Scanned PDF vs text PDF: why conversion and search behave differently

The short answer

A scanned PDF often contains images of pages. A text PDF contains a text layer that a viewer can interpret, select and search. Some documents contain both. This distinction explains why a PDF can look perfectly readable while converting poorly to Word or returning no search results. FOLD’s documented PDF-to-Word workflow does not perform OCR. Start by inspecting the source rather than assuming that any visible paragraph is already stored as usable text.

Use selection as a first check

Open the PDF in a viewer and try selecting a sentence. If individual words highlight and copy sensibly, that page has a usable text layer. If the selection acts like a whole image or nothing can be selected, the page may be image-only. This is a practical first check, not a complete classification of the file. A document can mix scanned appendices with text-based main pages, and copied text can still have a poor reading order even when selection works.

Why the distinction affects conversion

Converting a text PDF to Word involves reconstructing paragraphs and layout from the PDF representation. That is already different from recovering the original Word document. A scanned page adds another task: recognising characters from the image before they can become editable text. Without OCR, an image-only page can produce an empty or unusable editable document. A clean-looking scan does not remove that requirement. Review the tool’s scope before sending the file through a conversion service.

Choose the workflow for your goal

A worked example: a mixed report

Imagine a ten-page report created from Word with two scanned receipts added at the end. The first ten pages may copy and search well while the receipts do not. A successful search on the cover does not establish that every attachment is searchable. Similarly, conversion may recover the report’s text without producing useful editable text for the receipts. Inspect representative pages from each source type. Keep the original pack so that any later conversion can be compared with what was actually submitted.

What can change the text layer later

Page rasterisation can turn a previously text-based page into an image. In FOLD’s documented implementation, aggressive compression and page redaction can do this for their respective outputs. That may be an intentional trade-off, but it means the downloaded file can behave differently from the source. Reopen it, test selection and search where those matter, and compare the pages affected by the operation. Changing the filename does not tell you whether the internal representation stayed the same.

Common questions

Why can I read the page but not copy the words?

The words may be pixels in a scanned image rather than stored text.

Does Images to PDF make a searchable document?

Not by itself. It puts images into PDF pages without performing OCR.

Does FOLD’s PDF-to-Word tool recognise scans?

The documented scope does not provide OCR, so scanned inputs can produce unusable results.

Can one PDF contain both kinds of page?

Yes. Review sections individually.