Scanned or Digital? How to Tell What Kind of PDF You Have
A scan and a digital original can look identical and have almost nothing in common inside. Ten-second tests to tell them apart, and why the answer decides which checks are worth running.

A lender sends a file back to the broker with a one-line note: "Statements must be originals downloaded from the bank, not scans." The broker opens the PDF. It looks like a bank statement. It is slightly crooked, there's a faint grey shadow down the left edge, and the text is a touch soft. Is that a scan? Does it matter?
It matters more than most people expect, because the two kinds of file can be checked for completely different things. A lot of bad calls on documents, in both directions, come from applying the wrong test to the wrong kind of page.
Key takeaways
- A born-digital PDF holds real text and fonts. A scanned PDF holds a photograph of a page.
- You can usually tell which you have in ten seconds by trying to select a word and zooming to 400%.
- Many scans carry an invisible text layer from character recognition, which makes them searchable and also makes them look digital to a quick test.
- The combination to look at twice is a scanned page with live, visible type on top of it.
Two files that look the same
When your bank generates a statement, software writes a PDF directly. The file contains the letters themselves, the fonts they are set in, and lines drawn as geometry. Zoom in as far as you like and the edges stay sharp, because nothing is being magnified. The page is redrawn at every size.
When someone puts paper on a scanner, or photographs it with a phone, the PDF contains one large image per page and nothing else. There are no letters in it. There are only pixels that happen to be arranged like letters.
On screen at normal size these can look identical. Inside, they have almost nothing in common.
The ten-second tests
Try to select a word. Click and drag across a line. If individual words highlight, the file contains text. If nothing highlights, or the whole page selects as a single block, you are looking at an image.
Zoom to 400%. Digital text stays crisp. Scanned text goes soft and blocky, and you'll see specks, paper grain and compression fuzz around the letters.
Search for a word you can see. If Ctrl+F finds it, there is a text layer.
Look at the fonts. In most viewers, File → Properties → Fonts lists the embedded typefaces. A list of fonts means digital text. An empty list means pictures.
Check for a tilt. Scanned pages are almost never perfectly square. Text that climbs by a fraction of a degree was on a glass plate at some point.
The case that fools the quick tests
Most office scanners now run character recognition and store the result as an invisible layer on top of the image. You can select words. Search works. By the first and third tests above, the file looks digital.
It isn't. Zoom in and the letters are still soft, and the selection highlight often sits slightly off the visible characters. The font list is empty or names a single stand-in font. This is a searchable scan, and there is nothing wrong with it. It does mean the text you copy out is the scanner's best guess at what the page says, which is not always what the page says.
That gap matters. If the invisible layer and the image disagree, software reads one thing while people read another. We covered that problem in how to find hidden text in a PDF.
Why it changes what you can check
On a born-digital file you can ask structural questions. Which program produced it? Are the fonts consistent from the first page to the last? Was it saved more than once, and is an earlier version still inside? These are the checks in our guide to telling whether a PDF has been edited, and they are strong, because they rest on how the file was written rather than on how it looks.
On a scan, nearly all of that is gone. Scanning flattens history. Whatever happened to the document before it reached the glass is now just pixels. What you can examine instead is the image: whether the noise and grain are even across the page, whether one region is sharper or cleaner than its surroundings, whether the same patch of texture appears twice. Those are different tools with different error rates.
This is also why people who alter documents like scans. Print the edited PDF, scan it, and the tidy structural evidence disappears. It doesn't make the document clean. It moves the question from the file to the pixels, where the answers are less certain. A reviewer who knows that treats "it's only a scan" as a reason to look harder, not a reason to relax.
The combination to look at twice
Sometimes a page is a scanned image with real, selectable, visible text placed on top. There are honest reasons for it. Someone typed into a scanned form. A stamp or a page number was added. A signature block was dropped in.
There are also dishonest ones. A scanned bank statement where one balance is live text in a slightly different weight from its neighbours is a page where somebody typed over a picture. The rule of thumb we use: added text in margins, headers and signature areas is routine. Added text sitting on a value, a date or a name deserves a question.
What to ask for instead
If the document started out digital, ask for the digital original. A bank statement downloaded from online banking is a stronger thing than a scan of its printout, and any applicant can produce one in a couple of minutes. A request for "the PDF from your banking portal, not a photo" is reasonable, and a reluctance to supply it tells you something.
Be fair about the opposite case too. Plenty of real documents only exist on paper. A scan of a genuinely paper document is not suspicious. It is just harder to examine. Our free PDF X-Ray reports which kind each page is, working from the page contents and not from whatever the file's metadata claims, so you know which set of questions applies before you start asking them.
Frequently asked questions
Is a scanned document less trustworthy than a digital one?
Not in itself. It is less checkable. A scan of a real paper document is completely normal. A scan of something that should exist as a digital original, such as a statement from online banking, is worth asking about.
Can a scanned PDF be edited?
Yes. The image can be changed in any photo editor, and text can be placed over it. Edits to a scan leave different traces from edits to digital text: uneven noise, mismatched sharpness, repeated patches.
Why can I select text in a PDF that is obviously a scan?
Because the scanner ran character recognition and stored its reading as an invisible layer over the image. You are selecting that layer, not the picture.
Does the Producer field tell me whether a PDF was scanned?
It tells you what the file claims. That field is a line of text anyone can change, and a scanner's name in it proves nothing. Look at what the pages are made of.
Check the PDF you are holding
Run a free PDF X-Ray in your browser — it recovers text from the file’s earlier revisions, so you can see what a value was before it was changed. No account needed.
Keep reading
How to Find Hidden Text in a PDF
A PDF can carry text nobody reading it will ever see: white on white, under a box, or switched off. How to find it in thirty seconds, and how to tell the innocent kind from the kind written for a machine.
How to Tell if an Image Is AI-Generated
There is no single test, and detector scores are guesses. A ranked set of checks that still work, from content credentials to the physics a generator gets wrong, and the question that matters more than 'is it AI?'
What the Standards Prove: PAdES, eIDAS, C2PA and PDF/A
Only one of these detects a change to a document. What each standard actually asserts, where signatures stop covering the file, and why a missing set of content credentials proves nothing at all.