From: scott@bme.ri.ccf.org (Michael Scott) Newsgroups: comp.sys.ibm.pc.hardware.video,comp.answers,news.answers Subject: comp.sys.ibm.pc.hardware.video FAQ, Part 1/4 Followup-To: poster Reply-To: scott@bme.ri.ccf.org (Michael Scott) Distribution: world Approved: news-answers-request@MIT.EDU Summary: This is a monthly posting containing a list of Frequently Asked Questions (and their answers) pertaining to video hardware for IBM PC clones. It should be read by anyone who wishes to post to the comp.sys.ibm.pc.hardware.video newsgroup. Expires: 17 Feb 1997 00:00:00 GMT Archive-name: pc-hardware-faq/video/part1 Posting-Frequency: monthly (second Monday) Last-modified: 1997/02/20 Version: 1.0 URL: http://www.heartlab.rri.uwo.ca/vidfaq/videofaq.html ********************************************************************** COMP.SYS.IBM.PC.HARDWARE.VIDEO Frequently Asked Questions ********************************************************************** This FAQ was compiled and written by Michael Scott with numerous contributions by others, most notably Ralph Valentino who does a great job of keeping up the main csiph FAQ, Sam Goldwasser who has developed and now maintains the majority of diagnostic and repair FAQs for sci.electronics and sci.electronics.repair, Bill Nott of Compaq Computer Corporation and Dylan Rhodes of Hercules Computer Technology. Acknowledgments are listed at end of this FAQ. ********************************************************************** Posting to comp.sys.ibm.pc.hardware.video - please read! ********************************************************************** For general information and rules on posting to the c.s.i.p.hardware hierarchy, please refer to the main csiph FAQ, sections 1.2 - 1.6Jump to:list of FAQ questions to ensure your question hasn't already been answered here! If it has not been answered: Be as specific as possible. If you are having video problems, please include the following information: Symptoms - What exactly are the symptoms? Where do the symptoms exhibit themselves? i.e. only in Windows, or with certain applications. When did the problem start? Did it ever work properly? If so, what has changed since? Under what circumstances are the symptoms seen? Is the problem repeatable or intermittent? What have you tried and with what results? Hardware Configuration - CPU, RAM, bus type (ISA, VLB, PCI), video card model, amount and type of video RAM, monitor model if appropriate, video extension cables if used, resolution/colour depth/refresh rate if using SVGA or better resolutions. Software Configuration - operating system and version, video card driver and version, name and version of conflicting software. Anything else unique about your system? *** Email - make sure the email address in the From: field of your posting is valid. - as a courtesy to others, keep your .signature to 4 lines or less. Remember, if you include all of the right information the first time, you'll get an answer back faster, _and_ reduce unnecessary traffic on the net! Remember to try official channels first - often the manufacturer can answer common questions quickly. ********************************************************************** Copyright notice: ********************************************************************** The comp.sys.ibm.pc.hardware.video Frequently Asked Questions is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY. No author or distributor accepts responsibility to anyone for the consequences of using it or for whether it serves any particular purpose or works at all, unless he says so in writing. Refer to the GNU General Public License for full details. Everyone is granted permission to copy, modify and redistribute this FAQ, but only under the conditions described in the GNU General Public License. Among other things, the copyright notice and this notice must be preserved on all copies. Where section authors are noted, the copyright is held by that author. Where no author is noted, the copyright is held by the FAQ editor Michael Scott (scott@bme.ri.ccf.org) or (mjscott@heartlab.rri.uwo.ca). Unless otherwise specified, contributors are speaking for themselves in a personal, not professional capacity, and do not represent their employers or any other organization. If you'd like to contribute to the FAQ via comments, additional sections, posing questions or (ulp!) corrections, please email: Michael Scott (scott@bme.ri.ccf.org) or (mjscott@heartlab.rri.uwo.ca). ********************************************************************** Latest and Greatest: ********************************************************************** If you are concerned that this copy of the FAQ is out-of-date, copies are archived at rtfm.mit.edu or its mirrors in the /pub/usenet/comp.sys.ibm.pc.hardware.video directory. Alternatively, you can browse the latest FAQ and download text or compressed versions at: http://www.heartlab.rri.uwo.ca/vidfaq/videofaq.html or get a compressed text-only version from: ftp://ftp.worcester.com/pub/PC-info/pc-hardware-video-faq.Z ftp://ftp.worcester.com/pub/PC-info/pc-hardware-video-chipsetlist.Z For additional ways to retrieve the latest version of this FAQ, refer to question 1.2 in the main comp.sys.ibm.pc.hardware FAQ, part1: ftp://rtfm.mit.edu/pub/usenet/news.answers/pc-hardware-faq/video/part1 ********************************************************************** New: ********************************************************************** New sections have been added to the PC Video FAQ Web Site. These contain information of a graphical or web-centric nature, and so haven't been included in the text version of the FAQ. The additions include: Identifying video card components Video related sites on the web Circuits for driving fixed frequency monitors ********************************************************************** Table of Contents:
What does the csiph.video FAQ cover?
What are true color and high color?
(Extremely low memory, some icons may not be drawn)?
In addition, you may be interested in the PC Video Chipset List, which was originally compiled by Boogyman. It contains a list of common video chipsets with a brief description, and a list of video card models and the video coprocessors that they use.
