Image::MetaData::JPEG::Structures(3pm) User Contributed Perl Documentation Image::MetaData::JPEG::Structures(3pm)
NAME
Image::MetaData::JPEG::Structures - This document describes the structure of a JPEG file; it is an appendix to the main manual page of the
Image::MetaData::JPEG module, which the reader should refer to for further details and the general scope.
THE STRUCTURE OF JPEG PICTURES
The JPEG (Joint Photographic Experts Group) ISO standard is a lossy image compression mechanism developed before 1990, designed for
compressing colour or gray-scale photo-realistic images (it performs poorly on line-art drawings). The JPEG mechanism is designed to
exploit known limitations of the human eye, in particular the fact that colour changes are perceived less accurately than changes in
brightness. Note that although the "baseline" JPEG format is patent-free, there are patents associated with some optional features, namely
arithmetic coding and hierarchical storage. For this reason, these optional features are never used on the Web (good job, patent fans ...).
JPEG refers only to a class of compression algorithms, not to a specific file format. In order to produce files with embedded JPEG streams,
a number of file format (semi-)standards have been adapted or devised, some of which are listed in the format section. The structure of a
well formed JPEG file can be described by the following pseudo production rules (for sake of simplicity, some additional constraints
between tables and SOF segments are neglected).
JPEG --> (SOI)(misc)*(image)?(EOI)
(image) --> (hierarch.)|(non-hier.)
(hierarch.) --> (DHP)(frame)+
(frame) --> (misc)*(EXP)?(non-hier.)
(non-hier.) --> (SOF)(scan)+
(scan) --> (misc)*(SOS)(data)*(ECS)(DNL)?
(data) --> (ECS)(RST)
(misc) --> (DQT)|(DHT)|(DAC)|(DRI)|(COM)|(APP)
(SOI) = Start Of Image
(EOI) = End Of Image
(SOF) = Start Of Frame header (10 types)
(SOS) = Start Of Scan header
(ECS) = Entropy Coded Segment (row data, not a real segment)
(DNL) = Define Number of Lines segment
(DHP) = Define Hierarchical P??? segment
(EXP) = EXPansion segment
(RST) = ReSTart segment (8 types)
(DQT) = Define Quantisation Table
(DHT) = Define Huffman coding Table
(DAC) = Define Arithmetic coding Table
(DRI) = Define Restart Interval
(COM) = COMment segment
(APP) = APPlication segment
This package does not check that a JPEG file is really correct; it accepts a looser syntax, were segments and ECS blocks are just
contiguous (basically, because it does not need to display the image!). All meta-data information is concentrated in the (COM*) and (APP)
Segments, exception made for some records in the (SOF*) segment (e.g. image dimensions).
Reference: B<"Digital compression and coding of continuous-tone still
images: requirements and guidelines", CCITT recommend. T.81, 1992,
The International Telegraph and Telephone Consultative Committee,
standard ISO/IEC IS 10918-1 or ITU-T Recommendation T.81>.
Also: B<"The JPEG still picture compression standard", G.K.Wallace,
1991, IEEE Transactions on Consumer Electronics>.
A SHORT HISTORY OF JPEG FILE FORMATS
TIFF [not supported]
The TIFF (Tagged Image File Format) is one of the most popular and flexible raster file formats for digital images, and is the de-facto
standard graphics format for high colour depths. It was initially developed by Aldus (in 1986) and is now maintained (or neglected) by
Adobe. Multiple images (e.g., fax pages) can be stored in a single TIFF file, each page in a separate IFD (Image File Directory), all IFD's
being linked in a linear structure. The designers of TIFF wanted to be able to include all sorts of image information in the file (tags),
like geometry, size, data arrangement and various compression options. However, lack of full TIFF support by the first publicly available
readers resulted in a "reduced standard", so that even today most TIFF files contain only 32-bit uncompressed images. Support for JPEG
streams was included in version 6.0 (in 1992), but it was badly designed, so that JPEG/TIFF files are nowadays very uncommon (some changes
were proposed in various technical notes in 1995 and 2002, but they never made it to version 7.0).
Reference: B<"TIFF (TM)", revision 6.0, June 1992, Adobe Developers
Association, Adobe Systems Incorporated, Mountain View, CA, USA>.
