan) or the page must be re-recognized and re-checked as text.
Doing it properly on Android
PDFMason for Android has two entrances to the same deletion pipeline. Manual redaction: draw boxes over what you know is sensitive. Rule-assisted redaction: let the app propose candidates — phone and ID numbers, IBANs, bank cards, emails, custom terms — and tick what goes from a per-page review list. Either way, you confirm a written summary of what will happen, nothing under the marks survives, and the copy is verified before it is saved. The full workflow is on Redact PDF on Android.
For a related decision — deleting a passage you can see, versus destroying data so it cannot resurface — the guide to deleting text from a PDF separates the two jobs.
FAQ
Does putting a black box over text in a PDF hide it?
How do I check if a PDF I received is properly redacted?
Is cropping a safe way to remove content?
Why do so many tools only draw boxes?
What does redaction do besides deleting the visible text?
Can redacted text be recovered from the file?
The deletion pipeline behind all of this — marks, summary, sweep, verification — is documented on Redact PDF on Android.
PDFMason for Android is coming
The browser tools handle whole pages. The app edits the text inside them — on your device, with nothing uploaded.
- Edit the text and images already in the PDF
- Convert to Word, Excel and PowerPoint — no upload
- OCR, redact, fill and sign — everything on the phone