Questions marked with an asterisk (*) will be answered in a future release of this FAQ. ********************************************************************** S) PC Video Frequently Asked Questions ********************************************************************** Q) What does the csiph.video FAQ cover? Issues related to pc compatible video systems are covered here. This FAQ is primarily intended for hardware, but some software issues are also considered. The hardware components that are dealt with include, but are not necessarily limited to: video adapters monitors video terminology video capture cards video playback add-in cards i.e. hardware MPEG decoder Return to Table of Contents Q) Are there other sources of info on video related subjects? Some information is available on-line. Because some sites are less stable than others, you may have to try a given site a couple of times. For best results, try contacting on off-peak hours. If you try a site at three or more times and can't connect, please email the FAQ maintainer and that site will be removed from this list. If you find a useful site that isn't listed here and seems to be fairly stable, please send it in. Last checked: 96/02/28 http://hawks.ha.md.us/hardware/monitor.html : Monitor info http://www.devo.com/video : Fixed frequency PC video FAQ http://www.cviog.uga.edu/monitors : Info on 480 monitors Also contains links to other monitor resources http://www.cviog.uga.edu/monitors/monitors/manufacturers.html: : List of phone numbers and WWW sites for 60+ monitor companies http://www.cs.columbia.edu/~bm/3dcards/3d-cards1.html & http://www.cs.columbia.edu/~bm/3dcards/3d-cards2.html : FAQ for 3D graphics accelerators http://www.dfw.net/~sdw/index.html : System Optimization Information http://www.garlic.com/sid/ : The Society of Information Display http://www.hercules.com/knowbase/ : Bug report and fixes for Hercules http://www.hercules.com/monitors : WWW Monitor Database by Hercules http://www.noradcorp.com : NoRad Corporation - info on EMF's, standards and filters http://www.paranoia.com/~filipg/HTML/FAQ/BODY/Repair.html : Part of sci.electronics FAQ which contains monitor repair info http://www.vesa.org/ : Video Electronics Standards Association includes various VESA standards documents in Adobe Acrobat files but is only accessible to VESA members! Non-members can order standards - a price list is available. http://www.ziff.com/~cshopper : Has a variety of articles from back issues of Computer Shopper related to PC's including video cards and monitors Also see the references at the start of Appendix A - Glossary. Return to Table of Contents Q) Can I use two video cards in the same system? [From: rbean@execpc.com (Ron Bean)] Q) Can I use two video cards in the same system? There are several ways to do this. You may have heard that it's only possible if one of them is the old Hercules monochrome card, but other options have been available for several years. It is being widely reported that "Windows 97" (or whatever they end up calling it) will include dual-monitor support, using any two PCI video cards. This will be slower than other methods due to the complexity involved. Apparently this will also appear in some future version of NT. The reason this is possible is that the PCI bus allows the cards to map themselves onto different addresses and interrupts, avoiding the built-in conflicts of earlier fixed-address designs. You will have to disable "VGA emulation" on all but one of the cards, and this may not be possible on some cards (one VGA card is still needed to boot the system). A few PCI video cards already have special drivers available that allow more than one (identical) card to be used at once. The Matrox Millenium is often mentioned for Win95, and a few others have drivers for NT (usually aimed at the high-end CAD or DTP markets). Several companies make special video cards that contain two (or four) VGA chipsets, so that several monitors can be attatched to one card. These also usually allow for more than one card to be used at once-- in one case up to 16 monitors using 4 cards! ISA, VLB, and PCI cards are available. They cost more than two regular cards (especially if you're adding to a system that already has one card), but they should run faster and they come with more sophisticated control software. These are the manufacturers I'm aware of: Appian Graphics, Colorgraphics, STB, Tridium DuoGraphics, Diamond Multimedia, Datapath and Miro. To view product information use your WWW browser to access their sites - pointers are available at: http://www.heartlab.rri.uwo.ca/vidfaq/vendors.html For Linux (and other unix variants) there are commercial X servers that support multiple PCI video cards, from Metrolink (MetroX) and Xi Graphics (Xinside). Xfree86 does not support multiple monitors at this time, unless one of them is the old Hercules mono card. Pointers to these sites are available at the web page above. There is also a program called x2x which allows the keyboard and mouse from one X display to control another X display (eg, an X terminal or another computer). See: ftp://gatekeeper.dec.com/pub/DEC/SRC/x2x/ The original PC design used separate address spaces for the CGA (color) and MDA (monochrome) cards, and both could be installed at once, although DOS will only use one at a time (switch with the MODE command). The Hercules monochrome card used the same address space as the MDA card, and all subsequent color cards have used the same address space as CGA, which means they don't conflict with the Hercules card but do conflict with each other (until the PCI bus came along and changed that). Debuggers and CAD programs and often took advantage of this by including support for displaying information on both monitors at once. This still works if your application supports it, but there are a couple of catches. First, the Hercules mono card and all of its clones are 8-bit cards, and they require the other card to be in 8-bit mode also (you may have to set a jumper for this, or it may auto-detect). Note also that many cheap clone monochrome cards include CGA emulation, and there may be no way to disable it. Windows 3.x can also be set up this way. Include the line DualDisplay=TRUE (or ON) in your SYSTEM.INI file, in the 386enh section. If you open a DOS shell window and type MODE MONO, the shell will appear on the monochrome monitor (I don't know if this still works in Win95). If you just want to display the same image on several monitors, there are (expensive) signal splitters that will do this (computer stores often use them for their demo monitors). See "How can I hook more than one monitor to my video card?" Return to Table of Contents Q) How can I hook more than one monitor to my video card? [ From: Sam Goldwasser (sam@stdavids.picker.com) with a bit from Michael Scott (scott@bme.ri.ccf.org) and Bill Nott (BNott@bangate.compaq.com)] The following discussion assumes that you want to display the same video signal on a number of monitors. If instead you want use 2 or more monitors to increase your screen real estate, refer to the section "Can I use two video cards in the same system?". The best way to do this is to purchase a commercial VGA signal splitter or video distribution amplifier.. These are not cheap, but they will provide the best results. A video splitter designed for VGA or SVGA will include the proper high bandwidth video amplifiers as well as the proper cable termination and shielding. Someone may