Also: B<"Adobe PageMaker 6.0: TIFF Technical Notes", September 1995>,
B<"Adobe Photoshop: TIFF Technical Notes", March 2002>.
JFIF
The JFIF (JPEG File Interchange Format) was created by Eric Hamilton at C-Cube Microsystems in 1991 for storing JPEG-encoded data, and put
into the public domain together with example reference software. JFIF is designed to allow files containing JPEG-encoded data streams to be
exchanged between otherwise incompatible systems and applications. A JFIF file basically consists of a JPEG file with an APP0 segment of
JFIF or JFXX type at the beginning, providing information missing from the JPEG stream: version number, horizontal and vertical pixel
density, pixel aspect ratio and an optional thumbnail (extended JFIF files can contain also compressed thumbnails). JFIF files conform to
the general file interchange specifications in ISO/IEC 10918-1.
Reference: B<"JPEG File Interchange Format", version 1.02, September 1992,
Eric Hamilton, C-Cube Microsystems, Milpitas, CA, USA>.
SPIFF [not supported yet]
The SPIFF (Still Picture Interchange File Format) is an official file format released in 1996 by the Joint Photographic Experts Group. It
can contain image data streams encoded with a variety of compression mechanism, including JPEG and JBIG. When JPEG was standardised,
disagreements among ISO committees prevented a standard JPEG file format from being created. The de-facto format that appeared was JFIF,
with which a JPEG/SPIFF file is backward compatible. In addition to the image data, SPIFF includes information necessary to render it on
common output devices, within the constraints imposed by that device. SPIFF appears not to be widely adopted; most producers prefer simple
JFIF files.
Reference: B<ISO/IEC IS 10918-1, part 3> (extensions to the standard).
CIFF [not supported yet]
The CIFF (Camera Image File Format) is a JPEG-based image file format developed by some camera makers in 1997, and constitutes an evolution
of the JFIF. A CIFF file has a standard JFIF APP0 segment, immediately followed by another APP0 segment holding information specific to a
particular digital camera in a particular format. CIFF was not at all successful and was rapidly replaced by DCF/Exif.
Reference: B<"CIFF: Specification on Image Data File",
version 1.0 rev.4, December 1997, Canon corporation>.
Exif and DCT
The Exif (Exchangeable image file format), developed by the JEITA (1996 v.1.0, 1997 v.1.1, 1998 v.2.1, 2002 v.2.2) standard was aimed at
realizing a common format for the image files used with digital still cameras and other related equipment. Exif is paired with a set of
naming conventions and directory layout for files in a camera memory (DCF or Design rule for Camera File system, 1999). An Exif file is a
valid JPEG file with application marker segments (APP1 and APP2) inserted. Uncompressed files are recorded in TIFF-6 format. Related
attribute information is stored in the tag information format defined in TIFF-6 (i.e., in a chain of IFD's in the APP1 segment).
Information specific to the camera system and not defined in TIFF is stored in private tags registered for Exif. The Exif image file
specification also specifies the method for recording thumbnails. The APP2 segment is used when recording Flashpix extensions. Exif is the
most common JPEG file format used by digital cameras nowadays.
Reference: B<JEITA CP-3451, "Exchangeable image file format for digital
still cameras: Exif Version 2.2", April 2002, Japan Electronics
and Information Technology Industries Association>.
