HOW-TO GUIDE
Convert PDF to EPUB Without Losing Formatting
“I converted my PDF to EPUB and the formatting is a mess” is one of the most common complaints about ebook converters. Sentences break in half, page numbers appear mid-paragraph, chapter titles vanish into body text. This guide explains what can be preserved, how to fix the usual problems in calibre, and a quicker route that lets you review the result before you read it.
What “formatting” survives a PDF to EPUB conversion
| Usually kept well | Needs checking | Not kept (by design) |
|---|---|---|
| Paragraphs, headings, bold and italics, lists, ordinary images | Tables, footnotes, captions, multi-column pages, scanned pages (OCR) | Exact line breaks, page numbers, page size, decorative positioning |
If the exact page appearance matters — forms, magazines with complex spreads, comics, sheet music — keep the PDF and read it on a larger screen instead.
The five problems you will see, and why
- Broken lines. A PDF stores each printed line separately. If the converter does not join them, every line becomes its own paragraph.
- Page headers and numbers inside sentences. Running titles, page numbers and footers sit on every page and get pulled into the text flow.
- Hyphenated words split in two. “impor- tant” appears where a word was hyphenated at the end of a printed line.
- Wrong reading order. On two-column pages the converter reads across both columns, or puts boxes and captions in the wrong place.
- Missing chapters in the table of contents. Headings that were just “large bold text” in the PDF are not recognised as headings.
Fixing it in calibre
calibre is free and powerful, and its manual is candid that PDF is one of the hardest formats to convert. These settings are in the Convert books dialog:
- PDF Input → Line un-wrapping factor. Controls when calibre joins short lines into paragraphs. The default is 0.45; lower it if lines are still broken, raise it if separate paragraphs are being merged.
- Heuristic processing → Enable, then “Unwrap lines” to repair hard line breaks using punctuation and line length.
- Heuristic processing → “Remove unnecessary hyphens” to rejoin words split at line ends. It uses the document itself as a dictionary.
- Search & replace for headers and footers. Write a regular expression that matches the repeated title or page number, and use the wand button to test it — matches are highlighted before you convert. The manual notes that leftover headers and footers also confuse paragraph unwrapping, so remove them first.
- Structure detection to tell calibre which text should start a chapter, so the table of contents is built correctly.
Expect a few rounds of convert, check and adjust. Each PDF needs its own regular expression and settings, which is why calibre works best when you convert many books from the same source.
The review-first alternative
If you would rather check a result than tune settings, TidyReads PDF to EPUB handles the same problems in a different order:
- Layout-aware extraction. Pages are analysed before text is joined, so lines become paragraphs and multi-column pages are read column by column.
- A cleanup list instead of regular expressions. Repeated page headers, footers, page numbers and similar clutter are detected and shown to you. Untick anything you want to keep; your original file is never changed.
- A free sampled preview of up to 15 pages, so you can compare headings, tables and images against the source before exporting.
- Several formats from one pass: EPUB for e-readers, a newly typeset PDF in five layouts, Word and Markdown.
It is not magic. Scanned pages depend on OCR quality, and complex tables or formulas may remain images or lose detail. The difference is that you see those issues in the preview, not halfway through chapter four.
A quick checklist before you start reading
- Open the table of contents: are the chapter titles all there?
- Jump to the middle of the book: do paragraphs run on naturally, with no stray page numbers?
- Find a page that had a table or figure: is it complete and in the right place?
- Search for a common word split at a line end, such as “ment” or “tion”, to catch broken hyphenation.
Common questions
Is there a converter that keeps the PDF layout exactly?
Some converters produce a “fixed-layout” EPUB by turning each page into an image. It looks like the original, but the text no longer resizes — you have the same small-text problem in a bigger file.
Why is my scanned PDF worse than a digital one?
A scan is a photo of each page. The text must be recognised (OCR) first, which can misread characters, numbers and tables. Check names and figures in the preview.
Can I fix the EPUB after converting?
Yes, with an EPUB editor such as calibre’s editor or Sigil. For a one-off book, fixing the cleanup before export is usually quicker.
Next: how to read the result on your Kindle.
TidyReads is an independent reading tool and is not affiliated with Amazon, Kobo, Onyx Boox, arXiv or calibre. Product names belong to their owners. Device features vary by model and software version.