suggest that you just cut and splice a couple of VGA cables together, but this won't provide good results. Major problems relate to cable termination and interference. In order for the video to be sharp and clear without ghosting or ringing, the video cable must be treated as a transmission line. What this means from a practical point of view is that it must use high quality coaxial cable, multiple monitors must be daisychained and not star connected, and the proper terminating resistors must be put only at the very end. Another problem is that video signals operate at high frequencies, and as a result they can cause interference with neighbouring electronic devices, and even the monitor itself. In fact, the video cable can, when designed improperly, act like a nice big antenna. To minimize the interference emanating from the cable, considerations like conductor material, length, shielding, connectors and chokes are taken into account. Chokes are those (usually cylindrical) objects that are located at the ends of many video cables. The result of a good cable design is an impedance matched circuit, which causes a minimum amount of interference, and provides a clean crisp signal to the monitor. If you know enough about electronics, and the monitors and video card in question, then go ahead and design and build a splitter. If you don't, you may cause additional problems. Basic rules for a cable-only solution: 1. Use high quality 75 ohm coax - RG59 is a generic part number but many variations are available. 2. Multiple monitors must be daisychained and not split in a star configuration. 3. Only the last monitor should have its 75 ohm terminating resistors in place. They should be removed from all other monitors or if they have switches, set for HiZ. 4. Pay attention to the grounds - signal returns. Keep the stubs - the connections to intermediate monitors - as short as possible. This will work quite well for workstation monitors - those with BNC coax connectors. Most PC monitors with the 15 pin VGA connectors do not have any means of disconnecting the terminating resistors without actually doing some desoldering - which you really should not attempt unless you are familiar with the safety issues involved in working inside a monitor. If you decide to build an active video splitter which uses video amplifiers, be aware that the video and sync voltage levels are different in a PC: The video is typically 0.7 V p-p, and the sync's are typically TTL level (5 V p-p), so the splitter or amplifier must be able to handle both levels. Finally, pay attention to the video bandwidth capability of the splitter/amp if you care about preserving image detail information. As noted a better solution is to buy an active video splitter. This will include the proper high bandwidth video amplifiers and termination. Return to Table of Contents Q) Can I use my TV as a monitor? Monitor prices are constantly dropping, but not as fast as many people would like. A nice 17+ inch monitor is still $800 or more. It would be nice if a cheap 20 or 27 inch television could be used instead. Well, there are products available which will convert a VGA signal to one that is compatible with your television set. Below, I will outline the limitations that this type of setup has compared to a proper computer monitor. This may or may not be a viable option for you depending on what types of applications you use most frequently. To understand what is and isn't possible, we need to know the differences between a computer's video signal and the one expected by your television (usually NTSC (North America) or PAL (Europe)). Typically, PC's display in one of 320x200, 320x240, 640x480, 800x600, 1024x768, 1280x1024 or 1600x1200. The lowest three pixel addressabilities are supported by VGA, and are the most common video modes used for VGA (often called DOS) games. The higher resolutions are typically used when in windowing environments like Windows 3.1, OS/2, Win95 or X-Windows. In addition, computers refresh the screen at varying rates, ranging from 50 to 75 or more Hertz (cycles/s). Most newer monitors support non-interlaced video at all or all but the highest pixel addressabilities. The digital signal is converted into an analog one that the computer monitor understands. If the signal is within the capabilities of the monitor, it will be displayed as a screen image. Televisions are also analog devices, like computer monitors, but they are designed to accept a broadcast television signal like NTSC or PAL. The NTSC standard supports a maximum of 525 lines, while PAL supports up to 625. The screen refresh rates are fixed for NTSC and Pal as 60 and 50 Hz respectively. In addition, both standards are interlaced. It is important to note that since computer monitors and televisions are both analog devices, the number of colours is not a factor here. To compare this to computer video modes, we have to do a little bit of hand-waving, but basically, the best North American televisions can't display more than about 500 lines. This roughly translates to a maximum of 500 vertical pixels. In addition, the video amplifiers used in televisions are fairly low bandwidth devices, and can't handle high horizontal resolutions like 1024 or 1280 pixels. What this boils down to is that standard televisions can display a video signal derived from up to a 640x480 pixel mode. To summarize, a VGA to TV converter can be used to translate the VGA's RGB signal to a TV compatible signal for resolutions up to 640x480. This means that the vast majority of VGA/DOS games will display reasonably well on a television since most use 320x200 or 320x240 video modes. Converters that claim to handle higher resolutions have scan converters in them that reduce the effective resolution to that of NTSC or PAL television. i.e. You simply can't display better than 640x480 on a TV. Some PC to TV converters are listed in the Chipsets section of this FAQ. Return to Table of Contents Q) Can I use my CGA/EGA/VGA monitor as a TV? CGA and EGA monitors are digital, rather than analog like televisions and more modern monitors, usually making them incompatible with TV. Television signals contain all colour information along with syncs on one conductor. In addition, there are two types of television signals - the RF that comes in from cable or an antenna, and composite. The line-in/out on a VCR is a composite signal, and doesn't contain all of the different channel information that an RF cable signal does. The original CGA monitors accept a composite signal, but it is TTL, which uses a different voltage from composite. Some CGA (and perhaps EGA?) monitors have composite-in jacks and circuitry inside them to display a composite signal. If you have one of these, then you can feed it a composite video signal from a VCR, laser disc player or other composite video source. Since the