Graphics Formats Explained
Bit depth: describes the unique number of colors that are available in an image's
color palette in terms of the number of 0's and 1's, or "bits," which are used to
specify each color. This does not mean that the image necessarily uses all of
these colors, but that it can instead specify colors with that level of precision. For
a grayscale image, the bit depth quantifies how many unique shades are
available. Images with higher bit depths can encode more shades or colors
since there are more combinations of 0's and 1's available. The most common
depths are 8 bit (256 colours) and 24 bit (16.8 million colours). Bit depths higher
than 24 provide a wider colour "gamut", so image manipulation software can pull
out otherwise invisible detail out of the image.
ADI - AutoCAD's Device-Independent Binary Plotter Format, a vector format
generated by AutoCAD.
AI - Adobe Illustrator's metafile format, which is actually a flavour of
AWD - Microsoft Fax At Work format, a black-and-white (one bitplane) format for
storing fax images.
BMP - This is the Microsoft Windows bitmap format, also used in OS/2. It's a
fairly compact (compression is optional, but usually turned on) format for images
up to 24 bit. BMP is the native bitmap format for the Windows environment.
CALS - Computer Aided Acquisition and Logistics Support Raster Format; a
longwinded, seldom-seen military-spec two colour document storage format.
Used in Pentagon archives, and that sort of thing.
CAM - Casio Camera, the native file format of Casio's QV-series digital cameras.
CGM - Computer Graphics Metafile, an American National Standards
Institute/International Standards Organization metafile format for images of pretty
much any kind.
CLP - This is the format you get when you save a file from the Windows 3.x
Clipboard. It is very, very large, and very very, inefficient - and, what's more, you
can only view a CLP file if you're in the same resolution as the person who made
it, and are using the same number of colours. CLP is an image format that should
never be allowed to touch a disk. Do not use it.
CT - The most popular of the Scitex image formats, Scitex Continuous Tone
images are very large and intended for use with Scitex's professional film-printing
units, which produce high-grade output for publication.
CUT - The orphaned 256 colour format used by the old Doctor Halo paint
DIB - This is an orphaned Windows image format. It stands for Device
Independent Bitmap and was part of Microsoft's Great Plan for Windows 95; the
DIB code in Win 95 is designed to simplify the creation of display drivers for new
video cards by doing most of the grunt work in the operating system instead of in
the driver. DIB never really took off.
DLG - Digital Line Graph, a vector format for storing geographical data.
EPS - Encapsulated Postscript is a flavour of Postscript (see below) which can
be included in other documents - if your software supports it.
FPX - The FlashPix format, codeveloped by Hewlett-Packard, Kodak, Microsoft
and LivePicture Corporation and now an open format administered by the Digital
Imaging Group. Kodak uses it in all of their digital cameras from the DC200
onwards. Flashpix's chief claim to fame is that it stores images in multiple
resolutions, so a huge, high resolution image can be quickly displayed in
miniature on-screen and changes made rapidly to the displayed data only, saving
the CPU-grinding full processing for whenever you actually view or output the
high resolution version. Of course, this only works if your image editing program
supports it, and is not useful for small images. FlashPix images can also be used
for Web graphics, because the server only has to send the data being viewed
(which seems to the browser like an ordinary JFIF), but since FlashPix doesn't
support progressive display like JFIF or GIF or PNG, it hasn't achieved much
popularity. FlashPix also has no zero-loss compression option - it either uses no
compression at all, and makes a vast file, or uses medium-loss JPEG-type
encoding. This makes it a clumsy format for professional use, since lossy
compression is a no-no for image editing.
GIF - Graphics Interchange Format (the acronym's officially pronounced "JIF", by
decree of the format's creator) is a very efficient, and still quite popular picture
format. There are two "flavours" of GIF, the old 87 and the newer 89a. 89a adds
several extra features like transparency (so background graphics can "show
through" the GIF in places) and animation. GIF animations are a very - some
would say excessively - popular form of Web multimedia, because they're small
and display on all current graphical browsers without needing a special plug-in or
taking up much CPU time.
Unfortunately, GIF pictures can only have 256 colours, or 256 shades of grey.
256 greys is photo quality so GIF is fine for any monochrome image, and 256
colour looks OK for many pictures, but it's no use for professional imaging.
GIF images can also be interlaced, so that you can see a low resolution version
of the picture before downloading very much of it. GIF interlacing has four
passes, which show one out of every eight lines, then another eighth of the
image, then another quarter, then the remaining half. GIF is a data-stream type
format, like JFIF, so you can view partially downloaded images whether or not
they're interlaced - without interlacing, a 25% downloaded picture gives you the
first 25% of the lines, starting at the top.
