LCD Specs: Useless?
Tested contrast ratios rarely conform to vendors' specs.
If you are shopping for an LCD monitor, high contrast ratio is a selling point vendors love to pitch. But if you rely on this specification, you'll have about as much luck picking a winner as you would if you bet on a racehorse because you liked its name.
Contrast ratio is the ratio of a screen's whitest white to its blackest black. A higher number generally means a better image and easier-to-read text. But PC World Test Center evaluations of 15 LCD units showed that in some cases, actual contrast ratios were below the vendors' specifications--by as much as 50 percent (see the chart).
Most companies actually erred in the opposite direction, with a published contrast ratio that fell below our measured result--something that most people probably would not mind. But since shoppers have no way to tell if the manufacturer has overstated or understated contrast ratio, the specification is essentially useless for comparison purposes.
Specsmanship
We found that a major reason for this problem is vague wording in the commonly used VESA (Video Electronics Standards Association) Flat Panel Display Measurement (FPDM) standard that gives vendors considerable latitude in adjusting monitor settings when measuring contrast ratio. But this situation may be addressed soon: At least partly because of PC World's findings, the VESA FPDM committee is already revising the standard to provide more-detailed testing instructions.
Still, testing variations alone can't account for the most dramatic differences between our measurements and the published contrast ratios. Some companies don't even use the VESA spec. Others don't do their own testing.
The PC industry is built on specs, and LCDs aren't the first products to have the usefulness of their numbers called into question (see Up Front). With competition in the LCD market fierce, even a small advantage on paper can mean the difference between making or losing a big sale.
"Some markets buy on specs," explains Todd Fender, senior product manager at NEC-Mitsubishi. "The government, for example: It would buy one monitor over another if its contrast ratio was 600:1 instead of 599:1."
So important are these figures to a company's bottom line that when NEC-Mitsubishi engineers began to suspect other manufacturers were inflating specs, they tested the competition. NEC-Mitsubishi subsequently filed a lawsuit alleging that rival ViewSonic had damaged it by misstating contrast ratios for certain ViewSonic monitors.
The companies recently settled out of court, and both declined to comment on the case. But to get an idea of how widespread the problem of inaccurate specs was, the PC World Test Center ran a group of LCD monitors we'd recently evaluated for our Top 100 charts through a test based upon the VESA standard's instructions.
Wild Results
We developed our tests in collaboration with physicist Raymond Soneira, president of the video diagnostics company DisplayMate Technologies and a well-known expert in monitor testing, whose DisplayMate utility we use in our Top 100 monitor tests.
We went to considerable lengths to ensure a level playing field by subjecting all test units to a precise set of adjustments designed to standardize the LCDs (see "How We Test" in the chart). To ensure accurate measurements, we borrowed a research-class Pritchard PR-880 photometer from Chatsworth, California-based Photo Research, which develops high-precision electro-optical equipment.
Our findings? Thirteen test units posted numbers significantly different (by 10 percent or more) from their specifications, and only two LCDs--one from ViewSonic, the other from Eizo Nanao--landed within 5 percent of the stated number, close enough to be considered accurate. The good news: Nine offered a meaningfully better contrast ratio than listed. The bad news: Four were markedly worse.
Of the vendors for which we found results significantly below their reported contrast ratios, CTX was the worst offender with its S730, which achieved a contrast ratio of 252:1, 50 percent below its stated 500:1 specification. On the other hand, our test LCD from Dell notched an impressive 892:1 contrast ratio, a whopping 78 percent higher than its stated 500:1 rating.
Note that neither our measured contrast ratio nor the percentage by which a contrast ratio missed its published specification correlated with the display's overall quality as reflected by its PC World star rating--which is based not only on the LCD's image quality but also on features, controls, warranty, and other factors that are important to a buying decision.
For example, LG's monitor earned four stars even though our tests showed its 405:1 contrast ratio was 10 percent less than the published 450:1. Dell's 1702FP, with its 892:1 contrast ratio, still managed only three and a half stars.
Why did so many published contrast ratios differ from our measurements? Even if you believe the vendors overstate their results to try to make their products look better than they are, that wouldn't explain why so many models tested higher than their spec.
Specs Inspected
One reason is that some vendors don't use the VESA test to measure contrast ratio. The standard is voluntary, and the spec's primary author--Edward F. Kelley, a physicist at the National Institute of Standards and Technology's Flat Panel Display Laboratory--says it's not clear how many vendors have adopted it. (Manufacturers don't have to say which standard they use, a situation VESA is seeking to correct for its own standard.)