VGA/SVGA monitor was introduced, computers have used an RGB video signal, with separate horizontal and vertical syncs. This means that five separate wires are used to carry the video signal from the computer to the monitor. In order to display a TV signal on a VGA monitor, signals for all five wires have to be derived from one, the so-called composite TV signal. This involves some electronic circuitry, so it can't be accomplished simply by attaching all of the wires together. Because of the demands of higher pixel addressabilities and refresh rates, VGA and newer monitors run at horizontal refresh rates of 30 kHz or higher, which is double that of composite video (15.7 KHz). Basically, these newer monitors are unable to sync to a low enough frequency to display broadcast (NTSC or PAL) video. The end result is that it is not feasible to use a VGA or better monitor to display a television signal. The only real alternative is to purchase a TV card for your computer which allows you to display a television signal on your monitor. Personally, I'd rather spend the money on a small TV rather than look at a four inch window on my already cramped computer monitor. Return to Table of Contents Q) What kinds of monitors are available? Since there is a large variety of different types available, only some of the more common are listed here, along with their most common applications. In fact, it's difficult to define exactly what a 'kind of monitor' means. There are grayscale and colour, analog and digital, flat and not. I'll try to give some general answers. Monochrome, Grayscale and Colour This one's easy. Monochrome monitors can display two colours, usually black and one of white, green or amber. Grayscale monitors display only intensities between white and black. Colour monitors display combinations of red, green and blue, each in an independent intensity. Even though each colour is displayed only in one frequency (the frequency of light that a particular type of phosphor emits when excited) the combination of the three colours in different intensities fools the eye such that it perceives a full range of colours. Analog and Digital [From: Michael Scott (scott@bme.ri.ccf.org) and Sam Goldwasser (sam@stdavids.picker.com)] Today, digital monitors are much less common than analog though in the days of CGA and EGA the situation was reversed. Digital does _not_ mean that the monitor has digital controls. Rather, it indicates that the monitor accepts a digital input signal. Examples of digital monitors include early monochrome, the IBM EGA and CGA. Digital monitors are limited by their internal hardware as to the number of colours that they can display. Most digital monitors use TTL signals (Transistor Transistor Logic). Note that some sales persons will call a new analog monitor 'digital', in reference to the controls. Strictly speaking they are wrong - see "Analog vs. Digital Controls" below. Analog colour monitors can display an unlimited range of colours, since they accept an analog video signal. This means that the horizontal and vertical syncs, and actual video signals (usually red, green and blue) are analog. The total number of colours that a given computer system with an analog colour monitor can display is limited by the video card, not the monitor. It is rare for video cards to use digital-to-analog converters capable of generating more than 256 intensities per colour, so it is rare for systems to be able to display more than 256*256*256 equals 16.7 million colours. Analog monitors can have digital controls on the front panel, and have digital circuitry inside. The vast majority of monitors currently in use are analog, as they are more flexible than the digital variety and typically lower cost. Most graphics cards put out an analog _or_ digital signal but not both. Similarly, most monitors accept and analog _or_ digital signal. It is feasible, however, to convert a digital video signal to analog and vice versa, though building such a device requires considerable electronics knowledge. Shadow Masks and Aperture Grilles By far the most common type of monitor uses a shadow mask, which is a fine metal grid which enables the electron beams for red, green and blue to only impact their proper phosphor dots. One alternative to this design is the aperture grille, which uses fine vertical wires for the same purpose. Sony first used this aperture grille in their Trinitron line. Which one is better is not clear cut and is largely a matter of personal preference. Note that one complaint of Trinitron users is the presence of 1 or 2 very fine, almost invisible, horizontal stabilizing wires apparently needed to keep the fine aperture grill wires from moving out of place. You need to decide whether these will prove an unacceptable distraction. Trinitrons are usually considered to be brighter and sharper - but this is not always the case. Analog vs. Digital Controls [From: Michael Scott (scott@bme.ri.ccf.org)] An analog monitor can have either analog (dials or knobs) or digital (buttons, sometimes with a dial) controls for brightness, contrast, screen size and position, pincushioning and trapezoidal shape, among others. Also, digital controls tend to be associated with a monitor's ability to store factory and user calibrations for image size and centering when operated at common video modes. This is desirable for a user who may be switching between DOS and windows applications often, so they don't have to be bothered with readjusting these controls after each change. Analog controls have the benefit of being infinitely adjustable, while digital controls are limited to a number of discrete steps for each adjustment. Flat Panel vs. Conventional Tubes Cathode ray tubes (CRT's) are the most common, inexpensive and best performing displays available for most users. Variations of CRT's exist including older designs with double curvature, some with only curvature in the horizontal plane (like Sony Trinitrons) and others which are called flat screen. Flat panel displays are usually used in laptops because of their small size, but are expensive to manufacture and don't provide the high refresh rates and bright colours that conventional CRT technology provides. Flat panel displays range from monochrome LCD (Liquid Crystal Display) to dual scan colour to active matrix colour. Because of the difficulty of manufacturing these displays, and the fact that currently their primary application is in laptops where the maximum display size is usually less than eleven inches, high resolution flat panel displays are rare and expensive. In future, it's very likely that flat panel displays will replace conventional CRT technology for many home and business computer users. Return to Table of Contents Q) What types of flat-panel displays are available? [From: Michael Scott (scott@bme.ri.ccf.org) and some from