Structure of a JFIF APP0 segment
APP0 segments are used in the old JFIF standard to store information about the picture dimensions and an optional thumbnail. The format of
a JFIF APP0 segment is as follows (note that the size of thumbnail data is 3n, where n = Xthumbnail * Ythumbnail, and it is present only if
n is not zero; only the first 8 records are mandatory):
[Record name] [size] [description]
---------------------------------------
Identifier 5 bytes ("JFIF 00" = 0x4a46494600)
MajorVersion 1 byte major version (e.g. 0x01)
MinorVersion 1 byte minor version (e.g. 0x01 or 0x02)
Units 1 byte units (0: densities give aspect ratio
1: density values are dots per inch
2: density values are dots per cm)
Xdensity 2 bytes horizontal pixel density
Ydensity 2 bytes vertical pixel density
Xthumbnail 1 byte thumbnail horizontal pixel count
Ythumbnail 1 byte thumbnail vertical pixel count
ThumbnailData 3n bytes thumbnail image
There is also an extended JFIF (only possible for JFIF versions 1.02 and above). In this case the identifier is not JFIF but JFXX. This
extension allows for the inclusion of differently encoded thumbnails. The syntax in this case is modified as follows:
[Record name] [size] [description]
---------------------------------------
Identifier 5 bytes ("JFXX 00" = 0x4a46585800)
ExtensionCode 1 byte (0x10 Thumbnail coded using JPEG
0x11 Thumbnail using 1 byte/pixel
0x13 Thumbnail using 3 bytes/pixel)
Then, depending on the extension code, there are other records to define the thumbnail. If the thumbnail is coded using a JPEG stream, a
binary JPEG stream immediately follows the extension code (the byte count of this file is included in the byte count of the APP0 Segment).
This stream conforms to the syntax for a JPEG file (SOI .... SOF ... EOI); however, no 'JFIF' or 'JFXX' marker Segments should be present:
[Record name] [size] [description]
---------------------------------------
JPEGThumbnail ... bytes a variable length JPEG picture
If the thumbnail is stored using one byte per pixel, after the extension code one should find a palette and an indexed RGB. The records are
as follows (remember that n = Xthumbnail * Ythumbnail):
[Record name] [size] [description]
---------------------------------------
Xthumbnail 1 byte thumbnail horizontal pixel count
YThumbnail 1 byte thumbnail vertical pixel count
ColorPalette 768 bytes 24-bit RGB values for the colour palette
(defining the colours represented by each
value of an 8-bit binary encoding)
1ByteThumbnail n bytes 8-bit indexed values for the thumbnail
If the thumbnail is stored using three bytes per pixel, there is no colour palette, so the previous fields simplify into:
[Record name] [size] [description]
---------------------------------------
Xthumbnail 1 byte thumbnail horizontal pixel count
YThumbnail 1 byte thumbnail vertical pixel count
3BytesThumbnail 3n bytes 24-bit RGB values for the thumbnail
Structure of an Exif APP1 segment
Exif (Exchangeable Image File format) JPEG files use APP1 segments in order not to conflict with JFIF files (which use APP0). Exif APP1
segments store a great amount of information on photographic parameters for digital cameras and are the preferred way to store thumbnail
images nowadays. They can also host an additional section with GPS data. The reference document for Exif 2.2 and the Interoperability
standards are respectively:
B<"Exchangeable image file format for digital still cameras:
Exif Version 2.2", JEITA CP-3451, Apr 2002
Japan Electronic Industry Development Association (JEIDA)>
B<"Design rule for Camera File system", (DCF), v1.0
English Version 1999.1.7, Adopted December 1998
Japan Electronic Industry Development Association (JEIDA)>
The TIFF (Tagged Image File format) standard documents, as well as some updates and corrections, are also useful:
B<- "TIFF(TM) Revision 6.0, Final", June 3, 1992, Adobe Devel. Association
- ISO 12639, "Graphic technology -- Prepress digital data exchange
-- Tag image file format for image technology (TIFF/IT)"
- ISO 12234-2, "Electronic still-picture imaging -- Removable memory
-- Part 2: TIFF/EP image data format"
- DRAFT - TIFF CLASS F, October 1, 1991
- DRAFT - TIFF Technical Note #2, 17-Mar-95 (updates for JPEG-in-TIFF)
- "Adobe Pagemaker 6.0 TIFF Technical Notes", (1,2,3 and OPI), 14-Sep-1995>
Exif APP1 segments are made up by an identifier, a TIFF header and a sequence of IFDs (Image File Directories) and subIFDs. The high level
IFDs are only two (IFD0, for photographic parameters, and IFD1 for thumbnail parameters); they can be followed by thumbnail data. The
structure is as follows:
[Record name] [size] [description]
---------------------------------------
Identifier 6 bytes ("Exif 00 00" = 0x457869660000), not stored
Endianness 2 bytes 'II' (little-endian) or 'MM' (big-endian)
Signature 2 bytes a fixed value = 42
IFD0_Pointer 4 bytes offset of 0th IFD (usually 8), not stored
IFD0 ... main image IFD
IFD0@SubIFD ... Exif private tags (optional, linked by IFD0)
IFD0@SubIFD@Interop ... Interoperability IFD (optional,linked by SubIFD)
IFD0@GPS ... GPS IFD (optional, linked by IFD0)
APP1@IFD1 ... thumbnail IFD (optional, pointed to by IFD0)
ThumbnailData ... Thumbnail image (optional, 0xffd8.....ffd9)
So, each Exif APP1 segment starts with the identifier string "Exif 00 00"; this avoids a conflict with other applications using APP1, for
instance XMP data. The three following fields (Endianness, Signature and IFD0_Pointer) constitute the so called TIFF header. The offset of
the 0th IFD in the TIFF header, as well as IFD links in the following IFDs, is given with respect to the beginning of the TIFF header (i.e.
