Search the FAQ Archives

3 - A - B - C - D - E - F - G - H - I - J - K - L - M
N - O - P - Q - R - S - T - U - V - W - X - Y - Z - Internet FAQ Archives FAQ, Part 4/4

( Part1 - Part2 - Part3 - Part4 )
[ Usenet FAQs | Web FAQs | Documents | RFC Index | Business Photos and Profiles ]
Archive-name: pc-hardware-faq/video/part4
Posting-Frequency: monthly (second Monday)
Last-modified: 1997/02/19
Version: 1.0

See reader questions & answers on this topic! - Help others by sharing your knowledge
COMP.SYS.IBM.PC.HARDWARE.VIDEO Frequently Asked Questions - Part 4/4

Q) What is the pinout for a standard VGA/PGA/EGA/CGA connector?

Standard 15 pin D-Sub VGA connector pinout
\                                                 /
 \        1       2       3       4       5      /
  \                                             /
   \  6       7       8       9       10       /
    \                                         /
     \   11      12      13      14      15  /

Pin #   Description

1       Red Video
2       Green Video
3       Blue Video
4       Sense 2  (Monitor ID bit 2)
5       Self Test (TTL Ground)
6       Red Ground
7       Green Ground
8       Blue Ground
9       Key - reserved, no pin
10      Logic Ground (Sync Ground)
11      Sense 0 (Monitor ID bit 0)
12      Sense 1 (Monitor ID bit 1)
13      Horizontal Sync
14      Vertical Sync
15      Sense 3 - often not used

Compaq (and perhaps some other companies) use the "Sense" lines as a
way of telling what kind of monitor is connected.  Newer monitors with
DDC (also called Plug-n-play) use some of these pins.

[From: Ashok Cates (]

The  ID bit pins in the 15
pin connector are shorted/left open to identify the type of monitor.  I
don't think they are very important anymore, as most cards have software
to set resolutions, refresh rates etc.  However, I think their functions

ID bit 0 and ID bit 2 grounded:
Dual frequency analog color interlaced (8514 or compatible) or variable
frequency analog color interlaced.

ID bit 0 grounded, ID bit 2 not connected:
Fixed frequency analog color (8512, 8513, or compatible) or variable
frequency analog color non-interlaced.

ID bit 0 not connected, ID bit 2 grounded:
Fixed frequency analog monochrome (8503 or compatible) or variable
frequency analog monochrome.

-ID bit 1 and ID bit 2 are usually connected together.
-Monitor model numbers are for IBM monitors.

Standard 9 pin D-Sub PGA/EGA/CGA connector pinout
[From: Michael Scott (]

\                     /
 \ 1   2   3   4   5 /
  \                 /
   \ 6   7   8   9 /

                IBM Adapters

Pin Assignment  CGA             EGA             PGA             VGA
                TTL 16 colours  TTL 16/64 col.  Analogue        Analogue

1               GND             GND             Red             GND
2               GND             Secondary Red   Green           GND
3               Red             Primary Red     Blue            Red
4               Green           Primary Green   Composite Sync  Green
5               Blue            Primary Blue    Mode Control    Blue
6               Intensity       Secondary Green Red GND         GND
7               not used        Secondary Blue  Green GND       not used
8               H. Sync         H. Sync         Blue GND        H. Sync
                                                                /Comp. Sync
9               V. Sync         V. Sync         GND             V. Sync

Q) What are VGA/SVGA/UVGA/8514/a/XGA?

The wonderful thing about PC's is that there are standards for so many
different things.  The problem is that every company has their own
standards ;-).  The lack of a widely accepted standard for >VGA pixel
addressabilities is causing plenty of problems for manufacturers, system
builders, programmers and end users.  As a result, each vendor must
provide specific drivers for each supported operating system for each
of their cards.  In the list above, VGA, 8514/a and XGA are standards
established by IBM, and have been accepted to a greater (VGA), lesser
(XGA) or even much less (8514/a) degree.  The reason for this may be a
backlash against IBM (due to royalty demands) or that video card vendors
were not satisfied with the suggested standards.

For a more detailed discussion of VGA, see 'What is VGA, and how does it

The 8514/a was the next graphics offering from IBM and provides three new
video modes that are not available from the VGA controller.  Computers
with 8514/a hardware must also have a VGA controller, as the 8514/a does
not support VGA video modes.  The additional modes are:

Type    Pixel           Max. # Colours  Characters
gfx     640x480         256             80x34
gfx     1024x768        256             85x38  (interlaced)
gfx     1024x768        256             146x51 (interlaced)

The 8514/a also has some smarts, as it is capable of performing video
memory transfers, drawing lines and extracting rectangular areas of
the display image.  These are so-called accelerated features.

The XGA has superseded the 8514/a.  It was the first IBM display adapter
to use VRAM, and can be configured with 500k or 1 Meg.  Like the 8514/a,
the XGA has accelerated features which make it faster than standard VGA
for some operations.  The new modes XGA introduced are:

Mode    Type    Pixel           Max. # Colours  Characters
14      text    1056x400        16              132x25
-       gfx     640x480         256/65535*      -
-       gfx     1024x768        16/256*         -

*500k/1 Meg configurations


SVGA and UVGA are not established standards, and so their meanings vary
depending on manufacturer.  VESA VGA BIOS Extensions are the closest
thing to an 'SVGA' standard.  Most video cards currently available are
called SVGA (Super VGA), which basically means that the card provides a
superset of standard VGA calls and capabilities.  This means that
anything better than 640x400 and 16 colours is an SVGA mode.  Some
suggest that SVGA covers 800x600 modes, while UVGA (Ultimate VGA) refers
to 1024x768.  However, the absence of any real standard renders the term
SVGA quite useless, and the term UVGA is not used frequently.

