| ZVON > RFC Repository > RFC 2301 |
| Frontpage / Contents |
| Prev | Next | RFC index | RFC search | Download as zip/tar.gz |
TIFF provides a means for describing, storing and interchanging raster image data. A primary goal of TIFF is to provide a rich environment within which applications can exchange image data. The current TIFF specification [TIFF] defines a commonly used, core set of TIFF fields known as Baseline TIFF. The current specification and TIFF Technical Notes 1 and 2 [TTN1, TTN2] define several TIFF extensions. The TIFF- based specification for fax applications uses a subset of Baseline TIFF fields, with selected extensions, as described in this document. In a few cases, this document defines new TIFF fields specifically for fax applications.
TIFF is designed for raster images, which makes it a good match for facsimile documents, which are multi-page raster images. Each raster image consists of a number of rows or scanlines, each of which has the same number of pixels, the unit of sampling. Each pixel has at least one sample or component (exactly one for black-and-white images).
A TIFF file begins with an 8-byte image file header. The first two bytes describe the byte order used within the file. Legal values are "II" (0x4949) when bytes are ordered from least to most significant (little- endian), and "MM" (0x4D4D), when bytes are ordered from most to least significant (big-endian) within a 16- or 32-bit integer. Either byte order can be used, except in the case of the minimal black-and-white mode, which SHALL use value "II". The next two bytes contain the value 42 that identifies the file as a TIFF file and is ordered according to the value in the first two bytes of the header. The last four bytes give the offset that points to the first image file directory (IFD). This and all other offsets in a TIFF file are with respect to the beginning of the TIFF file. An IFD can be at any location in the file after the header but must begin on a word boundary.
An IFD is a sequence of tagged fields, sorted in ascending order by tag value. An IFD consists of a 2-byte count of the number of fields, a sequence of field entries and a 4-byte offset to the next IFD. The fields contain information about the image and pointers to the image data. Each separate raster image in the file is represented by an IFD.
Each field entry in an IFD has 12 bytes and consists of a 2-byte Tag, 2 bytes identifying the field type (e.g. short, long, rational, ASCII), 4 bytes giving the count (number of values or offsets), and 4 bytes that either contain the offset to a field value stored outside the IFD, or, based on the type and count, the field value itself. Resolution and metadata such as dates, names and descriptions are examples of "long" field values that do not fit in 4 bytes and therefore use offsets in the field entry. Details are given in the TIFF specification [TIFF].
A TIFF file can contain more than one IFD, where each IFD is a subfile whose type is given in the NewSubfileType field. Multiple IFDs can be organized either as a linked list, with the last entry in each IFD pointing to the next IFD (the pointer in the last IFD is 0), or as a tree, using the SubIFDs field in the primary IFD [TTN1]. The SubIFDs field contains an array of pointers to child IFDs of the primary IFD.
Child IFDs describe related images, such as reduced resolution versions of the primary IFD image. The same IFD can point both to a next IFD and to child IFDs, and child IFDs can themselves point to other IFDs.
All fax modes represent a multi-page fax image as a linked list of IFDs, with a NewSubfileType field containing a bit that identifies the IFD as one page of a multi-page document. Each IFD has a PageNumber field, identifying the page number in ascending order, starting at 0 for the first page. While a Baseline TIFF reader is not required to read any IFDs beyond the first, an implementation that reads the files that comply with this specification SHALL read multiple IFDs. Only the Mixed Raster Content fax mode, described in Section 8, requires the use of child IFDs.
The following figure illustrates the structure of a multi-page TIFF file.
