computerwriter.com

LCD monitors: Of pixels and pitches

Toronto Star Fast Forward column for Sept 9/98

Back to Computer Wares index

Back to White Pages main article index

© Copyright, Myles White, 1998

Tip of the Week: FAT32 slows Explorer

A number of weeks ago I ran a series of columns under the Back to Basics banner on computer monitors which drew a question from one reader about the physical characteristics of the new "LCD" flat panel displays for desktop computers.

Although I answered part of the reader's question in a subsequent mail column, there were still some details about the technology that I frankly didn't understand, partially because of some funny information I'd received a number of years ago from an engineer who should have known better and partially because the technology has evolved over time. In the meantime, I've been doing more homework, primarily with the aid of Peter Wheeler and Vijay Sharma from ATI Technologies, Lance Bailey at NEC Technologies, James Pyler at IBM, and Max Yonezowa from Panasonic's factory in Japan. All errors below are mine, not theirs.

Desktop flat-panel monitors are invariably referred to in ads as "LCD" monitors. LCD, or liquid crystal display, is an umbrella term that covers two basic technologies that first emerged in the notebook screen market. The first has a variety of names: passive matrix, super-twist nemonic (STN), fast super-twist nemonic (FSTN) and dual super-twist nemonic (DSTN). The second also has a variety of names: thin-film transistor (TFT) and active matrix. Both active and passive matrix technologies share some characteristics.

Each picture element (pixel) on an LCD screen is made up of three crystal "cells" representing red, green, and blue. Light is provided by one of several backlighting schemes, but the essence of the trick is that when power is applied to each cell, it twists to allow light through it. The amount that the crystal can twist determines how much light gets through, depends on the amount of power applied to it, and determines the number of colour shades it can display. If you want to, you could imagine that as your screen image changes, the tiny crystals are writhing like crazy, trying to follow it.

One of the primary differences between active and passive matrix screens is the speed with which each cell gets the power required to change it. In a passive matrix screen, there is a row of transistors across the top of the screen and another down the side. To change a cell in the middle, the transistors corresponding to the exact location at top and side need to send a pulse to the appropriate cell (e.g., 300 columns over and 200 down). If you're changing a 640 by 480 image (three cells each, and two transistors per cell, remember), there will be 1,843,200 possible combinations of signals to be organized. Early passive matrix screens were notable for the lag in screen updates (typically, the cursor would disappear or drag a "ghost" behind it).

To speed up the process, some manufacturers place transistors at top and bottom and each side (Fast STN). Later innovations allowed for a row of transistors across the middle of the screen so that both halves could be updated simultaneously (Dual STN). Nevertheless, while screen updates were faster, there was still a delay, the picture wasn't very bright, and the viewing angle is quite narrow (you'll want to view such a screen head-on, not from the side).

In an active matrix screen, there is a separate transistor for each cell; screen updates are much faster; the screens themselves are brighter and may be viewed from a much wider angle. Today's desktop flat panel LCD displays use active matrix technology, not passive matrix. Why didn't the industry use the more familiar thin-film transistor (TFT) or even "active matrix" label for these monitors so consumers would be less confused? Darned if I know.

The questions that spawned the research actually focussed in on another oddity in LCD technology and the industry's quirky use of some terms led to more confusion. The original question was about a "blurry" TFT screen on a notebook system and the writer wanted to know if there was an equivalent measure of sharpness to the standard monitor's "dot-pitch."

A quicky review: on a standard cathode ray tube (CRT) monitor, the inside of the tube is coated with phosphor dots which light up when excited by focussed beams from the electron guns in the tube. The beams may be focussed through either a shadow mask (with round holes) or an aperture grille mask (rectangular holes). Dot pitch (the measurement for shadow mask monitors) or stripe pitch (for aperture grille models), measures the distance between two phosphor dots of the same colour. The typical dot pitch range is from 0.26 mm (crisp and sharp images) to 0.43 mm or more (grainy indistinct images), while aperture grille monitors typically have a stripe pitch of 0.25 mm. Is there, the reader wondered, a similar measurement for LCD monitors?

Indeed there is and the range is about the same as it is for shadow mask CRT monitors - 0.26 to 0.31 mm, but the portion of the industry that makes these devices has also added another term, "pixel pitch" that can be confusing because some companies use it in place of dot pitch. As far as I can tell, they use it incorrectly, too.