The result of having no SVGA standard is that there are many (>10 !)
different SVGA chipsets available, and none of them use a common
programming interface.  Many provide video acceleration capabilities,
which free the system CPU to do other tasks,  i.e. hardware cursor,
BitBlt, etc.  However, to use the SVGA video modes and advanced
features, each chipset requires its own driver.  This is why video
drivers are required for Windows 3.1, Windows 95, OS/2 & XFree86.  These
drivers, combined with accelerated hardware, can provide enormous
increases in video performance.

If you are looking for a machine and would like SVGA capabilities,
don't accept that a given video card or monitor is adequate just
because it is advertised as supporting SVGA.  Instead, decide what
maximum pixel addressabilities and colour depths you want to use, and at
what vertical refresh rates, and ensure that the models you are looking
at provide those capabilities, and that software drivers are available
for the operating systems and programs you will be using.

Q) What is VESA SVGA?

While some vendors of video hardware decided to provide support for 8514/a
or XGA standards set by IBM, most defined their own 'SVGA' modes.  As a
result, no common programming interface was available which would allow
generic SVGA code to be written.  In order for programmers to be able to
write code which would work on a wide range of 'SVGA' hardware, VESA
(Video Electronics Standard Association) defined a standard interface
for SVGA functions.  It's more correct title is 'VESA VGA BIOS Extensions'
and it incorporates functions which allow a program to determine what
video modes (pixel addressabilities and number of colours) and other
functions are available and how the video memory is accessed.

Because many vendors already had proprietary extensions to the VGA
standard implemented in their hardware, VESA VGA extensions use a software
interrupt to access all of the programming routines.  This means that a
video card vendor can provide a VESA video driver (also called a TSR -
Terminate and Stay Resident program) which can fill the role of inter-
preter between VESA VGA compliant software and proprietary SVGA hardware.
As a result, programmers can now write software that will work on a range
of SVGA hardware, taking advantage of more colours and higher pixel
addressabilities than are available with VGA.  The video modes defined by
VESA are:

Mode #  Pixel           Colours
100h    640x400         256
101h    640x480         256
102h    800x600         16
103h    800x600         256
104h    1024x768        16
105h    1024x768        256
106h    1280x1024       16
107h    1280x1024       256

Nuts & Bolts

Specifically, the VESA VGA extension provides information and hardware
setup to the application program.  It has six functions:
Function 0 - Return Super VGA Information
Function 1 - Return Super VGA mode information
Function 2 - Set Super VGA video mode
Function 3 - Return current video mode
Function 4 - Save/Restore Super VGA video state
Function 5 - CPU Video Memory Window Control

These functions are all accessed by placing 4Fh in the AH CPU register,
the desired function in the AL register, then generating an interrupt

While this VESA standard doesn't define how 'accelerated' functions
like hardware mouse cursors, BITBLT or typical GUI windowing operations
should be accessed, it does provide a common set of instructions for
determining information about and programming of higher pixel
addressabilities and colour depths for video cards that have a superset
of standard VGA functions.

For more information, contact VESA at:

Video Electronics Standard Association
2150 North First Street
San Jose, CA 95131-2029
(408) 435-0333
(408) 435-8225

Q) What should I consider in buying a video capture card?

There are several factors that will determine which video card is the
best for your purposes.  It will depend on the number and type of video
inputs, AD (Analog to Digital) conversion and system noise, frame rate,
video overlays and whether video capture is to be integrated with other

Grayscale and colour video capture cards are available.  Grayscale
cards are usually 8 bit, but some are available for 12 bit conversion.
This means that the video intensity is sampled temporally, measured as
a voltage, then divided into 2^8 (2^12) or 256 (4096) discrete levels.
8-bit provides enough gray levels for most applications and approaches
the noise threshold in most video systems.  Noise can be reduced in
this or any colour system by frame averaging.

Colour capture cards are available in 16, 24, 32 and more bit models.
They convert the individual red, green and blue video streams into
digital values separately, each stream being treated similarly to
grayscale digitization.  16-bit cards discretize RGB into 5, 5 and 6
bits, and so can record 65535 different colours.  24-bit cards provide
8 bits for each pixel for a total of up to 16.7 million colours.  24-
bit cards are also called Truecolour because most humans can distinguish
5-6 million colours.  At 16.7 million, 24-bit colour can display more
different colours than anyone can perceive.  Cards that provide 32 bits
or more of colour depth are usually Truecolour cards with overlay
capabilities.  The overlay planes (8 bits in the case of 32 bit) can
be used to contain text or graphics overlays, or can store depth
information (z-buffer).  In addition, extra video memory can be used
to double buffer the incoming digitized signal, up to doubling the
frame capture rate.

Video capture cards can often digitize different image sizes , though
the most common is 640x480.  640x480 is the maximum image size that is
meaningful for NTSC video signals.  Keep in mind that while the horiz-
ontal resolution of a television signal is quite high, the vertical
resolution is limited to the number of scan lines displayed.  A VCR
provides ~250 lines, while S-video or laser disc provide over 300.
Many video cameras provide more - closer to the 525 that the NTSC
standard can handle.  This means that the capture card has to integrate
vertically (or subsample) to get 480 pixels vertically.  This
introduces a smoothing effect in the vertical direction and results
in a less sharp picture.  Capture cards are available which will
digitize larger images, but they require special-purpose video
equipment to be used to any advantage.  Ensure that the resolutions you
use maintain the screen aspect ratio.

Various types of input signals can be digitized including NTSC, PAL,
S-video and RGB.  Some cards can handle all types, but most of the
less expensive ones can only understand NTSC.  Boards that can
capture separate RGB signals can often be used to connect up to 3
grayscale video inputs.

Many video cards come with simple frame capture programs, but
if you are planning to integrate video capture with other operations
on the computer, like collecting data from an AD card, adding text
data as an overlay or changing video-in channels on-the-fly, you
will have to do some programming.  In this case you will need good
programming libraries in a language you are familiar with for the
video card.  Some companies include libraries with their cards, but
most charge extra.  Most often libraries, when available, are for
C or BASIC, and sometimes Pascal.