+-----------------------+
| Header |------------+
+-----------------------+ | First IFD
| IFD (page 0) |<-----------+ Offset
+---| |------------+
Value | +-----------------------+ |
Offset +-->| Long Values |--+ |
+-----------------------| | Strip |
| Image Data |<-+ Offset |
| strip 1 page 0 | | |
+-----------------------+ | |
| : | : |
|
+-----------------------+ | Next IFD
| IFD (page 1) |<-----------+ Offset
+---| |------------+
Value | +-----------------------+ |
Offset +-->| Long Values |--+ |
+-----------------------| | Strip |
| Image Data |<-+ Offset |
| strip 1 page 1 | | |
+-----------------------+ | |
| strip 2 page 1 |<-+ |
+-----------------------+ | |
| : | : |
|
+-----------------------+ | Next IFD
| IFD (page 2) |<-----------+ Offset
| : |
An IFD stores an image as one or more strips, as shown in the preceding figure. A strip consists of 1 or more scanlines (rows) of raster image data in compressed form. An image may be stored in a single strip or may be divided into several strips, which would require less memory to buffer. (Baseline TIFF recommends about 8k bytes per strip, but existing fax usage is typically one strip per image.)
Each IFD requires three strip-related fields: StripOffsets, RowsPerStrip and StripByteCounts. The StripOffsets field is an array of pointers to the strip or strips that contain the actual image data. The StripByteCounts field gives the number of bytes in each strip after compression. TIFF requires that each strip, except the last, contain the same number of scanlines, which is given in the RowsPerStrip field. This document introduces the new StripRowCounts field that allows a variable number of scanlines per strip, which is required by the Mixed Raster Content fax mode (Section 8).
Image data is stored as uninterpreted, compressed image data streams within a strip. The formats of these streams follow the ITU-T Recommendations. The Compression field in the IFD indicates the type of compression, and other TIFF fields in the IFD describe image attributes, such as color encoding and spatial resolution. Compression parameters are stored in the compressed data stream, rather than in TIFF fields. This makes the TIFF representation and compressed data format specification independent of each another. This approach, modeled on [TTN2], allows TIFF to gracefully add new compression schemes as they become available.
Some attributes can be specified both in the compressed data stream and within a TIFF field. It is possible that the two values will differ. When this happens for values required to interpret the data stream, then the values in the data stream take precedence. For informational values that are not required to interpret the data stream, such as author name, then the TIFF field value takes precedence.
The TIFF specification has a very flexible file structure, which does not specify the ordering of IFDs, field values and image data in a file. Individual applications may require or recommend an ordering.
This specification recommends that when using a TIFF file for facsimile, A multi-page fax document SHOULD be represented as a linked list of IFDs. It also recommends that a TIFF file for facsimile SHOULD order pages in a TIFF file in the same way that they are ordered in a fax data stream. In a TIFF file, a page consists of several elements: one or more IFDs (including subIFDs), long field values that are stored outside the IFDs, and image data (in one or more strips).
The minimal black-and-white mode (Profile S) specifies a required ordering of pages and elements within a page (Section 3.5). The extended black-and-white mode (Profile F) provides guidelines for ordering pages and page elements (Section 4.4.6). Other profiles
SHOULD follow these guidelines. This recommendation is intended to simplify the implementation of TIFF writers and readers in fax applications and the conversion between TIFF file and fax data stream representations. However, for interchange robustness, readers SHOULD be prepared to read TIFF files whose structure is consistent with [TIFF], which supports a more flexible file structure than is recommended here.
This specification introduces an optional new GlobalParametersIFD field, defined in Section 2.2.4. This field has type IFD and indicates parameters describing the fax session. While it is often possible to obtain these parameters by scanning the file, it is convenient to make them available together in one place for fast and easy access. If the GlobalParametersIFD occurs in a TIFF file, it SHOULD be located in the first IFD, immediately following the 8-byte image file header.
The TIFF specification [TIFF] is organized as a baseline set and several extensions, including technical notes [TTN1, TTN2] that will be incorporated in the next release of TIFF. The baseline and extensions have required and optional fields.
