TIFF MIME Type

What does image/tiff mean?

The image/tiff media type identifies a TIFF raster image. TIFF is a tag-based container rather than one fixed pixel encoding. Its image file directories, or IFDs, describe dimensions, color interpretation, compression, resolution, and the locations of the image data. A file can also contain more than one IFD, which is how multipage scans and related images fit into one container.¹

The current IANA registration points to TIFF Revision 6.0 and defines no required media-type parameters. It permits an optional application parameter as a hint about a TIFF profile, but defines no values for it. The registered extension is .tif; .tiff is also widely used by software and web documentation.²

TIFF has two standard four-byte signatures because the file itself declares its byte order:

Both are ordinary TIFF 6.0 headers. The next four bytes point to the first IFD.¹²

Why can two TIFF files behave so differently?

TIFF was designed for interchange among scanners, image editors, and publishing tools. The format can describe bilevel, grayscale, palette-color, and full-color raster data in several color spaces. It can store pixels in strips or tiles and can use different compression schemes.¹

That flexibility is useful, but it is also TIFF's main interoperability trap. Supporting “TIFF” does not guarantee that an application understands every compression method, color space, bit depth, private tag, layer convention, or specialized profile. The image/tiff registration explicitly warns that a receiver may recognize the container yet still be unable to render a particular file.²

Common choices include:

• Uncompressed data for simple, predictable interchange at the cost of large files.
• LZW, ZIP/Deflate, or PackBits when a workflow needs lossless compression.
• JPEG compression inside TIFF when smaller size matters and lossy encoding is acceptable.
• Multiple IFDs for pages, reduced-resolution images, or other related raster representations.
• Private tags for application-specific information that another program may ignore.

Adobe Photoshop exposes several of these choices when saving TIFF, including LZW, ZIP, JPEG, byte order, transparency, image pyramids, and optional layer data. Adobe also cautions that many applications skip Photoshop's layer data even when they can read the composite image.³

History and standardization

Aldus Corporation published the first TIFF specification in the fall of 1986 after meetings with scanner manufacturers and software developers. The document later described that release as effectively Revision 3.0 because two major drafts preceded it. Revision 4.0 followed in April 1987, Revision 5.0 in October 1988, and Revision 6.0 on June 3, 1992.¹

Revision 5.0 added palette-color images and LZW compression. Revision 6.0 expanded the format with definitions for CMYK, YCbCr, CIE L*a*b*, tiled images, and other capabilities. RFC 3302 updated the standard image/tiff registration in 2002 so that it referred generally to TIFF 6.0 instead of the narrower fax-oriented definition used by an earlier registration.¹⁴

Classic TIFF uses 32-bit offsets and therefore cannot address beyond 4 GiB. BigTIFF is a later extension that changes the header version from 42 to 43 and uses 64-bit offsets. It is structurally related to TIFF, but software support must be checked separately.⁵

How to open a TIFF file

TIFF is common in desktop imaging and publishing software, but it is a poor web-delivery format. MDN lists native TIFF display only for Safari among current mainstream browsers. Chrome, Edge, Firefox, and Opera should not be expected to render an image/tiff response directly.⁶

For desktop work:

• Adobe Photoshop can open TIFF files and save them with detailed compression, color, layer, and byte-order options.³
• Apple Preview can open images and export them as TIFF, PNG, JPEG, JPEG 2000, HEIC, or PDF.⁷
• Specialist scanning, publishing, GIS, and preservation tools may support profiles and tags that general-purpose viewers ignore.

If a trusted application rejects a TIFF, do not assume the file is necessarily corrupt. It may use an unsupported compression method, a specialized profile, BigTIFF offsets, an unusual color space, or private application data.

Support in Poly

Poly recognizes both .tif and .tiff filenames as image/tiff. Its image loader fetches the complete file for sizing, decodes a raster representation, applies the encoded orientation, and generates lossless WebP thumbnails. It also records image dimensions and available TIFF/EXIF fields.

The rendered result powers Poly's image viewer and visual-similarity search. The visual embedding is generated from a WebP derivative, not from the original TIFF byte stream. Filename and extracted metadata are searchable too, but words visible inside an ordinary TIFF image are not OCR'd.

The Poly agent follows the same rendered-image model. Once thumbnails exist, it receives the largest suitable WebP thumbnail up to 1536 pixels. This supports visual reasoning, but it does not expose the original compression, every TIFF tag, Photoshop layers, or a document-style sequence of all pages in a multipage file.

Converting TIFF to a more practical format

Choose the destination based on what must survive:

The Library of Congress reports extensive use of TIFF in digitization and preservation workflows, including primary or master files. That does not make every TIFF automatically archival. Long-term usefulness still depends on choosing a documented, well-supported profile and retaining the metadata needed to interpret the pixels.⁸

Conversion works on the decoded image an application can see. It cannot recover an omitted page, unsupported private tag, discarded layer, or detail already lost through JPEG compression inside the source TIFF.

Is TIFF lossless?

Not necessarily. TIFF can store uncompressed pixels or use lossless schemes such as LZW and ZIP, but it can also contain JPEG-compressed image data. The .tif or .tiff extension does not reveal which choice was used.¹³

To verify a file, inspect its compression tag with a TIFF-aware metadata tool or open it in software that reports the encoding. If you are preparing a master image, choose a lossless option explicitly and keep the original source until the result has been validated.