94
because it cannot be guaranteed that the
two variants have exactly the same con-
tent.
Resolution is not part of the PDF/A standard
Image resolution does not have a role to
play when it comes to compliance with the
PDF/A standard. 周is is because there is no
single image resolution that is considered
to be universally ‘correct’. For example,
screenshots tend to have a resolution of of
72 or 96 ppi (pixels per inch). A common
resolution for printing is 300 ppi, but it
would not make sense to increase the reso-
lution of a screenshot to the normal print-
ing resolution because it would not convey
any additional information to the user. If
the worst comes to the worst, ill-considered
increases in resolution can cause fuzzy
edges. 周e maximum sensible image reso-
lution for screenshots is 72 ppi.
Another area that o晴en deals in low res-
olutions is astronomical photography.
Some of the images transmitted to Earth by
telescopes such as the Hubble Space Tele-
scope are extremely grainy images of dis-
tant stars or galaxies. 周ese low-resolution
images are the best that can be achieved,
and it is, of course, quite possible to create
valid PDF/A documents from them.
Image resolution is not regulated by the
PDF/A standard – it is le晴 to the decision of
the creator of each PDF/A file. Users must
decide for themselves whether or not the
image resolution used is the best resolution
possible.
Permitted and prohibited compression
types
周e choice of image compression type – the
procedure used to minimize the quantity of
image data – is not entirely down to the user.
周ere are two types of pixel image: Half-
tone images (grayscale and color images)
and line art images that consist of only two
colors. Line art images can be compressed
for use in PDF/A using ‘CCITT Group 4’, a
technology that is effective and prevents loss
of data. Programs that use this compression
type include Acrobat Distiller and Acrobat
Professional.
周e choice of compression types for half-
tone images is greater, and not all types are
PDF/A-compliant.
Of the compression types that prevent
loss of data, LZW, a rather old compression
type, is prohibited. It was decided to pro-
hibit the use of this compression type be-
cause it was once protected by license. Since
the more modern ZIP compression type is
both permitted and prevents loss of data, it
is the recommended compression method
to be used for the compression of half-tone
images without data loss.
➔
‘X-Ray Stars in M15’ – N. White & L. Angelini (LHEA), GSFC, CXO, NASA; www.nasa.gov
Low image resolution is no problem for PDF/A: Since there are sub-
jects where only low resolution can be achieved, the PDF/A standard
does not regulate image resolution.
Color depths and grades
Black and white:
Line art image:
1 bit
2 grades
Continuous tone:
Grayscale:
8 bits
256 grades
Color/RGB:
24 bits 16.7 million grades
Color/CMYK:
32 bits 4.3 billion grades
Line art and black and white images only have two grades.
Half-tone images have different numbers of grades depending
on the color model (grayscale, RGB, or CMYK). The compression
options differ for black and white images and half-tone images.
There are two basic types of
compression: Lossless com-
pression and techniques that
can damage the quality of im-
ages to a lesser or greater ex-
tent (‘lossy’ compression).
PDF/A applications in everyday life
PDF/A
in a Nutshell
43
80
ZIP is not subject to license-related re-
strictions. LZW compression in a PDF can
also be replaced by ZIP compression later
on. 周e PDF Optimizer feature in Acrobat
is designed for this purpose.
JPEG was the first procedure to achieve
relatively high-quality results from com-
paratively small images, despite being sub-
ject to data loss. For this reason, the tri-
umph of PDF in some situations would not
have been possible without JPEG. JPEG en-
ables different file sizes. 周e image quality
can be set in steps from ‘minimum’ to
‘maximum’. If a high level of compression
is chosen, block artifacts form. Depending
on the nature of the image, they can be
clearly visible on the screen. In the case of
images with sharp edges (such as text), high
levels of compression can be particularly
awkward. However, just as for image reso-
lution, the user decides on the level of JPEG
compression; the PDF/A standard does not
make any specifications.
However, the standard does prohibit the
use of JPEG2000, a compression type that
was developed by the same group as JPEG
(the Joint Photographic Experts Group).