Q) What type of camera do I need for video capture?

The type and quality of camera you require depends on the application.
In general, most home hobbyists will opt for an inexpensive one-chip
CCD colour camera, while high-end video applications require a three-
chip colour CCD or tube camera.  Most scientific work requires the
high definition grayscale of a monochrome CCD or tube camera.

CCD vs. Tubes

Charge Coupled Device (CCD) cameras are a solid-state, inexpensive
and durable alternative.  The same technology as is incorporated
into camcorders is used in stand-alone CCD video cameras.  CCD's
consist of an array of light-sensitive material, which produces
an electrical signal when struck with a light photon.  As light
photons continually stream through the lens and strike the CCD,
they produce different voltages in corresponding CCD elements.  By
sampling the voltage generated at each element, an analog raster
representation of light intensity is collected.  This produces a
grayscale representation of the sampled light image, where the
maximum voltage corresponds to white, and the minimum corresponds
to black.  CCD's suffer from black noise (noise generated from an
element even when no light photons are striking it) and relatively
low light sensitivity, though newer CCD's are improving.  CCD's
have the advantage of low cost and high durability.

Tube cameras use older tube technology instead of solid-state
silicon.  They are very light sensitive, and so are useful for low-
light applications.  In general, most tube cameras are used when
CCD technology in inadequate.  They are more expensive than CCD's
and are more easily damaged by excessive light exposure.

Colour CCD

Two varieties of colour CCD's are available; one and three chip
implementations.  A three chip CCD uses three discrete CCD arrays,
each with a colour filter in front of it:  red, green or blue.
Each CCD is sampled in a raster fashion, the same as for the
grayscale device, above, and the result is a colour analog signal.
Because they require three discrete CCD's, the three chip models
are more expensive than one chip models and provide better colour
reproduction.  The latter use one CCD, and no colour filters.
They consider the energy of the incoming photons, which determines
their colour, to produce a red, green and blue value for each CCD
element in the array.  While cheaper, colour reproduction of one
chip CCD's is inferior to three chip.

Digital Cameras

Although expensive and used less frequently than analog video
cameras, digital cameras have the advantage of not requiring
dedicated frame capture hardware in the computer.  They are based
on the same CCD technology as is described above.  An example of
a digital camera is the IndyCam which come with SGI Indy
workstations.  Also, hand-held portable digital cameras are
available which can download images to your computer.

Q) I want to add an MPEG card to my system.  How does it work?

The Motion Pictures Experts Group (MPEG) has released a series of
standards which describe a lossy digital video compression technique.
In some cases, MPEG can reach compression rates of 100:1.  It works
by removing redundant information and details that most people would
generally miss, and in later versions storing only the differences
between successive frames.

When an MPEG video clip is viewed on the screen, the video stream
must be decoded on-the-fly.  If done in software, this operation can
be quite demanding of the system CPU.  An alternative is to have a
dedicated coprocessor do the MPEG decoding, then feed the resulting
video stream to the video card.  Because this type of coprocessor
is dedicated to MPEG decoding, it can be optimized to perform the
operation very fast, and can also be used to scale up the size of
the resulting video with little or no degradation in performance.
Even a relatively small 320x200 video displayed at 30 frames per
second requires a bandwidth of 15.4 million bits.  This would
seriously decrease available bandwidth for other purposes like
disk i/o if all of that data was dumped down the peripheral bus
(ISA, VLB, PCI, etc).  As a result, many video card manufacturers
incorporated a feature connector on their VGA cards.  This connector
gives direct access to video display memory, allowing high frame
rate video to be dumped to the monitor.  One limitation of this
adapter is that it can only provide 8-bit (256 unique) colour.

If you're planning on using your PC as a VCR, you'll be disappointed
with an MPEG card playing the cdrom version of your favourite film.
The resolution will be inferior to that provided by your television.
If you want to get smoother video playback and/or free-up your CPU
for other tasks, then the addition of an MPEG decoder card may be
worth the cost.

Q) What is the feature connector on my video card for?

A  The feature connector comes in two variants; VGA and VESA.
The basic idea is that video memory can be directly accessed by
using the feature connector, bypassing the CPU and peripheral
bus.  This reduces CPU load and avoids bandwidth bottlenecks.
In addition, this eliminates the need for a separate RAMDAC.
Typically, the feature connector is used by video capture cards
or MPEG decoder cards, as it provides high bandwidth which is
ideal for playing video clips.  The older VGA feature connector
is limited by the VGA itself, and can only display up to 256
colours at a pixel addressability up to 320x200.  While this can
provide reasonable images, they tend to look dithered, and the
quality is less than that of NTSC television.

A more recent standard is the VESA Media Channel (VMC) which
allows you to attach an MPEG or TV tuner card to your video card
thorugh a high-speed connector.  The VMC actually implements a
full bus system which allows up to 15 devices to share the
frame buffer and RAMDAC on the video card.  More information on
the VMC is available at the VESA WWW site at, though
the standards themselves are only accessible to VESA members.

Q) What is DCI?
[From: Dylan Rhodes (Formerly of Hercules)]

DCI stands for "Device Control Interface."  It's an Intel/Microsoft
standard, and exists primarily as a way for Windows 3.1 to exploit the
video acceleration features of a graphics card, and/or to provide fast
video when needed -- for example, the WinG games library uses DCI.
A DCI driver exists at the same software layer as the GDI.

Among DCI's capabilities are the ability to write directly to the
frame buffer (helpful for high-speed games) and the ability to
provide for on-board hardware acceleration of video scaling (i.e.
stretching a video window to a larger size) and color space
conversion (converting the YUV format color information in a video
file to the RGB format that a typical graphics card RAMDAC expects).
Note that support for DCI features doesn't need to be in hardware --
a graphics card vendor could provide a DCI driver that allowed
Windows 3.1 apps to speak DCI, but the graphics card could be
performing the DCI functions with a software driver.

