Skip to content

DjVu vs PDF - which should you use?

DjVu and PDF overlap enough to look like rivals, but they were built for different problems. PDF is a general document container that can hold anything. DjVu does one thing - scanned pages - and does it better than PDF does.

For most people the practical question is not which format is better in the abstract, but which one to keep a particular document in.

  • Keep the DjVu when you are archiving scanned books and care about disk space and page sharpness.
  • Export a PDF when the document has to leave your machine, be opened by someone else, be annotated in another app, or outlive your choice of reader.

Keeping both is a perfectly reasonable answer, and often the right one.

This is DjVu’s decisive advantage. It separates a scanned page into a high-resolution text mask and low-resolution background and foreground layers, then reuses repeated shapes across the document.

A scanned 300-page book is commonly:

  • 5-20 MB as DjVu
  • 50-150 MB as a PDF of page images

PDF can close some of that gap with JBIG2 or JPEG 2000 compression, but most PDFs of scans in the wild do not use them, and even when they do, DjVu is usually smaller at the same visual quality.

At equal file size, DjVu holds an advantage on exactly the kind of page it was designed for: black text on off-white paper. Its layered model spends its bits on the letterforms and almost none on the paper, so the type stays crisp where an equivalently sized JPEG-in-PDF goes soft and blotchy around the strokes.

For photographs, color plates, and illustrations, the advantage shrinks or disappears - those are exactly the images ordinary photographic compression is already good at.

Neither format guarantees it. Both can carry an invisible text layer behind the page image, and in both formats whether a given file has one depends entirely on whoever produced it. Plenty of DjVu files from library scanning projects have excellent text layers; plenty of PDFs are pure images.

If a document has no text layer, on-device text recognition can create one, and the text layer survives an export to PDF. See text recognition.

PDF wins outright, and it is not close.

PDF opens in Preview on every Mac, in every browser, on every phone, in Acrobat, on e-readers, and in every system that accepts document uploads. DjVu opens in none of those without an app you install on purpose - not on macOS, not on Windows, not on iOS or Android, not in any browser.

This is the reason to export a PDF even when the DjVu is technically the better file: a format nobody can open is not much use for sending to someone.

Both are documented, non-proprietary formats with open-source implementations, so neither is at real risk of becoming unreadable. PDF has a formal ISO standard and an archival profile (PDF/A) that institutions require; DjVu has a stable specification and DjVuLibre.

If you are depositing something in an institutional archive, that requirement will usually be PDF/A, and the decision is made for you.

PDF, comfortably. Annotation, form filling, redaction, page reordering, and signing are all routine in PDF tools and largely absent from the DjVu ecosystem. DjVu is a delivery format for scans, not a working format.

DjVu PDF
Scanned-book file size Much smaller Larger
Sharpness of scanned type Excellent Good, depends on compression
Photographs and color plates No real advantage Equal or better
Searchable text possible Yes Yes
Opens without installing anything No Yes
Annotation and editing tools Very few Many
Archival standard None formal PDF/A

If you have a library of scanned books in DjVu, there is no reason to convert them all - they are smaller and sharper as they are, and one reader opens them. Convert individual documents to PDF when you need to send, submit, annotate, or print through something else.

Niraqua DjVu reads DjVu natively on Mac, iPhone, and iPad and exports to PDF with a searchable text layer, all on your own device. See converting DjVu to PDF or what a DjVu file is.