DJVU Format Guide

DjVu (pronounced "deja vu") is a computer file format designed primarily for storing scanned documents, especially those containing a combination of text, line drawings, indexed color images, and photographs. Developed by Yann LeCun, Leon Bottou, Patrick Haffner, and Paul G. Howard at AT&T Labs in 1996, DjVu was created to address the need for a highly efficient format for distributing scanned books, magazines, and documents over the internet. The format achieves remarkable compression ratios by separating each page into distinct layers -- a high-resolution foreground mask for text and line art, a lower-resolution foreground color layer, and a separately compressed background layer for photographs and textures. This layered segmentation approach allows DjVu to compress a typical scanned color page to 40-70 KB, compared to 500 KB or more for JPEG at similar visual quality.

History of DjVu

DjVu was born out of research at AT&T Labs (later AT&T Labs-Research) in the mid-1990s, where the team explored neural network-based document analysis and compression techniques. The initial release in 1996 introduced the core concept of mixed raster content (MRC) compression, which intelligently segments page content into layers optimized for different compression algorithms. AT&T released the DjVu technology as open source under the GNU GPL license in 2000, enabling widespread adoption. LizardTech subsequently acquired the commercial rights and developed the DjVu software suite, while the open-source community maintained the DjVuLibre project as the reference implementation. The Internet Archive adopted DjVu as one of its primary formats for storing millions of scanned books, bringing the format to a massive audience. Over the years, DjVu gained popularity in academic institutions, national libraries, and digital archive projects worldwide, particularly in countries where large-scale book digitization efforts required extremely efficient storage of scanned pages.

Key Features and Uses

DjVu's most distinctive technical feature is its use of the IW44 wavelet-based compression algorithm for background layers, which provides progressive rendering and excellent quality at very low bitrates. The foreground text layer uses JB2 compression (an implementation related to JBIX2), which achieves extraordinary compression of bi-level text by identifying and matching similar character shapes across the page. DjVu supports multi-page documents natively, with shared dictionaries that further reduce file sizes when pages contain repeated elements such as fonts or logos. The format includes a hidden text layer that stores OCR results, making scanned documents fully searchable while preserving the original scan appearance. DjVu files also support hyperlinks, bookmarks, and annotations, making them suitable for interactive document viewing. The format excels at compressing documents that combine text and images -- a 300 DPI color scan of a magazine page that would be 25 MB uncompressed can be stored in just 50-100 KB as DjVu. This efficiency makes DjVu particularly valuable for digital libraries, online archives, technical documentation repositories, and any application where large volumes of scanned documents must be stored and distributed efficiently.

Common Applications

DjVu is widely used by digital libraries and national archives for storing millions of scanned books, manuscripts, and historical documents. The Internet Archive hosts a vast collection of DjVu files, making it one of the largest repositories of content in this format. Academic institutions and research libraries use DjVu to distribute scanned journal articles, theses, and rare book collections. Technical documentation teams use DjVu for storing scanned engineering drawings, schematics, and legacy manuals where the combination of line art and photographic content benefits from DjVu's layered compression. Government agencies in several countries have adopted DjVu for digitizing public records and legal documents. The format is also popular for sharing scanned comic books and graphic novels due to its excellent handling of mixed content. Converting DjVu files to formats like PDF, JPG, or PNG is essential for compatibility with devices and applications that do not natively support DjVu, while converting images to DjVu can dramatically reduce storage requirements for scanned document collections.

Advantages and Disadvantages

Advantages

Exceptional Compression: 5-10x smaller than JPEG for scanned documents with text and images
Layered Segmentation: Separates text, graphics, and photos for optimal per-layer compression
IW44 Wavelet Compression: Progressive rendering with excellent quality at very low bitrates
Searchable Text Layer: Hidden OCR text layer makes scanned documents fully searchable
Multi-Page Support: Native multi-page documents with shared dictionaries for repeated elements
Open Format: Open-source DjVuLibre reference implementation under GNU GPL
Fast Rendering: Efficient decoding allows smooth viewing even on low-power devices
Hyperlinks and Annotations: Supports bookmarks, hyperlinks, and document annotations

Disadvantages

Limited Native Support: Not supported by most browsers, image viewers, or operating systems natively
Niche Format: Primarily used for scanned documents, less suitable for general photography
PDF Dominance: PDF has become the universal standard for documents, limiting DjVu adoption
Fewer Editing Tools: Very few applications can create or edit DjVu files compared to PDF
Mobile Support: Limited native support on iOS and Android without dedicated apps
No Vector Graphics: Cannot store true vector data, only rasterized content
Declining Development: Less active development compared to modern formats like PDF 2.0
Accessibility Challenges: Less mature accessibility features compared to tagged PDF