Facsimile applications require (and recommend) a mixture of baseline and extensions fields, as well as some new fields that are not part of the TIFF specification and that are defined in this document. This sub- section lists the fields that are required or recommended for all modes. In particular, Section 2.2.1 lists the fields that are required by all modes and that have values that do not depend on the mode. Section 2.2.2 lists the fields that are required by all modes and that have values which do depend on the mode. Section 2.2.3 lists the fields that are recommended for all modes. Fields that are required or recommended by some but not all modes are given in the section (Section 3-8) that describes that mode. The sections for each fax mode have sub-sections for required and recommended fields; each sub-section organizes the fields according to whether they are baseline, extension or new.
The fields required for facsimile have only a few legal values, specified in the ITU-T Recommendations. Of these legal values, some are required and some are optional, just as they are required (mandatory) or optional in fax implementations that conform to the ITU-T Recommendations. The required and optional values are noted in the sections on the different fax modes.
This section describes the fields required or recommended by all fax modes. The pattern for the description of TIFF fields in this draft is:
FieldName(TagValueInDecimal) = allowable values. TYPE
WhetherRequiredByTIFForTIFFforFAX
Count = (omitted if =1) = (if not in current spec but available)
Explanation of the field, how it's used, and the values it can have.
Default value, if any, as specified in [TIFF]
When a field's default value is the desired value, that field may be omitted from the relevant IFD unless specifically required by the text of this specification.
The TIFF fields listed in this section SHALL be used by all fax modes, but have field values that are not specified by the ITU standards, i.e. the fields do not depend on the mode. The next sub- section lists the fields that SHALL be used by all fax modes, but which do have values specified by the ITU-specified or mode-specific values. Fields that SHALL be used by some but not all modes are given in the sections (3-8) which describe the modes that uses them.
ImageLength(257) SHORT or LONG
RequiredByTIFFBaseline
Total number of scanlines in image.
No default, must be specified.
PageNumber(297) SHORT
RequiredByTIFFforFAX, TIFFExtension
Count = 2
The first number represents the page number (0 for the first page);
the second number is the total number of pages in the document. If
the second value is 0, then the total page count is not available.
No default, must be specified
RowsPerStrip(278) SHORT or LONG
RequiredByTIFFBaseline
The number of scanlines per TIFF strip, except for the last strip.
For a single strip image, this is the same as the value of the
ImageLength field.
Default = 2**32 - 1 (meaning all scanlines in one strip)
StripByteCounts(279) SHORT or LONG
RequiredByTIFFBaseline
Count = number of strips
For each strip, the number of bytes in that strip after compression.
No default, must be specified.
StripOffsets(273) SHORT or LONG
RequiredByTIFFBaseline
Count = number of strips
For each strip, the byte offset from the beginning of the file to
the start of that strip.
No default, must be specified.
The TIFF fields listed in this section SHALL be used by all fax modes, but the values associated with them depend on the mode being described and the associated ITU Recommendations. Therefore, only the fields are defined here; the values applicable to a particular fax mode are described in Sections 3-8. Fields that SHALL be used by some but not all modes are given in the section (3-8) describing the mode that uses them.
BitsPerSample(258) SHORT
RequiredByTIFFBaseline
Number of bits per image sample
Default = 1 (field may be omitted if this is the value)
Compression(259) SHORT
RequiredByTIFFBaseline
Compression method used for image data
Default = 1 (no compression, so may not be omitted for FAX)
FillOrder(266) SHORT
RequiredByTIFFforFax
The default bit order in Baseline TIFF per [TIFF] is indicated by
FillOrder=1, where bits are not reversed before being stored.
However, TIFF for Fax typically utilizes the setting of FillOrder=2,
where the bit order within bytes is reversed before storage (i.e.,
bits are stored with the Least Significant Bit first).
Default = 1 (field may be omitted if this is the value)
Facsimile data appears on the phone line in bit-reversed order
relative to its description in the relevant ITU compression
Recommendation. Therefore, a wide majority of facsimile
implementations choose this natural order for storage. Nevertheless,
all readers conforming to this specification must be able to read
data in both bit orders.
ImageWidth(256) SHORT or LONG
RequiredByTIFFBaseline
The number of pixels (columns) per scanline (row) of the image
No default, must be specified.