Bill Nott (BNott@bangate.compaq.com)] Flat-Panel Display (FPD) technology is evolving rapidly, so I will only touch on the most common current types of displays. There are other types of displays still in use, though the most common ones are based on LCD (Liquid Crystal Display) or PDP (Plasma Display Panels) technology. Now, FPD's are expensive due to the difficulty in manufacturing (typically ~65% yield - ~4 in 10 are discarded) and relatively small number of units sold. As manufacturing techniques improve and volume increases, prices will drop. In fact, in 1995, yields are up, volumes are up, _and_ factory capacity has expanded to the point where prices are dropping significantly this year. It appears there will be an oversupply of panels this year. However, the prices are still not down to the point where they can compete with CRT monitors in desktop applications. [From: Michael Scott (scott@bme.ri.ccf.org)] The vast majority of FPD's are addressed in a matrix fashion, such that a given pixel is activated by powering the corresponding row and column. This means that an individual LCD element is required for each display pixel, unlike a CRT which may have several dot triads for each pixel. LCD displays consist of a layer of liquid crystal, sandwiched between two polarizing plates. The polarizers are aligned perpendicular to each other, so that light incident on the first polarizer will be completely blocked by the second one. The liquid crystal is a conducting matrix with cyanobiphenyls (long rod-like molecules) that are polar and will align themselves with an electric current. The neat feature of these molecules is that they will shift incoming light out of phase when at rest. Light exiting the first polarizer passes through the liquid crystal matrix and is rotated out of phase by 90 degrees, then it passes through the second polarizer. Thus, unpowered LCD pixels appear bright. When an electric current is passed through the crystal matrix, the cyanobiphenyls align themselves parallel to the direction of light, and thus don't shift the light out of phase, the light is blocked by the second polarizer and the LCD appears black. So, basic LCD technology can generate bright or dark pixels, like a monochrome (not grayscale!) monitor. In order for the eye to see shades of gray, the LC activation time is modulated. i.e. a pixel that is activated 50% of the time will appear as 50% gray. The number of shades that can be generated without visible flicker is limited by the response time of a LC element - typically 16 shades, although some display manufacturers claim 64 or more shades. Most colour LCD's use red, green and blue sub-pixels, similar to the way that CRT's use coloured dots of phosphor. The concept is the same; that when viewed from a distance, the human eye will perceive the three sub-pixels as a single colour. Obviously, this requires three times as many discrete elements as would a monochrome display of the same resolution. A second method of implementing colour uses a subtractive CYM (Cyan Yellow Magenta) system where white light is generated at the back plane. The light then passes through each of three LC layers, each one blocking one of the three colours. By activating the LC layers in different combinations, a variety of colours can be produced. Common to all LCD displays is the requirement for either high ambient light levels, or bright backlighting since liquid crystals don't generate light - they can only block it. Typically, LCD's allow 5-25% of incoming light (i.e. from the backlight source) to pass through. The result of this is that LCD technology requires a significant amount of energy, and this is an important consideration in light- weight laptop design. Specific type of LCD's Passive Matrix (twisted-nematic) LCD's PM LCD's come in several types including; supertwisted nematic, double supertwisted nematic and triple supertwisted nematic. The original PM LCD's had a very limited viewing angle and poor contrast. Super and double supertwisted nematic designs provide an increased viewing angle and better contrast. The triple supertwisted design implements the subtractive CYM colour model mentioned above. PM designs are addressed in matrix fashion, so a VGA PM display would require 640 transistors horizontally and 480 vertically. Rows of pixels are activated sequentially by activating the row transistors while the appropriate column transistors are activated. This means that a given row is activated for only a short time during a screen refresh, resulting in poor contrast. Some implementations of PM technology break the screen into two parts, top and bottom, and refresh them independently, resulting in better contrast. These are called Dual Scan PM LCD's. In addition, PM displays suffer from very slow response times (40-200 ms) which is inadequate for many applications. Aside from their performance shortcomings, PM displays are inexpensive - their relatively low number of discrete components reduces manufacturing complexity and increases yields. Note that while dual scan displays are better than the original PM LCD's, they still don't have the high refresh rates and brightness of active matrix LCD's. Active Matrix LCD's Instead of using one switch (transistor) for each row and column, AM LCD's dedicate one switch for each pixel. This results in a more complex display which requires a larger number of discrete components, and therefore costs more to manufacture. An AM display is basically a large integrated circuit (IC). The benefits are significant over the PM design. Pixels can be activated more frequently, giving better contrast and control over modulation. AM technology can produce higher resolution displays that can generate more, and brighter colours. The main types of AM LCD's are; TFT (Thin-Film Transistors), MIM (Metal- Insulator-Metal) and PALC (Plasma Addressed Liquid Crystal). Ferroelectric LCD's FE LCD's use a special type of LC which holds its polarization after being charged. This reduces the required refresh rate and flicker. Also, FE LCD's have a fast response time of 100ns. Although they are very difficult to manufacture, and therefore expensive, FE LCD's may provide AM quality at PM prices in future. Plasma Display Panels PDP's have been under development for many years, and provide rugged display technology. A layer of gas is sandwiched between two glass plates. Row electrodes run across one plate, while column electrodes run up and down the other. By activating a given row and column, the gas at the intersection is ionized, giving off light. The type of gas determines the colour of the display. Because it has excellent brightness and contrast and can easily be scaled to larger sizes, PDP's are an attractive technology. However, their high cost and lack of grayscale or colour have limited applications of PDP's. However, advancements