Note: with Windows 95, DCI will be replaced by an expanded interface
called DirectDraw.

Q) How do I contact my video card/monitor vendor?

A large list of vendors' phone numbers is distributed in the c.s.i.p.h.
FAQ section 9.8.  This FAQ is posted monthly to c.s.i.p.h.* groups, and
is available via FTP from: in /pub/usenet/
and its mirrors.

Many vendors are on-line.  See the following section "Is there an Internet
ftp/web site for my video card?" and/or these sites for listings:  : Guide to Computer Vendors by SBA Consulting.  : Computer Companies
                Phone List Pt 1 - HARRIS Mountaintop  : List of 60+
                monitor companies with phone numbers and WWW sites.

Q) I need new drivers.  Is there an Internet ftp/web site for my
        video card?

It's pointless for me to try to maintain a list of on-line services, since
they change so frequently, and others are already doing it!  Please refer
to: FAQ in sections 9.6 and 9.7.  This FAQ is posted
monthly to c.s.i.p.h.* groups, and is available via FTP from: in /pub/usenet/
and its mirrors.   : List of Windows95 drivers and updates

Here's a list of ftp sites for video card vendors on the 'net for those
too lazy to look up the other sources.  For a larger list of video
related WWW sites including vendors and information, refer to:

Most major vendors' www sites are listed there.

FTP Sites:

Alliance Semiconductor Corp.
ATI Technologies Inc. 
Boca Research Inc.    
Cirrus Logic          
Diamond Multimedia Systems, Inc
Genoa Systems Corp.   
Media Vision          
MIRAGE Video Solutions
miro Computer Products
Number Nine Visual Technology
S3, Inc.              
Software Integrators  
STB Systems, Inc.     

If you have any more sites to add, please email the FAQ maintainer.

In addition, video drivers are archived at:
        windows:        /.22/cica/drivers/video
        linux:          /.6/linux
        OS/2:           /.4/os2/drivers & maybe /.4/os2/warp

Other video utilities are available at SimTel, Cica and Garbo mirrors.


Abrash, Michael.  Demystifying 16-bit VGA, in Dr. Dobb's Journal,
May 1990

Abrash, Michael.  Mode X:  256-color VGA magic, in Dr. Dobb's
Journal, July 1991

Ericsson, Bo.  VESA VGA BIOS extensions, in Dr. Dobb's Journal,
April 1990

Howard, Christopher A.  Super VGA programming, in Dr. Dobb's
Journal, July 1990

McNierney, Ed.  New issues in PC graphics, Dr. Dobb's Journal,
November 1986

Myers, Ben.  Saving and restoring VGA screens, in Dr. Dobb's
Journal, July 1991

Norton, P.  Inside the IBM PC and PS/2, 4th Edition.  Brady,
New York, New York, c1991

Peddie, Jon.  High-resolution graphics display systems. Windcrest
(McGraw-Hill), U.S.A., c1994

Sanchez, J.  Graphics design & animation on the IBM microcomputer,
Prentice Hall, Englewood Cliffs, N.J., c1990

Sutty, G. & Blair, S. Advanced programmer's guide to the EGA/VGA,
Brady, New York, New York, c1988


Contributing Authors:
Ron Bean (
Sam Goldwasser (
Declan Hughes (
Bill Nott (
Dylan Rhodes (Formerly of Hercules)
Michael Scott (
Roger Squires (
Ralph Valentino (

Sam Goldwasser (
Andy Laberge (
Bill Nott (
Dylan Rhodes (Formerly of Hercules)
Ralph Valentino (

Appendix A - Glossary


If you don't find the definition you are looking for in this glossary,
try the resources below:

The "Free On-line Dictionary of Computing" is available via the web at:
This dictionary is compiled and maintained by Denis Howe (

A large list of COMPUTER ACRONYMS is defined in the Babel document
accessible via ftp or the web.  It is updated 3 times per year, so
you have to request the latest document.  It's of the format
babelYRP.html where

YR is the year, i.e. 95
P is the update period i.e. a, b or c:
         After May 1, 1995 request BABEL95B.
         After Sep 1, 1995 request BABEL95C.
         After Jan 1, 1996 request BABEL96A.  filename as above babelYRP.txt
                                          i.e. babel95b.txt
Babel is compiled and maintained by Irving Kind (