the address of the 'MM' or 'II' pair). This means that if the 0th IFD begins (as usual) immediately after the end of the TIFF header, the
offset value is 8. An Exif segment is the only part of a JPEG file whose endianness is not fixed to big-endian.
If the thumbnail is present it is located after the 1st IFD. There are 3 possible formats: JPEG (only this is compressed), RGB TIFF, and
YCbCr TIFF. It seems that JPEG and 160x120 pixels are recommended for Exif ver. 2.1 or higher (mandatory for DCF files). Since the segment
size for a segment is recorded in 2 bytes, thumbnails are limited to slightly less than 64KB.
Each IFD block is a structured sequence of records, called, in the Exif jargon, Interoperability arrays. The beginning of the 0th IFD is
given by the 'IFD0_Pointer' value. The structure of an IFD is the following:
[Record name] [size] [description]
---------------------------------------
2 bytes number n of Interoperability arrays
12n bytes the n arrays (12 bytes each)
4 bytes link to next IFD (can be zero)
... additional data area
The next_link field of the 0th IFD, if non-null, points to the beginning of the 1st IFD. The 1st IFD as well as all other sub-IFDs must
have next_link set to zero. The thumbnail location and size is given by some interoperability arrays in the 1st IFD. The structure of an
Interoperability array is:
[Record name] [size] [description]
---------------------------------------
2 bytes Tag (a unique 2-byte number)
2 bytes Type (one out of 12 types)
4 bytes Count (the number of values)
4 bytes Value Offset (value or offset)
The possible types are the same as for the Record class, exception made for nibbles and references (see "Managing a JPEG Record object" in
Image::MetaData::JPEG). Indeed, the Record class is modelled after interoperability arrays, and each interoperability array gets stored as
a Record with given tag, type, count and values. The "value offset" field gives the offset from the TIFF header base where the value is
recorded. It contains the actual value if it is not larger than 4 bytes (32 bits). If the value is shorter than 4 bytes, it is recorded in
the lower end of the 4-byte area (smaller offsets). For further details see the section "Valid tags for Exif APP1 data" in
Image::MetaData::JPEG::TagLists.
Structure of an XMP APP1 segment
XMP (eXtensible Metadata Platform) JPEG files use APP1 segments in order to store metadata information; the storage format (serialisation)
is RDF (Resource Description Framework) implemented as an application of XML. XMP Exif APP1 segments can store all information stored in
Exif and IPTC segments, as well as custom and future schemas. The reference document for XMP 3.2 can be requested from Adobe Systems
Incorporated:
B<"XMP Specification", version 3.2, June 2005, Adobe Systems Inc.,
345 Park Avenue, San Jose, CA 95110-2704, L<http://www.adobe.com>>
B<See also: L<http://xml.coverpages.org/xmp.html>>
B<See also: L<http://www.w3.org/TR/REC-rdf-syntax/>>
XMP APP1 segments are made up by an identifier and a Unicode XMP packet (the encoding is usually UTF-8, but it can also be UTF-16 or
UTF-32, both big-endian or little-endian). The packet cannot be split in multiple segments, so there is a maximum size of approximately
64KB. The structure is very simple: a fixed XMP namespace URI (null terminated, and without quotation marks) followed by the XMP packet:
[Record name] [size] [description]
---------------------------------------
Identifier 29 bytes http://ns.adobe.com/xap/1.0/ 00
<XMP packet> ... the actual Unicode XMP packet
The packet content is sandwiched between a header and a trailer, and may contain padding whitespaces at the end. The xpacket header has two
mandatory attributes, begin and id (order is important), separated by exactly one space (U+0020). Attribute values, here and in the
following, are enclosed by single quotes (U+0027) or double quotes (U+0022). The value of begin must be the Unicode zero-width non-breaking
space (U+FEFF), to be used as a byte-order marker during blind scans; an empty value is also acceptable (for backward compatibility), and
stands for UTF-8. The value of id is fixed. Other attributes, like (deprecated) bytes or encoding, may be ignored. A padding of 2KB or 4KB,
with a newline every 100 spaces, is recommended. The end attribute of the trailer may have a value of "r" (read-only) or "w" (modifiable).