CRT monitors, which tend to have a wider range of resolutions available than LCD models, have no distinct edges between pixels - the phosphor dots that are present are used in whatever combination is necessary, so long as there are at least three dots (one each for red, green and blue) within each pixel. In an LCD monitor, there will be as many pixels, each with its three cells representing red, green and blue, as the highest resolution the system can produce. In other words, if the LCD screen is able to render 1024 by 768, there have to be 786,432 pixels available for it to do so. And there is a space between them, generally ranging from 0.02 mm (20 microns) to 0.2 mm (200 microns). The higher the pixel pitch, the grainier the image will be.

So, how could our reader get a blurry screen? This is more likely to happen when you run an LCD screen at a resolution that is lower than its maximum. Depending on the monitor you own, one of two things can happen if you run it at a lower resolution than its upper limit. It will either use fewer pixels (and leave you with a nice black border around the smaller image) or the manufacturer may have built some math trickery into the monitor's driver to expand the picture to the full screen size. This may cause the image to be less distinct.

Almost home

As a consumer evaluating flat panel displays, the preceding will give you some specifications to check -- dot pitch and pixel pitch - but there's one more item you'll want to assess as well: whether the monitor is fully digital or analog. Fully digital LCD monitors are the newest breed and I'll try to keep the description of the difference as simple as possible. In your computer, the information sent to the graphics controller from the central processor is digital; descriptions of colour values and screen locations are expressed in bits and bytes. The graphics controller has to take the information, then convert it to an analog composite signal which it sends on to a standard monitor. The monitor decodes the composite signal and gives appropriate instructions to its electron guns. In an LCD analog monitor a roughly similar process takes place, except that the information is sent to each transistor for action. We'd have faster graphics and less expensive components if the analog portion of the process were eliminated. And that's approximately what happens with digital LCD screens. You need a special graphics controller and connector to achieve the end results, but in essence, the whole process becomes digital; the graphics controller doesn't require a digital to analog converter and the monitor can dispense with the analog to digital converter it would otherwise require.

In the final analysis, however, selecting a computer monitor should never be done through comparing specifications alone. The old eyeball test should still be the final determining factor.

Tip of the Week

Sleepy Explorer

If you have Windows 95 OSR/2 or newer (i.e., Windows 98) and have applied the new 32-bit file allocation table (FAT32) to your hard drive(s), you may have noticed some performance issues. When you open a folder in Microsoft Explorer on a drive using the FAT32 file system, it may seem to take an unusually long time before the file window is accessible and/or the "Working in Background"pointer may appear for a long time.

This is one aspect of FAT32 that caught Microsoft by surprise, and the programmers haven't found a way to fix it yet. However, the company has suggested some workarounds. The problems occur when the total space used by all entries in a folder exceeds 32 KB in size. The size of any particular entry is determined by the number of characters in the file or directory name, plus the items you see in the "Details" view (file size, file type, date of modification and attributes) and some you don't see (such as the starting cluster of the file on your hard drive). The total space used also includes hidden files and directories. Each character uses one byte. To exceed 32,000 bytes (32 KB) means you'll either have a lot of entries in a particular folder or a lot of entries with very long filenames.

The workarounds suggested include limiting your use of long file names. For instance, just because you can use up to 256 characters (one byte per character) doesn't mean you should use that many (Microsoft suggests keeping under 70 or less). When possible revert to the older 8.3 naming convention (eight characters for the file's name, three for its extension and note that using fewer than eight characters makes no difference -- Windows pads out the missing characters so that each filename occupies a minimum of 12 bytes). Another solution is to use more sub-folders and to move some files to them. You can do this without consequence (except for more clicking to find a file) in folders containing documents you've created, but should avoid arbitrarily moving files from program folders (the program may not work if a file it expects to find in one location is no longer there).

If you're having these problems and you don't have FAT32, it may still be because there are a large number of entries in the folder you're trying to open. According to Microsoft, Windows Explorer parses all the directory entries in the File Allocation Table completely (i.e., looks at them, analyzes them and places them in memory) before displaying the list, instead of displaying each name as it finds them.

Last, but not least, there may be times when your system seems to halt completely before anything happens. You may hear your hard drive spinning up while you wait until the system comes back to life. The cause of this one is probably that you have some type of power management turned on which is putting your hard drive(s) to sleep after a period of inactivity. You can either live with it, or check in both the system BIOS and the Power Management applet in Control Panel to either make the inactivity period longer or disable the feature altogether.

Back to top

Copyright Notice

This document, and all other articles found within the White Pages Web site, are protected by international copyright. All rights are reserved. You may download items for your personal use. You may place a link to the White Pages at your own web site. However, you may not post the articles at your web site, make paper or digital copies for your friends or your class, or quote any part of these documents in any medium whatsoever for any reason unless you ask me first. Okay?

Back to top