8514/a  IBM video graphics standard.  Supports pixel addressabilities
        up to 1024x768 and 256 colours.  It is _not_ a superset of VGA.
addressability (pixel addressability)
        This refers to the number of pixels that a video controller
        can display.  It is quoted as the (# horizontal pixels)
        by the (# vertical pixels).  Common PC pixel addressabilities
        320x200, 640x480, 800x600, 1024x768, 1280x1024 & 1600x1200
aperture grille
        An array of vertical wires which act in a similar manner as a
        shadow mask.  Their basic purpose is to permit the correct
        electron beam to strike its corresponding colour phosphor only.
        This results in crisp pixel definition, and superior colour
        brightness than is realized with more traditional designs.
        The aperture grille was first used by Sony in their Trinitron
AT bus  Advanced Technology (IBM) bus.  The standard PC compatible
        peripheral bus to which add-in cards like video, i/o, internal
        modems, sound are added.  Also called the ISA bus, it runs at
        a maximum of 8.33 MHz and has a 16-bit wide data path.
        Also called video bandwidth.  This is a measure of how much
        gross throughput a monitor can handle (in MHz).  Bandwidth at
        a given pixel addressability is a function of the vertical
        refresh rate and monitor timing.  see 'How do I calculate the
        minimum bandwidth required for a monitor?"
BIOS    Basic Input Output System.  The video BIOS basically tells
        the computer how to talk to the video subsystem at boot time.
        The video BIOS calls are used by DOS for VGA (and SVGA) modes.
BITBLT  A VGA video operation which copies an array of values to a
        rectangular region in video RAM.
bit planes
        This is the number of bits which are available to store colour
        information for each pixel displayed.  The number of colours
        which can be displayed is calculated as two to the exponent
        'n', where n is the number of bit planes.  i.e.  4 bit equals
        16 colours, 8 bit equals 256 colours and 24 bit equals 16.7
        million colours.  see "How does colour depth (bit planes)
        relate to the number of colours?"
colour depth
        Refers to the amount of memory (and therefore number of
        simultaneously displayable colours) available to store colour
        information for each pixel.  see 'bit planes'.
CPU     Central Processing Unit.  This is the heart and brains of your
        computer.  It is responsible for executing code, moving data,
        calculations, etc.  For PC's, this chip is a member of the X86
        family including 8088 through 80486, Pentium and Nextgen.
CRT     Cathode Ray Tube.  Basically the same technology as is in modern
        television sets.  One or more beams of electrons are focused onto
        phosphor, causing it to glow.  The phosphor is arranged into an
        array (usually close to rectilinear), and the electron beam scans
        the phosphor on the screen (similar to how you read text - left
        to right and top to bottom), usually 60+ times per second.
degauss Magnetic interference caused by a change in the position of a
        monitor in relation to the earth's magnetic field or the
        presence of an artificial magnetic field can cause discolour-
        ation.  To correct this, all colour monitors automatically degauss
        at power-on and some also have a manual degaussing button.
        This allows the monitor to compensate for the change in the
        magnetic field by realigning the electron guns.  In some low
        cost monitors without degauss buttons it is necessary to leave
        the power turned off for at least 20 minutes in order to get
        maximum degaussing.
display Usually used to indicate the monitor or flat-panel device used as
        the primary visual interface.
display adapter
        Usually this is the same as the video card, but some mother-
        boards have built-in video, and so don't require an additional
        card.  The display adapter contains video memory which stores
        what is displayed on the computer's monitor.  They have a
        wide range of features, from a basic frame buffer, to advanced
        3D geometric rendering engines.
dot clock
        Technically, this refers to the digital clock signal that
        transfers data into the video card's digital to analog converter.
        However, it has also become a measure of the maximum gross data
        throughput of a monitor.  It is measured in MHz, and indirectly
        determines the maximum pixel addressability and vertical refresh
        rate that a monitor can handle.  See "What do those monitor
        specifications mean?"
dot pitch
        The distance between a phosphor dot of one phosphor triad to
        its closest diagonal neighbour of the same colour on a monitor.
        Expressed in mm - i.e. .28 dot pitch means .28 mm between
        triads.  A smaller value indicates that the phosphor dots
        are more closely spaced, and that the resulting image displayed
        will be crisper.
dot stripe
        see 'stripe pitch'
DRAM    Dynamic Random Access Memory.  The vast majority of system RAM
        in modern computers is of this type because of it's low cost.
        It is also the most common type of RAM used for video cards.
        A specialized type of DRAM called VRAM is also used in higher
        end video cards.  see "What is the difference between VRAM
        and DRAM?"
EGA     Enhanced Graphics Adapter (IBM).  Precursor to VGA, all EGA
        video modes are supported in VGA, though register compatibility
        is not 100%.  EGA cards generate a digital signal, and thus
        will not drive a modern, analog monitor.
EISA    Extended Industry Standard Architecture.  This 32-bit bus
        standard was created primarily to compete with IBM's MCA bus.
        It runs at speeds of up to 8.33 MHz.  EISA is a dying standard.
graphics controller
        This is a generic term to describe the video hardware in a
        computer.  Sometimes it is built onto the motherboard, but
        usually it is a separate daughter card that fits into one
        of the expansion bus slots.  The interface between the graphics
        controller and the main processor is one of the ISA, EISA, MCA,
        VLB or PCI buses.  The graphics controller is responsible for
        generating the video signal that is sent to the monitor.
        Typically a graphics controller contains a graphics coprocessor
        which may be a graphics accelerator, video RAM and a RAMDAC.
graphics coprocessor
        A secondary processor dedicated to performing video display tasks.
graphics accelerator
        This is a highly misused and now almost meaningless term.  For the
        purposes of this FAQ, a graphics accelerator is a coprocessor which
        is capable of specific graphics operation, independent of the main
        system CPU.  See the section "How does a video accelerator
        work, and will one help me?"
GUI     Graphical User Interface.  In contrast to text-based interfaces like
        DOS or UNIX, GUI's provide more flexibility in terms of colour,
        pixel addressability and types of objects that can be displayed.
        Examples of GUI's include X-Windows, Microsoft Windows 3.1, OS/2.
        A monochrome display adapter which is MDA compatible and
        provides graphics modes up to 720x348
horizontal refresh
        see horizontal scan rate
horizontal scan rate (horizontal frequency)
        The frequency, expressed in kHz (thousands of times per second),
        at which the horizontal deflection circuit operates.  This roughly
        translates to the number of scanlines displayed on a monitor in
        one second.
        Standard NTSC television signals are interlaced, meaning that
        each video frame is divided into two separate fields of
        alternating scanlines.  The resulting fields are displayed
        sequentially, such that what was originally a 30 frame per
        second (fps) refresh becomes 60 Hz at half the vertical pixel
        addressability.  Thin horizontal lines will appear to flicker
        on an interlaced display since their effective refresh rate
        is only 30 Hz.
ISA     Industry Standard Architecture.  This is a 16-bit bus standard
        which runs at speeds of up to 8.33 MHz.  The vast majority of
        peripheral add-in cards like modems, sound cards, cdrom
        interfaces and other low-bandwidth applications are still ISA
        based.  VLB and PCI provide higher bandwidth for video and
        disk I/O operations.
Look-up Table (LUT)
        At higher pixel addressabilites, most graphics controllers can
        not simultaneously display as many colours as they are capable
        of generating.  Because of video card memory limitations, only
        a subset of all possible colours can be displayed at one time.
        A look-up table stores the mapping information which determines