JPEG2000 was introduced for PDF with
Acrobat 6 (PDF 1.5). Because PDF/A-1a and
-1b are based on PDF 1.4, JPEG2000 is pro-
hibited simply because it was introduced
too late. However, the version of the stan-
dard that is currently being compiled,
PDF/A-2, will include JPEG2000. For the
moment, if a PDF document contains im-
ages compressed using JPEG2000, the user
can replace the prohibited compression
type with JPEG or ZIP using the PDF Opti-
mizer in order to achieve PDF/A compli-
ance. 周is checklist lists specifications for
images in PDF/A documents:
■■All images must be an integral part of
the PDF file in which they appear.
■■Alternate images are not permitted.
■■周e user must decide on the image reso-
lution and compression level (both factors
influence the image quality).
■■周e compression types LZW and
JPEG2000 are prohibited.
■■For more information on colors (includ-
ing colors in images) see the information
starting on page 46.
Transparency
Transparent objects are not allowed in
PDF/A-compliant documents. At the point
when the PDF/A standard was adopted,
Blocks: Zoomed view of JPEG artifacts re-
sulting from heavy compression.
JPEG compressions (magnified)
JPEG minimum
285 KB
(US Letter page)
JPEG medium
325 KB
(US Letter page)
JPEG high
405 KB
(US Letter page)
JPEG maximum
509 KB
(US Letter page)
The JPEG compression rate affects the image quality. The compression rate is down to the decision of the user,
and is not regulated by the PDF/A standard.
Transparency: The upper image above is transparent. The transparent
object can be recognized easily, since the background is visible through
the image. The lower image has an opaque foreground image.
44
PDF/A
in a Nutshell
PDF/A applications in everyday life
56
Adobe had not yet completely formulated
the algorithms for evaluating transparency
in a completely clear manner. As a result,
transparency is currently prohibited in the
PDF/A standard. 周is will change in
PDF/A-2.
Transparency can effect images, graphics,
and text. Transparent objects are not 100%
opaque – instead, their background shows
through, as is the case for glass or thin
parchment. Transparent objects cannot al-
ways be detected with the naked eye, espe-
cially since opacity can be as high as 99%.
Transparent elements do not only occur
if they are explicitly created. Certain pop-
ular design functions such as drop shad-
ows and so晴 edges can create ‘sneaky’
transparent objects. For example, many
PowerPoint presentations contain trans-
parent objects, even if they cannot be de-
tected at a glance. If text or other elements
are given drop shadows, transparent ob-
jects are created when they are converted
to PDF.
Another widely used aid in Office envi-
ronments is the function that allows text
to be highlighted with a digital marker
pen. 周is function is also available in Ac-
robat Professional. However, this also
causes transparent objects to be created in
PDF files. So how can such transparent
objects be avoided or removed later on?
周is is normally done by transparency
flattening (transparency reduction). 周is
procedure involves merging the transpar-
ent area and the background in a way that
retains the appearance of the image. In
addition to certain professional layout
programs that can carry out transparency
flattening in advance, this process can
also be carried out during PDF optimiza-
tion in Acrobat Professional.
➔
Risk of transparent objects: Drop shadows
and soft object edges can cause the cre-
ation of transparent elements in a PDF file.
This example is from a PowerPoint presen-
tation.
Transparency resulting from the use of a
highlighter: Use of the Highlight Text Tool
can cause the creation of transparent ob-
jects in a PDF document.
PDF/A applications in everyday life
PDF/A
in a Nutshell
45
63
When flattening transparency, the user
can choose between different quality levels
(from low resolution to high resolution),
since this process generates new images out
of overlapping graphic objects.
However, users must be careful when re-
moving highlighted text. Instead of using
transparency flattening, which would make
the yellow highlighting opaque and hide
the text, the Acrobat PDF Optimizer func-
tion ‘Discard all comments, forms and
multimedia’ should be used. 周is function
can be called from the ‘Discard User Data’
area.
