Download TIFF sample images

The Tag Image File Format was originated in 1986 by Aldus Corporation under the leadership of Paul Brainerd. Designed as a universal format to standardize scanned desktop publishing imagery across competing desktop scanner hardware (such as Hewlett-Packard scanners) and computer operating systems, TIFF quickly replaced numerous incompatible proprietary scanner formats.

Aldus released TIFF Revision 5.0 in 1988, introducing support for palette color and Lempel-Ziv-Welch (LZW) compression. The current baseline specification, TIFF Revision 6.0, was finalized in June 1992, formalizing CMYK color separation, CIE L*a*b* colorimetry, tiling, and JPEG compression extensions. In 1994, Adobe Systems acquired Aldus and assumed copyright stewardship over the specification, issuing Technical Notes to refine and modernize container behaviors.

Because TIFF was engineered with an extensible tag-based architecture, industry consortiums adopted it as the foundation for specialized scientific standards. The Open Geospatial Consortium (OGC) created GeoTIFF to embed cartographic projection metadata, ISO 12234-2 established TIFF/EP for electronic digital photography (which formed the basis for Adobe DNG and camera RAW files), and BigTIFF expanded offsets from 32-bit to 64-bit to break the historic 4 GB file barrier.

A hallmark of TIFF is its support for both uncompressed raw storage and diverse lossless and lossy compression algorithms recorded in Compression tag 259. When uncompressed (Tag 259 = 1), TIFF stores exact numerical sensor data, ensuring zero generation loss across editing cycles at the expense of larger disk footprints.

For efficient storage, TIFF incorporates LZW compression (Tag 259 = 5) and Deflate / ZIP compression (Tag 259 = 8 / 32946). When compressing continuous-tone photographic imagery, TIFF encoders apply a horizontal differencing predictor (Predictor tag 317 = 2). The predictor replaces absolute color values with differences between adjacent horizontal pixels, dramatically increasing repetitive byte sequences and improving LZW/Deflate compression ratios by up to 30%.

For high-speed bilevel document scanning and fax transmission, TIFF incorporates CCITT Group 3 (Tag 259 = 3) and CCITT Group 4 T.6 (Tag 259 = 4) two-dimensional run-length encoding. For legacy graphics, PackBits (Tag 259 = 32773) provides rapid byte-level run-length encoding. Color bit depths range from 1-bit monochrome up to 8-bit, 16-bit integer, and 32-bit IEEE floating-point representations per channel.

TIFF files are structured around Image File Directories (IFDs) and tag-value pairs. Every TIFF file begins with an 8-byte header: a 2-byte byte-order indicator (either "II" 0x4949 for Little-Endian/Intel or "MM" 0x4D4D for Big-Endian/Motorola), followed by magic number 42 (0x002A), and a 4-byte offset pointing to the first IFD.

Each IFD contains a 2-byte count of directory entries, followed by a sequence of 12-byte field entries (comprising a 2-byte tag identifier, a 2-byte data type, a 4-byte count, and a 4-byte value or offset), concluding with an offset to the subsequent IFD. This linked-list directory architecture allows a single physical TIFF file to encapsulate multiple sub-files or pages, making it the bedrock standard for multi-page document scanning, medical DICOM imaging, and legal fax archiving.

Tags also define comprehensive prepress and spatial parameters: ImageWidth (256), ImageLength (257), BitsPerSample (258), PhotometricInterpretation (262, defining RGB, CMYK, Palette, or Bilevel), ResolutionUnit (296), and XResolution/YResolution (282/283) specifying exact print pixel densities (typically 300 to 600 DPI) for offset printing presses.

Standard modern web browsers (including Google Chrome, Mozilla Firefox, Microsoft Edge, and modern Apple Safari) do not natively render TIFF images within HTML <img> tags. Because TIFF encompasses dozens of optional compression modes, private tags, and CMYK color spaces, browser rendering engines deliberately omit native TIFF decoders to minimize attack surface and binary footprint.

Web applications requiring TIFF previews decode files client-side via JavaScript/WebAssembly decoders such as UTIF.js or tiff.js, drawing decoded pixel rasters onto HTML5 <canvas> elements or converting them into WebP or PNG on the fly.

On backend servers, software engineers process TIFF files using industry-standard libraries: Node.js applications use Sharp (powered by libvips), Python workflows rely on Pillow and tifffile, command-line automation employs ImageMagick, and native C/C++ applications link against the authoritative LibTIFF reference library.