        which subset of all possible colours are available at any given
MDA     Monochrome Display Adapter (IBM)
monitor Usually a CRT-based device which directs an electron beam onto
        coloured phosphor, causing it to glow.  Monitors use the same
        basic technology as televisions, but are capable of much higher
        pixel addressabilities and resolutions.
        The main component of the computer, which contains the CPU
        (brain), main memory slots, keyboard connector and expansion bus
        slots, among other possible components.
        This means that an entire frame is displayed with each screen
        refresh.  Non-interlaced displays produce a more pleasing screen
        image since thin horizontal lines don't flicker with each screen
OEM     Original Equipment Manufacturer.  Often manufacturers will produce
        versions of their products in large quantities for other companies
        who either stick their name on them or use them as components for
        their systems.  OEM products often make it to the retail sales
        arena where they are sold at lower prices.  An OEM version of a
        card _may not_ be equivalent to the retail version.
PCI     Peripheral Components Interconnect.  This is basically the Pentium
        equivalent to the VLB, but with improvements.  It is a 64-bit
        standard, but is currently only implemented as 32 bits - look
        for 64 bit PCI in the future.  It performs asynchronously to
        the main CPU, meaning that the PCI bus operates at 33 MHz
        regardless of the CPU clock.  It also allows more than two
        devices on the bus, unlike VLB.
phosphor triad (dot triad)
        This is the smallest dot that can theoretically be resolved on a
        colour monitor and consists of three phosphor dots - one each of
        red, green and blue.  When struck with the electron beam, these
        dots glow producing a bright spot on the screen.  Practically,
        1.2 or more dot triads comprise each pixel on the screen,
        although the pixel addressability of some monitors is greater than
        their resolution, and in this case a pixel can be smaller than a
        dot triad.  The result in this case is that small objects may not
        be resolvable.
pixel   This is the smallest addressable display unit available at a
        given video addressability.  There is no physical thing on a
        display that can be called a pixel.  Pixels exist only in the
        graphics controller bitmap.  The screen image in the bitmap is
        composed of an array of pixels, arranged in a rectilinear
        fashion, with the X axis running horizontally, perpendicular to
        the Y axis.  A pixel consists of intensity only (in grayscale
        monitors) or colour and intensity information (red, green & blue
        in colour).  While a pixel usually corresponds to a square or
        rectangular area, it is displayed as a number of spots on a
        CRT.  One pixel usually consists of 1.2 or more dot triads.
        Flat panel displays are a special case where individual pixels
        correspond directly to a picture element on the display.
pixel addressability
        see addressability
RAM     Random Access Memory.  RAM comes in different types, including
        DRAM (Dynamic RAM) and VRAM (Video RAM) among others.  DRAM is
        used as main system memory, while both DRAM and VRAM can be
        used on graphics cards.
RAMDAC  Random Access Memory Digital-to-Analog Converter.  This is part
        of the graphics card which converts the digital intensity values
        for each of the red, green and blue guns (usually an 8-bit
        number) to analog voltages which are sent to the monitor.  A
        RAMDAC can use its RAM to store look-up table (LUT) information.
refresh rate
        When referring to monitors, the number of times that the video
        card refreshes the entire screen in one second.  Expressed in Hz
        The most common misinterpretation of this term is that it is the
        same as pixel addressability.  In fact, resolution is more
        closely related to dot pitch, since it is a limitation of the
        monitor rather than of the graphics controller.  The resolution
        limits how small an object a monitor is able to display.
RGB     Red, Green and Blue.  By varying the intensity of each of these
        colours in a single pixel, the human eye can be fooled into
        seeing a wide range of colours.  For example, a combination of
        red and green appears as yellow, even though no light with a
        yellow wavelength is emanating from the screen.  This works
        because the optical system integrates the photons striking a
        region on the retina, and the combined impulses from green
        and red sensitive cones are seen as yellow.
        The movement of a monitor's electron gun from one side of the
        screen to the other results in the appearance of a horizontal
        line of varying intensity and colour.  Typically, 200 to 1200
        horizontal scan lines (lined-up vertically on top of each other)
        make-up the image you see on your display.
shadow mask
        This is usually an invar mask which acts to block the electron
        beam from striking the wrong phosphors in a CRT.  The beam
        passes through holes in the mask to strike the correct phosphor
        while shadowing neighbouring phosphor.  i.e. it prevents
        a beam intended to strike a red phosphor from striking a
        neighbouring green phosphor by causing an electron shadow
        over the green dot.
stripe pitch
        This is similar to dot pitch, but applicable to Sony Trinitron
        and similar tubes which use fine vertical wires (aperture
        grille) to separate phosphors.  Dot stripe is measured as the
        distance between the vertical stripes that result.  Measures
        of dot pitch and dot stripe are _not_ directly comparable.
        A common but proprietary picture tube design developed by Sony.
        Uses fine vertical wires instead of the more traditional
        shadow mask.  see "Why does my monitor have 1/2/3 faint
        horizontal lines on it?"
vertical refresh rate (vertical scan rate)
        The number of fields (on an interlaced display) or frames (on
        a non-interlaced display) that are displayed in one second.  A
        field or frame covers the entire screen area.  This is measured
        in Hz (cycles per second).  It is limited by the monitor and
        video card (pixel addressabilities and colour depths).  Modern
        monitors and video cards provide refresh rates of 60Hz+.
VESA    Video Electronics Standards Association.  This group has produced
        standards for the VLB (Vesa Local Bus), VESA SVGA video modes and
        standards for minimum screen refresh rates at various pixel
VGA     Video Graphics Array (IBM).  Supports pixel addressabilities of
        up to 640x480x16.  This is the de facto video standard and
        consists of a number of video modes.  It is still heavily
        supported by DOS-based applications and games.  see "What is VGA,
        and how does it work?"
video card
        A dedicated piece of hardware which performs graphics
        operations.  Also called a display adapter.  Consists of
        microchips and other electronic components mounted on a
        pc-board which connects into a slot (ISA, EISA, MCA, VLB or PCI)
        on the motherboard.
viewable area
        Typically monitors are advertised by the diagonal size of the
        picture tube in inches.  Common sizes are 14", 15", 17", 20"+.
        However, the amount of the screen that can be seen is usually
        less.  For example, most 17" monitors have only a 15.5" diagonal
        area used for display, in part because the actual phosphor area
        is only about 16" due to the glass thickness.  This is partially
        due to the fact that the monitor's case covers the edge of the
        tube, and partially because monitor manufacturers want to make
        you think you're getting a larger display than you are.
        see "Size" under "What do those monitor specifications mean?"
VLB     VESA Local Bus.  This 32 bit bus was originally designed to
        provide higher bandwidth for video cards than is available with
        the ISA bus.  It is optimized for the 486 CPU and can run at
        speeds up to 40 MHz with one card on the bus, or up to 33 MHz
        with two cards on the bus.  The speed of the VLB is dependent,
        and runs synchronously with, the main system CPU.  Some VLB cards
        are not designed to run faster than 33 MHz, though some mother-
        boards will clock the bus at up to 50 MHz!  VLB 2.0 has been
        written, but has not been implemented on many 486 motherboards.
VRAM    Video Random Access Memory.  A specialized type of DRAM, VRAM
        is dual-ported, meaning it can be read from and written to at
        the same time.  see "What is the difference between VRAM and

