Scanned PDF vs text PDF: why you sometimes cannot copy the text

You open a PDF, try to copy a paragraph, and nothing selects — the cursor just draws a box, as if you were dragging across a photograph. Meanwhile another PDF lets you select, search and copy freely. Both have the same file extension. What is the difference?

One of them contains text. The other contains pictures of text. Almost every confusing PDF behaviour traces back to this distinction.

What a text layer is

A born-digital PDF — one exported from a word processor, a browser, or an invoicing system — records each character as a character: which glyph, which font, which position on the page. Your PDF reader draws those characters when it renders the page, which is why you can select them, search them, and copy them out. The characters are in the file.

A scanned PDF records none of that. A scanner is a camera; its output is an image. The PDF wrapping that image knows it is displaying a picture that happens to depict words, but it has no idea which words. There is nothing to select because, as far as the file format is concerned, there is no text present at all.

How to check in five seconds

1. Drag across a sentence. Text highlights character by character in a text PDF. In a scan you get a rectangular selection box, like selecting a region of a photo.
2. Press Ctrl+F and search for a word you can see. Zero results for a word that is plainly on the page means there is no text layer.
3. Drop it into the PDF to text extractor. It reads whatever text layer exists. If the document is a scan, it will say the document looks scanned rather than silently returning an empty box.

What OCR does — and what it does not

OCR (optical character recognition) is the bridge between the two kinds of PDF. It examines the pixels of a scanned page, recognises letter shapes, and writes a new text layer containing its best guess, usually positioned invisibly behind the image so the page still looks like the original scan but becomes selectable and searchable.

Two properties of OCR are worth understanding:

It guesses. Extraction from a real text layer is exact — the characters are simply read out of the file. OCR is pattern recognition, and it makes mistakes: rn read as m, 0 as O, errors that multiply with poor scan quality, unusual fonts, or handwriting. Modern OCR is good; it is not exact.

It needs resolution. OCR engines work best on scans at 300 DPI. A heavily compressed scan may be unreadable to an OCR engine even when a human can still make it out — which is one more reason to run OCR before compressing a scan, never after.

Why extraction tools return nothing on scans

A text extractor is not doing OCR. It reads the text layer that already exists — which makes it exact, fast, and private, but also means it can only return what the file contains. Point it at a scan and the honest answer is “there is no text here”, which is exactly what the extractor on this site reports rather than returning a blank page and letting you wonder whether the tool is broken.

The rule of thumb: extraction for born-digital PDFs, OCR for scans.

The grey areas

A few documents are hybrids, and worth recognising:

• Scans that were OCRed select and search like text PDFs, but copy out with recognition errors. If pasted text contains odd character swaps, you are reading OCR guesses, not the original characters.
• Text PDFs “flattened” for security — sometimes done to stop edits — have had their text layer deliberately rasterised. They behave like scans because they now are.
• Mixed documents, such as a digital contract with a scanned signature page appended. Extraction returns everything except the scanned pages.

Why this matters beyond copying

The text layer is also what search indexers, screen readers, and automated document systems consume. A scanned invoice is invisible to full-text search; a scanned form cannot be read aloud to a blind user; an application processed by software may fail silently if the pages are images. When a document has any downstream life beyond human eyeballs, whether it carries a real text layer is the single most consequential thing about it.