Header <?xpacket begin="..." id="W5M0MpCehiHzreSzNTczkc9d" ...?>
Content ... serialised XMP data (see later) ...
Padding ... padding with XML whitespaces ...
Trailer <?xpacket end="w"?>
The structure of the packet content is as follows. There is an optional x:xmpmeta (or x:xapmeta for older files) element, with a mandatory
xmlns:x attribute set to "adobe:ns:meta/" and other optional attributes, which can be ignored (currently, Adobe's toolkit stores its
version in x:xmptk). Inside it (or at top level, if it is absent), there is exactly one rdf:RDF element with an attribute specifying the
xmlns:rdf namespace (other namespaces, whose role is not clear to me, can be listed here as additional attributes). Inside the rdf:RDF
element then, all XMP properties are stored inside one or more rdf:Description elements:
<x:xmpmeta xmlns:x='adobe:ns:meta/' ..opt.attributes..>
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#">
<rdf:Description ...> .... </rdf:Description>
<rdf:Description ...> .... </rdf:Description>
.........
<rdf:Description ...> .... </rdf:Description>
</rdf:RDF>
</x:xmpmeta>
rdf:Description elements and schemas are usually in one-to-one correspondence, although this is just a convention and not a rule. Each
element has two mandatory attributes, rdf:about and xmlns:NAME. The rdf:about attribute is usually empty (it can however contain an
application specific URI), and its value must be shared among all rdf:Description elements. The xmlns:NAME attribute specifies the local
namespace prefix (NAME stands for the actual prefix). Additional namespaces can be specified via xmlns attributes.
<rdf:Description rdf:about="" xmlns:NAME="...">
<NAME:propname1 ...> .... </NAME:propname1>
<NAME:propname2 ...> .... </NAME:propname2>
.........
<NAME:propnameN ...> .... </NAME:propnameN>
</rdf:Description>
Properties can come in various flavours. A simple property is just some literal value between opening and closing tags carrying the
property name; there exists also an abbreviated form where properties are listed as attributes of the rdf:Description tag (in this case
there is no closing rdf:Description tag, and the opening tags ends with the '/' character). Simple properties can have qualifiers
(attributes). This is an example of a description element with only simple properties:
<rdf:Description rdf:about="" xmlns:xap="http://ns.adobe.com/xap/1.0/">
<xap:CreatorTool>Adobe Photoshop Elements 3.0</xap:CreatorTool>
<xap:ModifyDate>2005-05-12T44:39:11+03:00</xap:ModifyDate>
</rdf:Description>
or
<rdf:Description rdf:about="" xmlns:xap="http://ns.adobe.com/xap/1.0/"
xap:CreatorTool="Adobe Photoshop Elements 3.0"
xap:ModifyDate="2005-06-13T23:38:11+02:00"/>
A structured property is characterised by a rdf:Description block instead of a simple content. The inner description block, which does not
allow for an rdf:about attribute, contains one or more named properties (which can, of course, contain further substructures). This is an
example of a description element with one structured property:
<rdf:Description rdf:about="" xmlns:xmpTPg="..an URI..">
<xmpTPg:MaxPageSize>
<rdf:Description xmlns:stDim="..another URI.."> ---. Content
<stDim:w>21</stDim:w> | of the
<stDim:h>27</stDim:h> | MaxPageSize
<stDim:unit>centimetres</stDim:unit> | structured
</rdf:Description> -----------------' property
</xmpTPg:MaxPageSize>
</rdf:Description>
An array property is, on the other hand, just an array of same-type properties. It can come in three varieties: as a Bag (an unordered
set), as a Sequence (an ordered list) and as an Alternative list (see later). As for structured properties, each item in the array is a
property and can contain further substructures. The item list is delimited by an rdf:Bag, an rdf:Seq, or an rdf:Alt tag pair, respectively.