For example, engineers at CTX haven't been using the VESA standard (though CTX spokesman Daniel Rhodes says the company is joining VESA and will be using its flat panel standard).
Even when vendors do follow the standard, results may vary due to some ambiguity in its wording, which says that testers should adjust an LCD for its intended use "and not optimize [the display] for each measurement separately."
Accordingly, most vendors adjust LCDs for real-world usage before testing. But DisplayMate's Soneira says that others use the wording as a loophole to achieve higher brightness and contrast ratios. "Some manufacturers turn up all of the controls to their maximums and then take their measurements. An LCD is actually unusable that way," Soneira says.
The VESA standard's principal author agrees. "The philosophy of the VESA FPDM was to set up the display as best possible for the intended task, for example in an office," Kelley says. "We never intended to allow anyone to misadjust the display out of its useful operating range. For anyone to interpret the FPDM any differently from that is a sad, sad mistake."
As a result of PC World's tests, Kelley says VESA will change the standard to precisely specify how an LCD should be adjusted for testing.
While we didn't test the monitors involved in NEC-Mitsubishi's lawsuit against ViewSonic, one of the two ViewSonic LCDs we did test exceeded its published contrast ratio slightly and another fell far below it. ViewSonic executives say that they take contrast ratio measurements seriously and use the full-white, full-black VESA test. They suggest the monitor that underperformed its contrast-ratio spec may have used a substandard LCD panel that slipped through the quality-control process at ViewSonic's panel supplier.
Duane Brozek, director of public relations at ViewSonic, says his company's specs are based on "rigorous testing.... That's not to say that sometimes something won't slip through, and we'll track down what the issue is here."
ViewSonic's practice of buying LCD modules from third-party suppliers, who provide their own specs, is fairly common in the LCD monitor business. But some vendors, including Hitachi and LG Electronics, say they use their suppliers' specs for their finished monitors without testing further--a practice that Soneira and others frown on.
A finished monitor's contrast ratio can vary from that of its LCD module, Soneira explains. "The values measured depend on the electronics, the factory calibration, and the user adjustments."
NEC-Mitsubishi uses the VESA test, but its unit tested much higher than its listed spec. That's intentional, says Richard Atanus, vice president of product development for NEC-Mitsubishi.
"We state that specification as guaranteed," Atanus explains. "There is no discrepancy and no misleading of the customer."
The bottom line for LCD shoppers? You're better off not putting too much faith in the listed numbers (see "Buying Tips: The Eyes Have It When You're LCD Shopping").
Help may be close at hand. When the VESA standard is updated--probably by midyear--vendors who adhere to it will be able to promote their specs as "VESA contrast" and "VESA brightness," which should make them more meaningful for comparisons.
At PC World, however, our monitor ratings depend heavily on "taste tests"--actual evaluations by a jury of human beings who view a variety of text and graphic images on side-by-side displays. We believe that this kind of review is more helpful than one that's based solely on specifications--even those we can measure ourselves.
If you've got your eye on an LCD monitor for which you can't find a review from a trustworthy source, don't be swayed by a sheet of technical specifications: You're better off relying on the evidence of your own eyes.
