EPUB vs PDF: which should you use?

Short answer: use EPUB for anything that is mostly words. Use PDF when the layout is part of the content, or when the page number has to be citable.

The one thing that actually differs

A PDF is a description of a printed page: this glyph, at these coordinates, in this size. It looks identical everywhere, which was the whole point when it was designed for prepress and office printing.

An EPUB is HTML and CSS in a zip file. It describes the text and lets your reader decide how to lay it out. Change the font size and the text reflows into new lines and new pages.

On a 30-centimetre sheet of paper the first approach is superior. On a 7-centimetre phone screen it’s a problem, because a fixed A4 page has to be scaled down until the type is unreadable, or panned around one column at a time.

Side by side

Where PDF is genuinely the right choice

• Anything where design carries meaning: art books, photography, graphic novels, sheet music, maps.
• Academic papers. Reviewers cite page numbers, and equations and figures need to stay where the author put them.
• Forms and documents you’ll print. That is what the format was built for.
• Anything legally fixed: contracts, official filings, records that must not reflow.

The mistake isn’t using PDF. The mistake is using it for a 400-page novel.

Why PDFs are hard to listen to

Text-to-speech has to know what order to read in. An EPUB says so explicitly. A PDF stores text as positioned fragments, so the reading order has to be reconstructed, and a two-column layout with running headers frequently gets read as: paragraph, page number, chapter title, other column.

Reflowing solves it. Lectern reflows a PDF into flowing text before narrating it, which rebuilds the reading order and drops the page furniture, so on-device voices read the actual paragraph. Detail in how to read PDFs aloud.

Converting between them

PDF to EPUB is the useful direction and the harder one. The converter has to guess the structure from visual layout: which text is a heading, where a paragraph continues across a page, what to do with a footnote. Calibre does a decent job on simple, text-heavy PDFs and struggles with columns, tables and scans. Always skim the result before deleting the original.

EPUB to PDF is easy and usually pointless: you throw away flexibility to gain a fixed page you probably don’t need. Do it only when you plan to print.

If the PDF is a scan, no conversion will help until OCR has added a text layer, because there is no text in the file at all.

A practical rule

Ask what would be lost if the text rearranged itself. For a novel, an essay, a biography or a manual: nothing. Use EPUB. For a score, a chart-heavy report or a paper you’ll cite: quite a lot. Keep the PDF, and reflow it in your reader when you just want to read the words.

Frequently asked