Tagged Image File Format
Tagged Image File Format | |
---|---|
File extension : |
.tiff, .tif
|
MIME type : | image / tiff, image / tiff-fx |
Developed by: | Aldus Corporation |
Current version: | 6.0 (as of June 3, 1992) |
Type: | Bitmap - graphic format |
Expanded to: | Exif, sDCF, TIFF / EP, TIFF / IT, GeoTIFF |
The Tagged Image File Format ( TIFF or TIF for short ) is a file format for storing image data . It was originally developed by Aldus ( taken over by Adobe in 1994 ) and Microsoft for color separation in scanned raster graphics . TIFF files have the MIME type image / tiff .
use
In addition to PDF and EPS , TIFF is an important format for the exchange of data in the prepress stage in publishing houses and printing companies because it supports the CMYK color model they use . In addition, TIFF images can be saved with a high color depth (up to 32 bits per color component), which is why TIFF is often used for data exchange during RAW conversion.
TIFF is used on the Internet to provide users such as publishers with high-resolution images in printable, loss-free quality. It is accepted that these files are several times the size of a lossy compressed JPEG image. TIFF has thus established itself as the quasi- standard for high-quality images.
For the area of grid-based geographic information , a TIFF variant with additional tags, the so-called GeoTIFF , is becoming more and more established . It allows - for example in the case of map images, aerial photos and similar information - to indicate where on the earth the situation shown in the image is exactly based on coordinates.
TIFF is also used for archiving monochrome graphics (e.g. technical drawings) and in the computer-to-plate process in prepress, as very compact files are created in conjunction with "Fax Group 4" compression.
In order to make the work of the print media easier, some organizations, for example ESA or NASA , offer high-resolution images in addition to the JPEG format in some places in TIFF, while at low resolutions only the lossy JPEG format is normally used.
properties
The coding of numbers ( byte order ) can be done either as big endian or little endian . Several images can be stored in one file ( multipage TIFF ). That can be B. be different versions of the same image, such as preview image ("thumbnail") and original image. TIFF knows different color spaces and algorithms for data compression . Most of them are lossless (e.g. LZW , run length coding ), but TIFF can also serve as a container format for JPEG images, which may be lossy ( DCT ) compressed. It is also possible to embed IPTC metadata in the TIFF file.
With TIFF, individual pixels can consist of any number of individual values (samples) . In addition to the standard case of "one byte equals one sample", samples can also take up parts of a byte (e.g. 1, 2 or 4 bits) or consist of several bytes. In addition to integers , floating point numbers can also be stored as image data. There is also the option of storing transparency information ( alpha channel ).
Image data is in groups of pixel lines, so-called stripes (stripes), or as rectangular tiles stored (tiles). The storage takes place for each strip or for each tile independently of the others, so that parts of the image can be loaded relatively quickly, depending on the size of the strips or tiles selected. Other formats require that all image data be loaded before the desired section. The main aim of the subdivision in the design was that individual parts could be kept completely in the memory. The 1992 specification recommends a maximum size of eight kilobytes.
Programs like Photoshop offer to create TIFF files with separate layers . There is also the option of saving TIFFs with an image pyramid . This means that multiple resolutions of the image are contained within one file. With this, for example, layout programs or image viewers can display a small preview of the image more quickly because they do not need to load the image in full resolution.
restrictions
The biggest disadvantage of TIFF is its complexity . The variety of possible valid TIFF files can only be supported with difficulty by individual programs. The specification of the file format therefore defines a subset of valid TIFF files that every TIFF-capable program should be able to process, called Baseline TIFF .
32 bits are generally used for offset values . This means that only places up to four gigabytes from the beginning of the file can be referenced. Since some programs implement it as a signed 32-bit offset, these programs even have the limit of 2 gigabytes. At the time TIFF was developed, this was not a limitation in practice, but recently, particularly in some scientific disciplines (e.g. astronomy ), very large images have been generated that TIFF cannot save. BigTIFF was introduced to get around this limitation. This format allows the use of 64-bit offsets.
A streaming TIFF files is not possible in all cases without backup of the entire file as TIFF does in many places offset values use, can reference the data that lies before the point at which they are referenced. Thus, random access or even buffering in memory is a necessity. For this reason, TIFF is also unsuitable for the WWW . With TIFF (unlike its PDF document format ), Adobe has never subsequently added a linearization method that guarantees a sequence in the converted file that makes recourse unnecessary.
Basic structure
The first eight bytes of a TIFF file contain a signature that contains the byte order and the magic number 42 , as well as an offset to the first Image File Directory (IFD).
Such an IFD gives information about an image in the TIFF file and consists of a list of markings ( English " tags "), individual information units. Such a mark can describe the width of the image in pixels or the name of the software with which the TIFF file was created.
Each tag has its own identification number (e.g. 256 for the image width) and a type (e.g. 16-bit integers, 32-bit floating point numbers, character strings and much more). Some of these tags must be available (e.g. the image width), others are optional (e.g. the name of the software). A number of applications use proprietary tags . The structure of the data that is saved or referenced in such a tag is then usually not documented. You can apply for a number for your own proprietary tags from Adobe.
At the end of the IFD there is an offset value that references the next IFD in the file, or 0 if the current IFD was the last. In this way, any number of images can be stored in the file, as long as the total amount of data does not exceed four gigabytes (see also restrictions above).
The IFD structure is reused in the Exif part of JFIF files.
specification
A detailed description of the format for developers is available as a PDF file. The latest version is 6.0 from June 3, 1992. It is supplemented by TIFF Technical Notes . These are texts that add individual capabilities to TIFF, including the deflate method for lossless data compression , which is already used in gzip and PNG .
See also
literature
- Thomas W. Lipp: Graphic formats. Microsoft Press, Unterschleißheim 1997, ISBN 3-86063-391-0 .
Web links
- LibTIFF - open source library for reading / writing TIFF files with many examples (English)
- TIFF information pages (English)
Individual evidence
- ↑ RFC 3302
- ↑ The BigTIFF File Format (English)
- ↑ TIFF specification (PDF) on itu.int . June 3, 1992. Retrieved March 31, 2017.