Test Report: A Study in Contrasts (Chart)
| 17-INCH LCD | Contrast ratio1: Tested | Contrast ratio1: Published | Tested versus published ratio |
Dell UltraSharp 1702FP (http://accessories.us.dell.com/sna/productdetail.aspx?Sku=320-4106&category_id=4009&customer_id=19&c=us&l=en&cs=19)  | 892:1 | 500:1 | 78% better |
Samsung SyncMaster 172B (http://pcworld.pricegrabber.com/search_getprod.php/masterid=609344)  | 590:1 | 350:1 | 69% better |
Planar PX171M (http://www.planar.com/monitors/PX171M.asp)  | 584:1 | 400:1 | 46% better |
| Hyundai ImageQuest Q17 (http://pcworld.pricegrabber.com/search_getprod.php/masterid=704372) | 683:1 | 500:1 | 37% better |
| Hitachi CLM174SXW (http://pcworld.pricegrabber.com/search_getprod.php/masterid=677663) | 525:1 | 400:1 | 31% better |
| AG Neovo F-17 (http://www.agneovo.com/products/F-17.htm) | 496:1 | 400:1 | 24% better |
| BenQ FP767 (http://pcworld.pricegrabber.com/search_getprod.php/masterid=653061) | 610:1 | 500:1 | 22% better |
NEC MultiSync LCD1760V (http://pcworld.pricegrabber.com/search_getprod.php/masterid=671759)  | 545:1 | 450:1 | 21% better |
| Iiyama ProLite AS4314UTG (http://www.iiyama.com/web2002/product/13p_4314.htm) | 550:1 | 500:1 | 10% better |
| ViewSonic VG700 (http://www.pcworld.com/reviews/article/0,aid,106070,00.asp) | 461:1 | 450:1 | 2.4% better |
| Eizo Nanao FlexScan L565 (http://pcworld.pricegrabber.com/search_getprod.php/masterid=614292) | 386:1 | 400:1 | 3.5% worse |
| LG Electronics Flatron L1710B (http://www.lge.com/c_product/catalog/product_list.jsp?ca_code=020103&navigation=1,18,3) | 405:1 | 450:1 2 | 10% worse |
| Hansol H750 (http://www.hansol-us.com/products/products.html) | 247:1 | 350:1 | 29% worse |
| ViewSonic VE175 (http://pcworld.pricegrabber.com/search_getprod.php/masterid=616979) | 352:1 | 500:1 | 30% worse |
| CTX S730 (http://pcworld.pricegrabber.com/search_getprod.php/masterid=687406) | 252:1 | 500:1 | 50% worse |
Lab Notes: What About Brightness?
Brightness or luminance--typically expressed in candelas per square meter (cd/m2)--is at least as important to many LCD buyers as contrast ratio. But in our tests, vendor brightness numbers proved as unreliable as their contrast ratios. More than half the vendors overstated brightness, as measured in our tests using maximized settings (which vendors typically use for luminance tests).
Vendors measure contrast ratio using full-white and full-black screens, and so did we. But few people actually use their monitors this way, and VESA's standard also calls for separate measurements using a checkerboard pattern, which tends to lower the contrast ratio because light bleeds from white boxes to dark boxes. Contrast ratios with our 15-by-15 checkerboard (the most granular our test setup permitted) were slightly lower, but we expect that they would have been lower still had we been able to use smaller boxes. However, even this arrangement is nine times more granular than the 5-by-5 pattern that VESA recommends.
Buying Tips: The Eyes Have It When You're LCD Shopping
When shopping for an LCD, your best bets are to check independent reviews such as PC World's Top 100 (see our Top 10 15-inch LCDs or the 17-inch displays we tested) and trust your own eyes. Here are some buying tips:
- Size: A 15-inch LCD offers about the same amount of viewable screen as a 17-inch CRT, a 17-inch LCD as a 19-inch CRT, and so on. Ross Young, president of the LCD research firm DisplaySearch, says you should expect to pay $209 to $620 for a 15-inch LCD, $314 to $815 for a 17-incher, or $559 to $994 for a 19-inch screen.
- Resolution: LCDs operate best at their native resolution--which is typically 1024 by 768 for 15-inchers, and 1280 by 1024 for 17-inchers. The higher the resolution, the more data the screen can display at once. However, icons and text can look small if a screen's native resolution is unusually high for its size.
- Brightness: For multimedia images in a well-lit room, brighter screens are better--say, 300 to 400 cd/m2 (a measure of luminance; see "Lab Notes: What About Brightness?") compared with the 200 cd/m2 (or so) of most mainstream LCDs. But as PC World's tests also called vendor brightness ratings into question, try to test brightness by looking at a blank white page in a word processor set to fill the page.
- Viewing angle: If you aren't always looking at your monitor head-on, or if you want several people to see the display simultaneously, you'll want to check out the viewing angle. But here, again, don't trust the specs: Move horizontally and vertically to check for subtle color shifts and loss of contrast.
- Response time: This is the time it takes to turn a pixel on or off, or to change its color. If you plan to use your LCD monitor to watch movies or play games, faster times are definitely better; don't settle for a response time of more than 16 milliseconds. However, there's a trade-off: Motion flicker increases on fast-responding LCDs.
- Interface: Most mainstream LCDs have only analog inputs, but for a premium you can find some that also offer a port for digital input, which generally should produce sharper images without requiring adjustment. But using that port also requires a graphics card with a digital-out port and an appropriate cable, and DisplaySearch's Young says most people are happy with analog.