in colouring technology have allowed some manufacturers to produce large full-colour PDP's. In future, large colour PDP's will be more common in workstation and HDTV applications. Return to Table of Contents Q) What do those monitor specifications mean? Refer to Appendix A - Glossary for definitions of terms not included in this section. Like so many other areas in high-technology, a bewildering array of models are available, and along with them comes a list of specifications. There are a few that will help you understand more about the differences between specific models. [Thanks to Bill Nott for straightening me out on bandwidth and dot clock] Bandwidth: This is a measure of the total amount of data that the monitor can handle in one second, and is measured in megahertz (MHz). The bandwidth of a monitor is limited by the design of the video amplifiers. It is generally desirable to match the bandwidth of the monitor with the dot clock of the video controller to take full advantage of both devices. see dot clock. see 'How do I calculate the minimum bandwidth required for a monitor?' Dot Clock: This is the clock frequency (in MHz) used by the video controller chip, sometimes termed pixel rate. Many newer graphics processors have variable dot clocks, but usually only the highest is quoted in specifications. It is a measure of the maximum amount of throughput that a video controller can sustain. A higher dot clock generally means that higher screen addressabilties, colour depths and vertical refresh rates are possible. If you want to know the _approximate_ maximum dot clock for your video card and it isn't specified, you can calculate an approximate value (which tends to overestimate) as outlined in "How do I calculate the minimum bandwidth required for a monitor?" Horizontal Scan Rate (HSR): This is a measure of how many scanlines of pixel data the monitor can display in one second. The electron gun has to scan horizontally across the screen and then return back to the beginning of the next line ready to scan again. It is controlled by the horizontal sync signal which is generated by the video card, but is limited by the monitor. If too much data (i.e. too high a horizontal pixel addressability) is sent to the monitor, it exceeds its ability to modulate the electron gun, and the signal will be displayed incorrectly and/or the monitor may be damaged. VGA and SVGA monitors must have a minimum HSR of 31.5 kHz to be able to display the corresponding horizontal resolutions. Now we begin to see how the vertical refresh rate and the horizontal scan rate are related. Refresh Rate (also Vertical Refresh Rate or Vertical Scan Rate): This measures the maximum number of frames that can be displayed on the monitor per second at a given pixel addressability (resolution). It is controlled by the vertical sync signal coming from the video card. The vertical sync tells the monitor to position the electron gun(s) at the upper left corner of the screen, ready to paint another frame. The maximum rate for a given monitor is dependent on the frequency capability of the vertical deflection circuit and the pixel addressability, since higher addressabilities require a higher horizontal scan rate. For example, a monitor which can provide 72Hz refresh rate at 800x600 may only be capable of 60Hz refresh at 1024x768. In order to be considered a VGA or SVGA monitor, the unit must provide a minimum vertical refresh rate of 60Hz. In general, higher is better, but there is no point in paying more for a video card and monitor which are capable of higher refresh rates if you won't notice a difference. 60 Hz is adequate for most people, but others are bothered by flicker and prefer 72 Hz or faster to reduce eye strain. The minimum acceptable refresh rate for you may also depend on the screen resolution and monitor size. In general, higher addressabilities require higher refresh rates to prevent flicker from becoming noticeable. A monitor's maximum vertical refresh rate is limited by how fast it can direct the electron beam over all of the picture elements on the monitor. This involves moving the electron beam in the same manner as you would read the words in a book, left to right, top to bottom. It is limited by the maximum HSR, which determines the maximum horizontal pixel addressability the monitor can display and the number of scanlines (i.e. vertical addressability). For example, to display a screen with an addressability of 640 pixels horizontally and 480 vertically, a monitor with a HSR of 31.5kHz would take 480/31.5k = 15.2 ms to scan the entire screen once. In one second, this monitor could be refreshed 1000ms/15.2ms = 65.6 times. However, the vertical sync - movement of the electron gun to the upper left corner of the screen - requires some time, so the resulting vertical refresh rate is only 60 Hz. Built into the HSR and vertical refresh rate are the horizontal and vertical blanking intervals, respectively. During horizontal blanking, the electron beam is moved back across the screen from the right end of one scan line to the beginning of the next scan line on the left of the screen. This occurs once for each scan line displayed. The vertical blanking interval occurs after the last scan line is displayed, and the electron beam is directed back to the upper left corner of the screen to begin displaying the next screen image. Interlacing: Interlacing is a holdover from television standards which use it as a way of putting more information on the screen than would otherwise be possible. Original television technology could handle thirty full frames of video per second. However, a 30 Hz refresh rate results in highly annoying flicker, so the video signal is divided into two fields for each frame. This is accomplished by displaying first the odd scanlines (i.e. 1,3,5, etc.) for 1/60 of a second, and then displaying the even scanlines for the next 1/60 of a second. Your brain can integrate the two fields, and the result is a higher effective resolution and lower flicker. Ideally however, you want to display a frame of video information at full resolution - i.e. have one horizontal scanline for each horizontal line of pixels and display it at a high enough refresh rate that flickering is not an issue. Fortunately, modern monitor technology is capable of non-interlaced (NI) display at high vertical refresh rates. Many non-interlaced monitors can only work in non-interlaced mode up to a maximum pixel addressability, above which they revert to interlaced mode. For this reason, it is important that you ensure that the monitor you buy is capable of non-interlaced display at the maximum addressability and vertical refresh rate that you want to use. Typically, interlaced computer monitors refresh at about 87Hz, or 43.5 full frames per second. Interlaced displays can result in annoying flicker, especially