NewSubFileType(254) LONG
RequiredByTIFFforFAX
A general indication of the kind of data contained in this IFD
Bit 1 is 1 if the image is a single page of a multi-page document.
Default = 0 (no subfile bits on, so may not be omitted for FAX)
PhotometricInterpretation(262) SHORT
RequiredByTIFFBaseline
The color space of the image data
No default, must be specified
ResolutionUnit(296) SHORT
RequiredByTIFFBaseline
The unit of measure for resolution. 2 = inch, 3 = centimeter;
Default = 2 (field may be omitted if this is the value)
SamplesPerPixel(277) SHORT
RequiredByTIFFBaseline
The number of color components per pixel; SamplesPerPixel is 1 for a
black-and-white, grayscale or indexed (palette) image.
Default =1 (field may be omitted if this is the value)
XResolution(282) RATIONAL
RequiredByTIFFBaseline
The horizontal resolution of the image in pixels per resolution
unit. The ITU-T Recommendations for facsimile specify a small number
of horizontal resolutions: 100, 200, 300, 400 pixels per inch, and
80, 160 pixels per centimeter (or 204, 408 pixels per inch). The
allowed XResolution values for each mode are given in the section
defining that mode. Per [T.4], it is permissible for applications to
treat the following XResolution values as being equivalent: <204,
200> and <400,408> in pixels/inch. These equivalencies were allowed
by [T.4] to permit conversions between inch and metric based
facsimile terminals.
TIFF for Facsimile Writers SHOULD express XResolution in inch based
units, for consistency with historical practice and to maximize
interoperability. See the table below for information on how to
convert from an ITU-T metric value to its inch based equivalent
resolution.
No default, must be specified
YResolution(283) RATIONAL
RequiredByTIFFBaseline
The vertical resolution of the image in pixels per resolution unit.
The ITU-T Recommendations for facsimile specify a small number of
vertical resolutions: 100, 200, 300, 400 pixels per inch, and 38.5,
77, 154 pixels per centimeter (or 98, 196, 391 pixels per inch). The
allowed YResolution values for each mode are given in the section
defining that mode. Per [T.4], it is permissible for applications to
treat the following YResolution values as being equivalent: <98,
100>, <196, 200>, and <391, 400> in pixels/inch. These equivalencies
were allowed by [T.4] to permit conversions between inch and metric
based facsimile terminals. TIFF for Facsimile Writers SHOULD express
YResolution in inch based units, for consistency with historical
practice and to maximize interoperability. See the table below for
information on how to convert from an ITU-T metric value to its inch
based equivalent resolution. No default, must be specified
+-----------------------------+-----------------------------+
| XResolution | YResolution |
+--------------+--------------+--------------+--------------+
|ResolutionUnit|ResolutionUnit|ResolutionUnit|ResolutionUnit|
| =2 (inch) | =3 (cm) | =2 (inch) | =3 (cm) |
+--------------+--------------+--------------+--------------+
| 100 | | 100 | |
+--------------+--------------+--------------+--------------+
| 204 | 80 | 98 | 38.5 |
| 200 | | 100 | |
+--------------+--------------+--------------+--------------+
| 204 | 80 | 196 | 77 |
| 200 | | 200 | |
+--------------+--------------+--------------+--------------+
| 204 | 80 | 391 | 154 |
+--------------+--------------+--------------+--------------+
| 300 | | 300 | |
+--------------+--------------+--------------+--------------+
| 408 | 160 | 391 | 154 |
| 400 | | 400 | |
+--------------+--------------+--------------+--------------+
The TIFF fields listed in this section MAY be used by all fax modes. However, Profile S writers (the minimal fax mode described in Section 3) SHOULD NOT use these fields. Recommended fields that are mode- specific are described in Sections 3-8.
DateTime(306) ASCII
OptionalInTIFFBaseline
Date/time of image creation in 24-hour format "YYYY:MM:DD HH:MM:SS".
No default.