Each item is delimited by a pair of rdf:li tags. The following example illustrates an unordered array property with three items:
<rdf:Description rdf:about="" xmlns:dc="..an URI..">
<dc:subject>
<rdf:Bag> ----------------. Content
<rdf:li>metadata</rdf:li> | of the
<rdf:li>schema</rdf:li> | unordered
<rdf:li>XMP</rdf:li> | array
</rdf:Bag> ----------------' dc:subject
</dc:subject>
</rdf:Description>
A simple property (but, currently, not a structured or array property) can have qualifiers. For instance, items in an array property can be
qualified; in this case, the content of rdf:li is not text, but an rdf:Description (without attributes) containing the actual value (as
rdf:value) and the qualifiers as additional properties; a qualifier name belongs to a secondary namespace, specified as an attribute of the
outer rdf:Description. The following example illustrates the use of property qualifiers:
<rdf:Description rdf:about="" xmlns:dc="..1stURI.." xmlns:ns="..2ndURI..">
<dc:creator>
<rdf:Seq>
<rdf:li> ----------------. An element
<rdf:Description> | of the
<rdf:value>William Gilbert</rdf:value> | ordered array
<ns:role>lyricist</ns:role> | dc:creator
</rdf:Description> | with a
</rdf:li> ----------------' qualifier
<rdf:li>
<rdf:Description>
<rdf:value>Arthur Sullivan</rdf:value>
<ns:role>composer</ns:role>
</rdf:Description>
</rdf:li>
</rdf:Seq>
</dc:creator>
</rdf:Description>
Text properties in an alternative array property may have a special qualifier (the xml:lang property) specifying the language of the text.
In this case, the serialisation is different: xml:lang for each item of the array becomes an attribute of the rdf:li tag. The default value
is indicated by a special value of the qualifier, or is taken from the first item in the list. The following example illustrates an array
of language alternatives:
<xmp:Title>
<rdf:Alt> ----------------. List of
<rdf:li xml:lang="x-default">Title</rdf:li> | alternatives
<rdf:li xml:lang="en-us">Title</rdf:li> | corresponding
<rdf:li xml:lang="fr-fr">Titre</rdf:li> | to xmp:Title;
<rdf:li xml:lang="it-it">Titolo</rdf:li> | the language
</rdf:Alt> ----------------' is xml:lang
</xmp:Title>
The interpretation of the values of properties conforming to the previously described syntax is done with the aid of the schema
corresponding to the property namespace. The most common schemas in JPEG files are the following ones (note, however, that XMP is
intrinsically extensible, so this list will never be exhaustive):
Dublin Core schema =item XMP basic schema =item XMP rights management schema =item XMP media management schema =item XMP basic job ticket
schema =item XMP paged-text schema =item XMP dynamic media schema =item Adobe Photoshop schema =item Camera row schema =item Exif schemas
=item IPTC schema
Structure of a Photoshop-style APP13 segment
The Adobe's Photoshop program, a de-facto standard for image manipulation, uses the APP13 segment for storing non-graphic information, such
as layers, paths, IPTC data and more. The unit for this kind of information is called a "resource data block" (because they hold data that
was stored in the Macintosh's resource fork in early versions of Photoshop). The content of an APP13 segment is formed by an identifier
string (usually "Photoshop 3.0 00", but also 'Adobe_Photoshop2.5:', used by earlier versions, is accepted; in this case some additional
undocumented bytes are read (resolution info?) and saved in a root 'Resolution' record) followed by a sequence of resource data blocks; a
resource block has the following structure:
[Record name] [size] [description]
---------------------------------------
(Type) 4 bytes Photoshop uses '8BIM' from ver 4.0 on
(ID) 2 bytes a unique identifier, e.g., "