noticeable with thin horizontal lines because the scanline is alternating between the line and background colours. It's very noticeable if you look at the top or bottom edge of a window on an interlaced monitor. Dot Pitch: Images on a computer monitor are made up of glowing blobs of phosphor. On colour monitors, the smallest discrete picture element consists of three phosphor blobs, one each of red, green and blue. These elements are called dot triads. On most monitors the blobs are arranged in rows and columns, often with every other row staggered: R G B R G B R - Red B R G B R G G - Green R G B R G B B - Blue B R G B R G So, in the above example, a shape like the following might be a dot triad: R G B The dot pitch is measured as the shortest diagonal distance between the centers of any two neighbouring dot triads. This is the same as the shortest diagonal distance between any two phosphor blobs of the same colour. As dot pitch decreases, smaller objects can be resolved. Resolution: First, the correct term that _should_ be used in place of resolution for most computer video discussion is pixel addressability. This is because in actuality, when we talk about 'resolution' being say, 640x480, we are referring to how many pixels can be addressed in the video frame buffer. Resolution should actually be defined as the smallest sized object that can be displayed on a given monitor, and so is really more closely related to dot pitch. So, two definitions are given here. The first is technically more correct, while the second is the more common interpretation (though strictly incorrect). The technically correct answer: [From: Bill Nott (BNott@bangate.compaq.com)] Resolution: The ability of a monitor to show fine detail, related mostly to the size of the electron beam within the CRT, but also to how well the focus is adjusted, and whether the video bandwidth is high enough. Note that the dot pitch of a CRT is generally an indication of the tube's resolution ability, but only because the manufacturers try to maintain a spot size enough larger than the dot pitch to prevent Moire' patterning from appearing. The more mainstream usage: This refers to the maximum number of pixels which can be displayed on the monitor at one time, and is expressed as (number of horizontal pixels) by (number of vertical pixels) i.e. 1024x768. While a higher maximum resolution is, in general, a good thing, keep in mind that as the resolution gets higher, the pixel size gets smaller. The resolution capability of a monitor puts practical limits on the maximum pixel addressability a user may want to use. You may notice that most addressabilities are in the ratio of 4:3. This is also a holdover from television technology which uses the same 4:3 aspect ratio. As a result, monitor size can be quoted with one diagonal measure, since the horizontal and vertical sizes can be calculated from the 4:3 ratio. In future, HDTV (High Definition Television) will use 16:9 (the same aspect ratio as used in movie theatres) and this may spill over into computer technology. The following are recommendations: Monitor Size 14" 15" 17" 20" Resolution 640x480 A A B B 800x600 C A A B 1024x768 D C A A 1280x1024 D D C A Legend: A - Optimal B - Grainy, pixels become visible C - Usable, but objects become small and fine detail becomes less distinct D - Not Recommended, objects are difficult to see and fine detail can not be perceived These are only recommendations. Personally, I can only afford a 14" NI monitor, and I run it at 1024x768. Objects are small, but my vision is 20/20 :-). [From: Sam Goldwasser (sam@stdavids.picker.com)] Keep in mind that there is also a very wide variation in the quality of the images between manufacturers and between models. Many factors contribute to this variation including video amplifier bandwidth, sharpness of the electron beam (focus), dot pitch of the CRT shadowmask (or line pitch of a Trinitron's aperture grill), stability of the power supplies, bandwidth of the video card, quality of the cables, etc. [From: Bill Nott (BNott@bangate.compaq.com)] Note: Many monitors are able to operate (synchronize, and present an image) at pixel addressabilities beyond their resolution capabilities. When operated in this way, fine detail (single pixels) within the image may not be perceptible by the user. [From: Bill Nott (BNott@bangate.compaq.com) and Michael Scott (scott@bme.ri.ccf.org)] Size: Monitor sizes are typically quoted in inches, and this is measured across the diagonal length of the monitor i.e. the longest possible measurement. Industry practice has been to list the size of the picture tube as the size of the monitor, but this has lead to some problems. For example, a tube may measure 17" across the diagonal, but due to glass thickness and that the tube is encased in the monitor housing, the viewable area is only 15.5". So, just because two monitors are advertised as being the same size doesn't mean that they have the same viewable area. Part of the source of this inconsistency is that the monitor _tube_ manufacturers do not specify image performance such as focus and convergence up to the extreme edge of the phosphor, so the image size is adjusted to that which the tube supplier specifies. (Many monitors today provide the possibility of adjusting the image size larger than this, but may neglect to tell the user to expect image quality degradation beyond the calibrated image size.) Some users may have allowed themselves to think (or wish) that the size designation should refer to the image size, but this has never been true. Regardless, within the US, the Federal Trade Commission (the body which brought standardization to the TV industry with use of the "V" terminology) is working to produce a standard for computer monitors. Some vendors actually quote viewable area in addition to the tube size, but this is not provided by all vendors yet. Until then, caveat emptor - take a measuring tape with you when you go shopping. Return to Table of Contents Q) What should I consider when buying a monitor? [From: Michael Scott (scott@bme.ri.ccf.org) with contributions from Andy Laberge (tic-toc@wolfe.net) & Bill Nott (BNott@bangate.compaq.com)] Your monitor may be the most expensive option of a new computer system, and is the part that you will be looking at most of the time, so it pays to get the right one for your purposes. You will have to decide what size is appropriate for your work - in general bigger is better, but do you really want to shell out $3000 for a huge 21" monitor that weighs 80 lbs and covers most of your desk? The most common monitor sizes are 14", 15", 17" and 21". See "What pixel addressabilities are best for my monitor?" Make sure that your monitor can display the highest screen addressability that you want to be able to use, and that the refresh rate at that addressability is reasonable (generally >=60 Hz). Note that VESA and European standards groups are moving towards 75 and 85 Hz recommendations, respectively. You should expect to pay more for a monitor capable of higher refresh rates because they use faster video amplifiers and deflection circuits. You also have to know whether the monitor is interlaced or not at the higher addressabilities. In addition, decide what features you would like in your monitor including: pincushioning and/ or trapezoidal controls, individual RGB gain and cut-off controls, remote control, programmable memory for presets, warranty & service, etc. Once you have decided what you want, and have narrowed the field to a few choices, you should go somewhere that you can compare the possibilities beside each other. Typically, CRT manufacturers today do not specify the image performance such as focus, convergence, and geometry, out to the edges of the tube. As a result, you have to evaluate these parameters for yourself. Also, users typically do not want their images overscanned as in TV displays, especially when using GUI's. If monitors overscanned, parts of the image near the edges of the screen may not be visible. Thus, the useable image size of a monitor will be smaller than the maximum useable phosphor area which may be specified for the FTC. VESA has already established and published a standard for useable image size in a computer monitor. Comparing and Testing Monitors First make sure that the monitor(s) has warmed up for at least ten minutes. The heat escaping from the rear of the monitor should not be much more than that generated by a colour television. Some monitors are now coming with fans installed for positive ventilation. Next, adjust the brightness so that the illuminated part of the screen has the same brightness as the unilluminated border. Increase contrast to a reasonable level (fairly high) and reduce screen glare as much as possible. Now you're ready to check the following: Focus: It is important that the electron gun be focused in the center of the screen and near the corners. The corner areas are typically problematic. Look at bright text on a dark background in the center, and in the corners of the screen. Letters should be quite legible, and pixels shouldn't bleed into each other at the screen edges. Bill Nott suggests looking at lower case e's and m's to see if they're readable everywhere. Convergence: Look closely at white lines on a black background. If the lines are white along the edges, convergence is good. If, however, a band of another colour is visible along the line, then colour reproduction of small objects such as characters or lines may be poor. Even if color banding is present, the monitor may still be within the manufacturer's specification. If you can see distinct differently colored lines, chances are the monitor does not meet the specification, but color fringing, while possibly considered objectionable, is likely to be present in almost every monitor built. Pincushioning: Hold something straight (like the edge of a piece of paper) up to the edge of the screen image while viewing the display straight on, from a typical viewing distance. If the image edges bow away from the straight edge, the monitor is exhibiting pincushioning or barreling. Barreling occurs when too much pincushion correction is applied, such that the display bulges outward. Some monitors provide a pincushioning adjustment, but if one is unavailable and pincushioning is severe, significant geometric distortion is likely. Check the pincushioning for different screen addressabilities/refresh rates, as it may vary. Geometric Distortion: Move an object of consistent size ( a window works well) around the screen and measure its height and width with a ruler. Significant variations in the size at different locations indicate geometric distortions that may not be correctable. Colour Purity: Display pure red, green and blue and for each look for colour inconsistencies in the display that may indicate poor colour reproduction. [From: Sam Goldwasser (sam@stdavids.picker.com)] White Purity: Display a totally white screen. The brightness should be reasonable uniform and there should be no objectionably obvious coloured or tinted splotches. Color Bleeding: Display bright primary colored object - red, green, and blue. There should be no colored trails off to the right of the bright areas. Moire: This will depend on resolution and size. There should be no objectionable contour lines visible in the background or smooth areas of the image. [From: Andy Laberge (tic-toc@wolfe.net) and Michael Scott (scott@bme.ri.ccf.org)] Overall Impression: Is the image clear, bright and sharp? Remember that you will be looking at the monitor for hours at a time, and that a minor flicker may become irritating over time. When possible, look at the specific monitor you want to buy, as each monitor has undergone a calibration procedure and some may be better than others - even of the same model. This is one big advantage that local stores have over mail order companies - you can look at the monitor before paying. Failure Rate: Inquire about failure and repair rates for each model. Sometimes retailers stop carrying products because of high returns. How long has the manufacturer been in business? Do they have a good reputation for reliability and performance? Will the retailer deal with any warranty claims, or do you have to go directly to the manufacturer? Will the manufacturer supply parts and schematics for your monitor in future? You may not be doing the work yourself, but a monitor repair technician may need these sometime after the warranty period expires. Return to Table of Contents Q) What pixel addressabilities are best for my monitor? There is no right answer to this question, because it is subjective. However, my recommendations are: Monitor Size Aspect 14" 15" 17" 20" Ratio Screen Addressabilities 640x480 4:3 O O G G 800x600 4:3 A O O G 1024x768 4:3 NR A O O 1280x960 4:3 NR NR A O 1280x1024 5:4 NR NR A O 1600x1200 4:3 -- -- NR A 1600x1280 5:4 -- -- NR A O - Optimal G - Pixels are large enough to appear grainy A - Acceptable NR - Not Recommended - unless you like looking through a magnifying glass Keep in mind that the aspect ratio can be important. Standard televisions and computer monitors are designed to work with a 4:3 aspect ratio. If you use a 5:4 aspect ratio on a monitor with a 4:3, your screen image will be compressed vertically, making circles appear as ellipses. The error associated with using a pixel addressability aspect ratio of 5:4 with a monitor os 4:3 is about 6%. Return to Table of Contents ********************************************************************** END of comp.sys.ibm.pc.hardware.video FAQ - Part 1/4 **********************************************************************
Go Back Go Forward Top of Document Home