DocumentName(269) ASCII
OptionalInTIFFExtension(DocumentStorageAndRetrieval)
The name of the scanned document. This is a TIFF extension field,
not a Baseline TIFF field.
No default.
ImageDescription(270) ASCII
OptionalInTIFFBaseline
A string describing the contents of the image.
No default.
Orientation(274) = 1-8. SHORT
OptionalinTIFFBaseline
1: 0th row represents the visual top of the image; the 0th column
represents the visual left side of the image. See the current TIFF
spec [TIFF] for further values; Baseline TIFF only requires value=1.
Default = 1.
Note: It is recommended that a writer that is aware of the
orientation will include this field to give a positive indication of
the orientation, even if the value is the default. If the
Orientation field is omitted, the reader SHALL assume a value of 1.
Software(305) ASCII
OptionalInTIFFBaseline
The optional name and release number of the software package that
created the image.
No default.
The new TIFF fields listed in this section MAY be used by all fax modes, but their support is not expected for the minimal fax mode described in Section 3. In addition, support for these new TIFF fields has not been included in historical TIFF-F readers described in Section 4 and [TIFF-FY]. These fields describe "global" parameters of the fax session that created the image data. They are optional, not part of the current TIFF specification, and are defined in this document.
The first new field, GlobalParametersIFD, is an IFD that contains global parameters and is located in a Primary IFD.
GlobalParametersIFD (400) IFD
An IFD containing global parameters. It is recommended that a TIFF
writer place this field in the first IFD, where a TIFF reader would
find it quickly.
Each field in the GlobalParametersIFD is a TIFF field that is legal
in any IFD. Required baseline fields should not be located in the
GlobalParametersIFD, but should be in each image IFD. If a conflict
exists between fields in the GlobalParametersIFD and in the image
IFDs, then the data in the image IFD shall prevail.
Among the GlobalParametersIFD entries is a new ProfileType field
which generally describes information in this IFD and in the TIFF
file.
ProfileType(401) LONG
The type of image data stored in this IFD.
0 = Unspecified
1 = Group 3 fax
No default
The following new global fields are defined in this document as IFD
entries for use with fax applications.
FaxProfile(402) = 0 - 6. BYTE
The profile that applies to this file; a profile is subset of the
full set of permitted fields and field values of TIFF for facsimile.
The currently defined values are:
0: does not conform to a profile defined for TIFF for facsimile
1: minimal black & white lossless, Profile S
2: extended black & white lossless, Profile F
3: lossless JBIG black & white, Profile J
4: lossy color and grayscale, Profile C
5: lossless color and grayscale, Profile L
6: Mixed Raster Content, Profile M
CodingMethods(403) LONG
This field indicates which coding methods are used in the file. A
bit value of 1 indicates which of the following coding methods is
used:
Bit 0: unspecified compression,
Bit 1: 1-dimensional coding, ITU-T Rec. T.4 (MH - Modified Huffman),
Bit 2: 2-dimensional coding, ITU-T Rec. T.4 (MR - Modified Read),
Bit 3: 2-dimensional coding, ITU-T Rec. T.6 (MMR - Modified MR),
Bit 4: ITU-T Rec. T.82 coding, using ITU-T Rec. T.85 (JBIG),
Bit 5: ITU-T Rec. T.81 (Baseline JPEG),
Bit 6: ITU-T Rec. T.82 coding, using ITU-T Rec. T.43 (JBIG color),
Bits 7-31: reserved for future use
Note: There is a limit of 32 compression types to identify standard
compression methods.
VersionYear(404) BYTE
Count: 4
The year of the standard specified by the FaxProfile field, given as
4 characters, e.g. '1997'; used in lossy and lossless color modes.
ModeNumber (405) BYTE
The mode of the standard specified by the FaxProfile field. A
value of 0 indicates Mode 1.0; used in Mixed Raster Content mode.
| ZVON > RFC Repository > RFC 2301 |
| Frontpage / Contents |
| Prev | Next | RFC index | RFC search | Download as zip/tar.gz |