Colors
周e colors of illustrations and graphics in a
document should always appear exactly the
same – whether displayed on one’s own
monitor, on a colleague’s monitor, or viewed
as a printout. Nothing is more annoying
than a company logo that, when used in a
presentation or brochure, fails to depict the
corporate identity because, for example, it
appears orange rather than magenta.
周anks to PDF/A, such problems are a
thing of the past, since the PDF/A stan-
dard guarantees the reliable reproduction
of colors for text, image, and graphical el-
ements.
Color management
PDF/A uses color management to safely de-
pict colors. Color management is based on
the use of color profiles that are appended
to image files, graphical documents, and
PDF files to act as a kind of instruction
manual.
周e RGB color space is widespread in Of-
fice environments. sRGB (‘Standard RGB’)
is now being used to enable colors to be dis-
played or printed as reliably as possible on
different devices and printers. 周e sRGB
profile is suitable for images, graphical ele-
ments, and text in Office documents. It was
developed by Hewlett-Packard and Micro-
so晴 in 1996 to make printed pages as simi-
lar to those displayed on the screen as pos-
Which color should it be? Without color management, the correct
depiction of colors in company logos is a question of luck.
PDF Optimizer: This Acrobat function en-
ables transparency flattening with differ-
ent quality settings. To ensure PDF/A
compliance, it is important to select a
compatibility level no higher than Acro-
bat 5 (PDF 1.4).
Hidden text: Transparency flattening
should not be used for highlighted text. It
is better to use the ‘Discard all comments,
forms and multimedia’ function, since the
Highlight Text Tool is a comments tool.
46
PDF/A
in a Nutshell
PDF/A applications in everyday life
61
sible. Common modern monitors and
printers support sRGB color adjustment.
Adobe RGB is another widespread RGB
profile. It was published by Adobe Systems
in 1998. 周is profile is most useful to peo-
ple who work with digital photographs,
since cyan and green tones appear to be
more natural with Adobe RGB than with
sRGB. For documents always intended for
four-color printing (production or digital
printing), the ISO Coated color profile con-
stitutes a good choice.
Output intent (output condition)
周e profiles named above (and other pro-
files) can be passed to the conversion pro-
cess along with each individual object
placed in a document, but there is another,
more practical procedure, that is applied to
the entire PDF/A file. An output intent (the
intended output condition) can be speci-
fied for the PDF/A conversion process. For
example, if a PDF/A file is to be archived for
the purpose of being displayed on a moni-
tor later on, the sRGB profile, which is a
standard part of PDF/A converters such as
Acrobat, Preflight, and pdfaPilot, is ideal.
On the other hand, PDF/A files that are in-
tended for printing can be given an ISO
Coated profile.
If the output condition of a PDF/A file
changes at any point in the future, color
conversion processes can be triggered. ■
Output intent: What is the purpose of the PDF/A document? In this
case, sRGB is the chosen output intent. Acrobat (Preflight) and other
converters deliver a range of profiles. In addition, users have access
to other profiles stored on their computers.
The incorrect reproduction of colors can
sometimes affect the message of an im-
age: Was the evening spent at the lake de-
picted in these two photographs a warm
evening or a cool one?
Safe depiction of colors in PDF/A
- If device-dependent colors are used, an output intent
must be specified.
- If there is already a source profile for all colors used,
there is no need to specify an output intent.
- If an output intent is used, if must have one output
profile only.
- Objects such as images and graphics can exist in different
color spaces (RGB, CMYK, spot colors, grayscale and Lab).
- It is also possible to use a single device-dependent col-
or space (with no ICC profile).
- Device-dependent CMYK and device-dependent RGB
may not be used together. If device-dependent colors
are used, there must be an output intent for the same
color space (RGB, CMYK, or Gray). However, only one
output intent color space may be used.
PDF/A applications in everyday life
PDF/A
in a Nutshell
47
48
Fonts
If a PDF file contains text that uses fonts –
that is, text that has not been converted to
paths/pixel-image text – there are a range
of specifications for achieving PDF/A-com-
pliance. 周e specifications for PDF/A-1a
and PDF/A-1b are different. However, we
shall first deal with the common specifica-
tions for both standards.
