Scanned PDF to Word: Why the Formatting Breaks
By the FreePDF team ·
A scanned PDF isn't a Word document waiting to be unpacked. It's a stack of page photos, so before anything can be edited, something has to read the letters off the images and then guess where paragraphs, columns and tables belong. That guessing is where the formatting goes wrong, and with a scan it pays to pick your route before you start.
First, check what kind of PDF you have
Try selecting one sentence in your PDF viewer. If individual words highlight and copy cleanly, the file already has a text layer and PDF to Word can rebuild it as an editable document. If the whole page highlights like a single picture, it's a scan, and the rest of this post applies.
What our PDF to Word does with a scan
Put a scan straight into PDF to Word and you get a .docx with each page placed as an image. It looks right, but there's nothing to type into. Running the file through OCR first doesn't change that: OCR adds an invisible text layer behind each page, and the Word converter ignores that hidden layer. So for scans, skip PDF to Word and use one of the routes below.
Routes that do work for a scan
- You need the words in Word: use PDF to Text and tick the Run OCR option. You get a .txt with the recognised text, which you paste into a new Word document and format there. You rebuild headings and tables yourself, but you're not retyping.
- You only need to search and copy: OCR PDF is the final step. The page keeps its exact look, and the words become selectable.
- It's mostly a table of figures: run OCR PDF, then PDF to Excel. The table reader can use the OCR layer, but check the columns, since neighbouring cells sometimes get merged.
Why the layout breaks either way
OCR recognises characters; it can't know the design decisions behind the page. A faint scan can turn an 8 into a 3. Two columns may be read straight across. A table can come out as values separated by spaces, and handwriting is unreliable. Our OCR also reads English only, so accented letters get mangled and non-Latin scripts aren't recognised at all.
- Clean, straight, high-contrast scans give OCR its best chance.
- Tables and multi-column pages need the closest review.
- Stamps and handwritten notes should be checked character by character where accuracy matters.
A better source beats any amount of fixing
If the paper original is still around, rescan it square to the page at around 300 DPI, with no fingers, shadows or folded corners in the frame. If the PDF came out of an office system, ask for the original digital file or the source document instead. A real digital PDF goes through PDF to Word properly, and one clean original saves more time than an hour of formatting repair.
Common questions
Why does my Word file only contain pictures of the pages?
Because the PDF was a scan. The converter found no real text to rebuild, so it placed each page as an image. Use PDF to Text with OCR switched on to get the words out, then paste them into Word.
Can I fix a blurry scan by converting it again?
No. Repeating a conversion can't add detail the scan never captured. Rescan the original more clearly if you can, then run OCR once.