Appendix B - Popular Video Chipsets

Some of the information in this section was taken from Boogyman's video
FAQ.  For more detailed chipset info, or info on a particular  video
card model, refer to the Chipsets document which is distributed as a
supplement to this FAQ.  The relevant sections are included
with the Video Chipset Information List.  Note that while much of the
information in this section is fact, by its very nature, some of the
following is based on opinion.  Please don't submit flames - any messages
that are of the 'my video card is the best' will be sent to /dev/null.
If you disagree with any of the comments below and can provide reasonable
justification, feel free to email the FAQ maintainer.  Note that 'my
card is _really_ fast' isn't good enough evidence.

Alliance Promotion
   32, 64 and 128 bit chipset with DRAM, EDO and video acceleration
capabilities.  Their 6422 chipset is faster than the CL5434 and some S3
chips for Windows (according to Alliance) but DOS/VGA speed is poor.

ARK 1000PV, 2000PV
   A relative newcomer to the video arena, the ARK2000 based cards are
the fastest DOS/VGA performers available.  They are inexpensive, 32
(1000) & 64 bit (2000) processors that provide some acceleration
capabilities, but are only available in DRAM versions.  Of course, for
any VGA application, VRAM would provide no benefit over DRAM anyway.
(i.e. Hercules Stingray 64/Video, Actix Picasso 64)

ATI (see Mach 32 & Mach 64)

Avance Logic Inc. -ALI (ALG2301/ALG2228)

   32 bit DRAM based chipset with limited acceleration features.  A 64-bit
chip is also available.

Cirrus Logic (542x)
   C.L. based cards have become the de facto entry level video adapters
for VLB and PCI.  While somewhat faster than the older Trident 8900,
C.L. chipsets are economy models, low on price and acceleration.  These
chipsets are 32 bit and support DRAM up to 2 MB for some models.
Because they are so common, they are well supported across different
operating systems.

Cirrus Logic (543x/544x)
   This is the 64 bit replacement for the 542x series of chips.  Has
some acceleration features and in general are good cards for their
modest price.  The GD5430 has only 32 bit DRAM interface even with
2 MB installed, but other models have full 64 bit interface.

   While no longer in production, this is a popular 32 bit chip.  Its
speed has been surpassed by many newer chipsets, but it still provides
good performance, and is supported widely.  Supports up to 2 MB of VRAM
or DRAM and 64 bit memory transfers (interleaved).
(i.e. ATI Graphics Ultra Pro)

Mach64 (88800)
   This 64 bit chip was designed by/for ATI.  It provides accelerated
GUI performance and respectable VGA speed.  Support across many OS's
and most buses is available.  Mach64 based cards have been given
first place honours in many PC magazine video card rankings due to
good performance and excellent drivers and utilities.  Newer versions
of the Mach64 provide video acceleration, while the Rage 3D accelerator
adds 3D acceleration.
(i.e. ATI Graphics Pro Turbo, WinTurbo, Xpression, Video Xpression)

Matrox MGA
   Typically, Matrox cards are blisteringly fast for GUI's (typically
Windows 3.1) and are considered a high-end chipset because of their cost.
However, the VGA chipset used on Matrox cards is abysmal, and is
usually much slower than even the cheapest VGA cards.  Matrox has
released a new card called the Millenium which is very fast for GUI's
and has fast VGA performance.  The newer Mystique also provides fast
GUI and DOS/VGA speed.  It uses the new high-bandwidth, low-cost
WRAM technology.

Oak Technologies Inc.
   Provide low end SVGA chipsets, some available with up to 2 MB.
Performance isn't spectacular, and are fairly inexpensive.