HRF - Hitachi Raster Format, an obscure, proprietary, one bitplane format used
for storing scanner data.
IFF - This is Electronic Arts' Interchange File Format, and is the image format
used by Amiga and Atari ST personal computers. There are multiple IFF formats,
but by far the most popular are the image and sound file ones. A file with the .IFF
suffix may, therefore, be a sound, not a picture - and it might be any one of a
number of other types of data. IFF images may also, uncommonly, have the
suffix .ILBM, for InterLeaved BitMap, or just .LBM on DOS-based systems.
IFF pictures are not at all efficient, spacewise, but they're fast to display, which
was important for poor little Amigas with a 0.7 million instruction per second
(MIPS) processor. With current PCs steaming along at hundreds and hundreds
of MIPS, this no longer matters at all.
IFF is peculiar in that it has two odd variants - HAM and HAM8. HAM stands for
Hold And Modify, and is a technique the original Amiga designers came up with
for getting 4096 colours from hardware which, traditionally, can only display 32 at
once. HAM8 is the updated version, which displays 262,144 colours on 256
colour hardware. No non-Amiga computer can display HAM images exactly as
they're meant to be seen, but some conversion programs can display them as
256 or higher colour images. If your display program isn't smart enough to do
this, it'll assume it's loading an ordinary 32 or 256 colour image and give you a
distinctive multicoloured porridge on screen. There are some very strange IFF
variants which use whole different palettes on every line; pray you never meet
All IFF images can be compressed or uncompressed; just about all are
compressed. The compression, like the whole format, is built for speed, not
efficiency, and so doesn't reduce the size much.
ILBM - See IFF.
IMG - This is the format used by the old GEM Paint program; it only works in 256
shades of grey.
IMG - See "PIC".
IMJ - A proprietary variant of the JFIF format created by Pegasus Image
JBG - Also suffixed JBIG, this is the Joint Bilevel Image Group's data
compression and transmission format. JBG is a way of sending one-bitplane
document images so that a low resolution version arrives first, then extra data to
"fill in" more and more detail. Not an image format as such - a JBG "file" is just a
JBG data stream dumped to disk.
JFIF - The JPEG (Joint Photographic Experts Group) File Interchange Format,
commonly called JPEG and with the filename suffix .JPG, can be the most
efficient image storage method of all - at a price. First to the name. Everybody
might call these images JPEG, but that just describes the type of compression
used for the data; it doesn't describe how the compressed data is sorted and
stored. Calling JFIF "JPEG" is like calling a Ford Falcon "internal combustion".
The idea of JPEG is that as it compresses the data it throws some of it away -
technically, this is called "lossy compression". You can configure how lossy you
want your JFIFs to be (well, you can if you're using even a slightly well written
JFIF saver); 100% quality gives you almost exactly the same result as the
original picture but also gives you a gigantic, uncompressible file. 10% quality
takes up much less space but looks dodgy. You have to strike a balance.
JFIF can store up to 24 bit colour, so it's suitable for professional use, and it can
do interlaced display like GIF (called "progressive" JFIF), which along with its
small file sizes makes it the standard format for Web graphics. Like GIF, JFIF is a
data-stream format - you can view images before you've got all of the data. Also
like GIF, JFIF supports interlacing.
The JFIF format also supports CMYK (process colour - Cyan, Magenta, Yellow
and blacK in a subtractive colour model, as against the additive Red, Green and
Blue more commonly used) images, which makes it suitable for use in publishing
applications. CMYK this support was added in a later version of the standard,
though. This means that quite a few JFIF display applications, including Web
browsers, do peculiar things when fed CMYK images. There's no reason to use
CMYK JFIFs unless you're sending the image to a CMYK output device, which a
monitor definitely isn't. Usually, CMYK ones get through because someone's
converted a CMYK image of some other format, like TIFF, without changing the
JFIF quality comparison
Reasonably high quality
9 kilobyte version
2 kilobyte version
JPG - See JFIF and SPF.
LBM - See IFF.
MacPaint (3k) - Usually suffixed .MAC, this is the format used by the ancient
original black and white Macintosh paint program. Two colours only, 576x720
resolution only, thankfully rare.
MNG - The proposed Multiple Network Graphics (pronounced "ming") format is a
multi-image extension of the existing PNG format - or it will be, if it ever makes it
out of the design stage.
MSP - Microsoft Paint was the early PC answer to MacPaint, and its format is
just as boring. Two colours only.
PCC - see PCX.