Embedding fonts
周e following applies to both compliance
levels – PDF/A-1a and -1b: All used fonts
must be embedded into the PDF file in
question. If this were not the case, text dis-
played on a computer that does not have
the font used might not be displayed in its
entirety. 周is is incompatible with the re-
quired visual reproducibility. 周e entire
font does not need to be embedded; it is
sufficient to embed only the characters that
are used in the document. 周is is known as
‘embedding subsets’.
In the light of the international exchange
of documents containing special characters
that the recipient might well not have on
his or her computer, the use of embedded
fonts is a significant advantage. Modern
operating systems increasingly provide Cy-
rillic, Asian, and Eastern-European fonts
in order to enable the display of interna-
Global depiction of characters: PDF/A en-
sures that international texts are dis-
played properly, since the embedding of
texts in documents means that all required
characters are actually available within
the document in which they are used.
photocase.com/de
Because entire fonts can con-
stitute large datasets, PDF/A
permits the embedding of sub-
sets. This means that only
characters used in a particular
PDF document are embedded.
This limits the file size.
48
PDF/A
in a Nutshell
PDF/A applications in everyday life
77
tional Internet pages, but there is no guar-
antee that the fonts delivered will be the
fonts used by the creator of a particular
PDF document.
Unlike in the early days of PDF, embed-
ding fonts with the current program ver-
sions of Acrobat and many other profes-
sional tools for creating PDFs is not diffi-
cult. However, even today there are solu-
tions – even at industry level – that do not
fulfill the font handling specifications stip-
ulated by the PDF/A standard.
Fonts must be uniquely encoded
Problems with character set encodings can
cause individual font glyphs to be displayed
incorrectly or not at all in documents in-
cluding Word documents and e-mails.
‘Glyphs’ are the graphical depictions of
characters.
What might happen if character set en-
coding is inconsistent? When the euro was
introduced, problems were o晴en experi-
enced with the € sign. 周e characters ‘ä’,
‘ü’, and ‘ö’ o晴en cause difficulties in inter-
national communication. 周e PDF/A stan-
dard now requires glyphs that are used in
documents to be uniquely encoded in order
to guarantee correct reproducibility.
Overlapping letters such as those that
can occur when copying text are also elim-
inated by compliance with the PDF/A stan-
dard. 周e gobbledegook shown here is
caused by missing tracking information.
周is problem cannot occur if PDF/A is
used.
Unique characters with PDF/A-1a – thanks to
Unicode
In addition to the points mentioned above,
a further font requirement applies to
PDF/A-1a. All of the characters embedded
in a PDF/A-1a file must be uniquely identi-
fiable by means of their Unicode name.
Unicode is an international standard that
assigns a unique ID number to every char-
acter and symbol that exists worldwide
(even for historic script). 周e Unicode Con-
sortium and ISO work together on this
project. Unicode encodes only abstract
characters, not glyphs (the various graphi-
cal depictions of letters).
周e use of Unicode encodings for PDF/A-
1a brings the advantage of all character-
based text being completely unique. 周is
enables text to be searched precisely and re-
liably for content as well as allowing con-
tent to be reused. 周is is not completely
guaranteed in the case of PDF/A-1b docu-
ments, although it should usually be the
case.
■
The letter ‘O’ or the digit ‘0’? This example
shows that it is not always possible to dis-
tinguish between certain characters with
the naked eye. This is where Unicode
comes into play, since it defines a Unicode
name for each individual character. (Illus-
tration: Linotype FontExplorer X)
Missing characters: In this case, it is impossible to tell whether the
transfer should be 100 €, £, or ¥. This cannot occur with PDF/A.
Tracking information: The information on tracking has been lost in
the case of the overlapping letters.
U+0061: All of these letter ‘a’s have the same Unicode numbers, re-
gardless of the font.
PDF/A applications in everyday life
PDF/A
in a Nutshell
49
Documents you may be interested
Documents you may be interested