S3 ViRGE, Vision864, Vision868, Vision968, Trio64, 805, 911, etc
   One of the most popular chipsets, S3-based cards seem to appear at
or near the top of most Windows 3.1 accelerator top ten rankings.
The S3 family enjoys good support across most operating systems and
the 864/964 and Trio64 provide very fast performance for GUI's and
respectable VGA speed.  Typically the 8xx series are DRAM based, while
the 9xx are for VRAM.
   The 911 and 924 were the first generation chips which came with
VRAM.  Provide good GUI acceleration but poor VGA performance.
   The 32 bit chipsets consist of 801, 805, 928 and Trio32 (732).  The
801 is a low end chip which is faster than comparable C.L. chips.  The
805 supports VLB and the 805i supports interleaved DRAM.  The 928 is
a high-end 32 bit card and the Trio32 is an attempt to dominate the
low-end 32 bit market - used DRAM.
   The 64 bit chipsets are ViRGE, 864, 964, Trio64 (764),V+ 868 and 968.
These chipsets dominate many top 10 listings for economical but fast GUI
accelerators.  The Trio64 is basically an 864 with integrated RAMDAC
but isn't software compatible.  All have a 64 bit memory interface.
The x68 chips have additional video acceleration capabilities, as does
the Trio64V+.
   The ViRGE has a Trio64V+ core with additional 3D acceleration features.
(i.e. Hercules Terminator 64/DRAM (Trio64), STB Powergraph 64, Diamond
Stealth 64 (Trio64), rPC FireStorm64, Paradise Bahamas 64 (Vision864),
Diamond Stealth Video (Vision868), Hercules Terminator Professional

   For a long time, Trident chipsets (89xx) were the most common entry
level chips used on ISA boards.  They aren't very fast for VGA or
GUI but enjoy broad-based SVGA mode support.  The newer Trident chips
(94xx) are faster and provide some acceleration features, but are
still one of the slowest chipsets available.  The more recent 9680 is
a respectable GUI accelerator.

   Tseng's latest chipset gives blazing DOS/VGA performance and fast
Windows performance.  Many vendors offer ET6000 based cards.  It
supports DRAM and EDO DRAM sporting a 128-bit accelerator and on-chip
ET4000 - W32, W32i, W32p
   Traditionally the fastest VGA chipset, the ET4000 has recently been
put into second place behind the ARK2000 chipset(s).  The original
ET4000 chipset is not accelerated, but the newer W32's are.  It enjoys
good support across most OS's, but lacks the bandwidth (being a 32 bit
chip) to provide high-resolution, high-refresh truecolour performance.
The 'p' denotes PCI, 'i' and 'p' are capable of memory interleaving.
(i.e. Hercules Dynamite Pro(W32i) or Power (W32p), Cardex Challenger,
Diamond Stealth 32 (discontinued))

UMC 8710
   This is a less common 32-bit interleaved DRAM-based chipset.

Weitek P90xx and P91xx
   Very fast VRAM-based accelerators have no VGA support which must
be provided by a separate chip.  32-bit.  Weitek is not providing
drivers for Windows95, so support is vendor-specific and spotty at

Western Digital (Paradise)
   The WD90C3x chipsets are 32-bit DRAM based and are available in up
to 2 MB version.  Have some acceleration capabilities.  Philips
Electronics purchased Western Digital's video chipset division in

Appendix C - Circuit for Converting from VGA to Fixed-Freq. RGB

More information and additional circuits are available on the PC Video
FAQ WWW Site.  They have not been included here for brevity and because
some of the information is presented in a graphical manner.  Please
refer to
for more information and superior circuits for driving workstation

The following is an ASCII file of a circuit which combines the vertical
and horizontal sync signals from a VGA card to output a composite sync
signal compatible with many fixed frequency monitors.  If necessary, the
Csync line can be connected with the green video line to produce a
sync-on-green signal.  This circuit produces a signal that is compatible
with many fixed-frequency monitors, but ensure that the signal you are
sending has the same vertical and horizontal frequencies as the monitor

For more information on how others have done this, read pertinent parts
of this FAQ and refer to:  - Experiences hooking up
                                                   a Hitachi monitor - More experiences with a Hitachi       - Fixed Frequency PC Video FAQ

*** USE AT YOUR OWN RISK - Others have used this circuit but I have not!

[From: Roger Wolff (]

   VGA connector                                        monitor

      R ----------------------------------------------------- R
    gnd ----------------------------------------------------- gnd

      G ----------------------------------------------------- G
    gnd ----------------------------------------------------- gnd

      B ----------------------------------------------------- B
    gnd ----------------------------------------------------- gnd

                   |           |
                   |  74HCT86  |
                   |           |
                   |           |
  hsync -----------|1          |
                   |          3|----------------------------- Csync
  vsync -----------|2          |
                   |           |
                   |           |
                   |           |
           gnd--+--|7        14|--+-- +5V
                |  |___________|  |
                |                 |
                        0.1 uF

Use COAX cables for the "data" (R, G, B) lines.  You can use just
about anything for the hsync and the vscyn lines but keep them as
short as possible. Around 30 cm (a foot) is fine.  I use a COAX cable
for the Csync line too, as I need a BNC connector at the end of the
monitor anyway. I didn't do anything about termination, and all seems
to be A-OK.

I gather the power for the 'HCT86 from inside the computer. You can
find something yourself (find a 5V powerline going to the leds or
something), or use something that resembles those "add-on" fans.
The latter usually use 12V, but they show the principle: a male and
a female cable connector and 6 wires should do the trick.

This worked for my Grayscale monitor (where the R and B lines are
not needed), and now works just fine for my 21" monochrome monitor.

I start X at boot-time, before ANYTHING runs: my Xserver owns PID 3.
I then quickly bring up the network and start an Xconsole. This
allows me to follow the rest of the bootsequence.

END of FAQ - Part 4/4

Michael J. Scott                       R.R.I.,  U of Western Ontario                 'Need a good valve job?' 
PC Video Hardware FAQ:
###############  Illegitimus non tatum carborundum.   ##############

User Contributions:

Comment about this article, ask questions, or add new information about this topic:


Part1 - Part2 - Part3 - Part4

[ Usenet FAQs | Web FAQs | Documents | RFC Index ]

Send corrections/additions to the FAQ Maintainer: (Michael Scott)

Last Update March 27 2014 @ 02:12 PM