PCD - Kodak's PhotoCD was going to set the world alight, with happy snappers
having their film scanned and the high-resolution images written to CD, to access
via PC or special PhotoCD players. Amazingly, it turned out that nobody was
very interested in viewing their photos on their TV, and PhotoCD flopped
miserably in the consumer market. It survives as a somewhat popular
professional image storage format; a genuine PhotoCD has a particular directory
structure containing the images, stored in five resolutions. An ordinary PCD file
can be read by any application that can read the format, but unless it's on a CD
with the right structure, a PhotoCD player won't recognise it.
The PhotoCD storage process is proprietary to Kodak, who no longer sell the
software to make full multi-resolution images.
PCX - The ZSoft Paint format, occasionally suffixed .PCC, is ancient but still
fairly widely used, simply because everybody understands it. There are three
common versions, 0, 2 and 5; 0 is the original two colour one (small but not
useful), 2 only does 16 colours and is hence also of little interest to owners of
rather old video cards, and 5 does 24 bit. All are large for what they do, but fast
to load on elderly computers. PCX is the IBM equivalent of Amiga IFF. The size
listed is for v5, at full 24 bit; v2 scored 216k and v0 48.1k.
PIC - A few proprietary (one company makes software that supports them, and
nobody else does) image formats use this suffix. They are not interchangeable.
Some programmers need a good slapping. PIC is most likely to be the 256 colour
format of the old PC Paint program, but it might also be a Micrografx Draw!
vector file, a Lotus vector file, a Pegasus Imaging Corporation image file or an
image file for General Parametrics' Video Show Film Recorder.
PICT - Pict is the all-in-one Apple Quickdraw metaformat. It can include
bitmapped or vector images, and can use different compression schemes.
PNG - The Portable Network Graphics format, pronounced "ping", was created
as a free replacement for GIF, whose LZW compression is owned by Unisys and
which can't be included in commercial software without paying license fees to the
owners. It handles 1 to 48 bit images, and is a lossless, well-compressed format
like GIF. It still isn't very popular, though.
PS - Adobe Systems' Postscript isn't an image format, per se - it's a page
description language, originally conceived so computers could send very
accurate page descriptions to the then-new high resolution laser printers. You
can save black and white or even colour pictures as Postscript, but you'll end up
with a very large file. Postscript is not a very efficient format, but its advantage is
it's all plain text - you can modify a Postscript file with any text editor, if you know
what you're doing.
PSD - Adobe Photoshop's native format, which stores all of its layer and
selection and miscellaneous other image data.
RAS - This is SUN Raster format, the default image format for monster SUN
workstations. Only lighhtly compressed and so a rather large format, but it
supports up to 36 bit images.
RAW - This may be a Photoshop RAW file, which is a PSD file with no identifying
header. Or it may be a minimally formatted image data dump - see PIC.
RGB - See "PIC".
RIX - The orphaned bitmap format of the old DOS ColoRIX paint program.
RLE - This is an antique CompuServe or Windows Run Length Encoded
compressed image format, which only support 256 x 192 black and white
RTF - Microsoft's Rich Text Format, which is normally used as a well-understood
cross-platform word processing document format, but which can store pictures as
well as text. As image storage formats go, though, this one's as bloated as
SPF - SPIFF, Still Picture Interchange File Format, the "official" International
Standards Organisation Joint Photographic Experts Group (ISO JPEG) image
format defined in the recent Part 3 extensions to the JPEG standard. SPIFF
offers more features than the current JPEG standard and is backwards
compatible (a JFIF decoder can understand most SPIFF images), but has not yet
achieved much popularity. SPIFF files may also be suffixed .JPG.
TGA - The real name for this format is just plain "TGA" or "Truevision File
Format", but a lot of people call it "Targa", after the Truevision video card that
first used it. There's a lot of this name confusion in image file formats. It supports
1 to 32 bit images and professional features like an alpha (mask) channel,
gamma settings and a built-in thumbnail image.
TIF - TIFF (to give the full acronym) stands for Tag Image File Format; many
people say Tagged for the first word, which is technically incorrect but minimally
important. TIFF was a large, unwieldy, 24 bit format until version 6 came out,
which supported compression and made it less painful. Mind you, the fact that its
compression was somewhat broken and might or might not be compatible with
different programs on different computers somewhat reduced the bonus, and the
further fact that the compression is LZW and thus owned and licensed out by
Unisys (see GIF) is another pain. TIFF is, nonetheless, a very popular
professional graphics format.
WMF - This is Windows Metafile format, which is an intermediate vector format
for Windows programs to use when interchanging data and, generally speaking,
should never be seen anywhere else.
WPG - This is the WordPerfect metafile format, used by WordPerfect software on
various platforms. It supports bitmapped, vector and Encapsulated Postscript