Display Technology Shoot-Out

Comparing CRT, LCD, Plasma and DLP Displays

 

Dr. Raymond M. Soneira

President, DisplayMate Technologies Corp.

Copyright © 1990-2005 by DisplayMate Technologies Corporation. All Rights Reserved.
This article, or any part thereof, may not be copied, reproduced, mirrored, distributed or incorporated
into any other work without the prior written permission of DisplayMate Technologies Corporation

 

Part IV  –  Display Technology Assessments

 

Article Links:  Overview  Part I  Part II  Part III  Part IV

Printing: If your browser is improperly printing some pages with text cutoff

 on the right edge then either print in Landscape mode or reduce the font

size (View Menu - Text Size) and margins (File Menu - Page Setup).

 

 

Introduction

This is Part IV of an article series describing an in-depth comparison between CRT, LCD, Plasma and DLP display technologies in order to analyze the relative strengths and weaknesses of each. In Part I we measured, analyzed and compared primary specs like Black-Level, Color Temperature, Peak Brightness, Dynamic Range, and Contrast for each display technology. Click this link to read Part I. In Part II we continued with Gray-Scale, Gamma, Primary Chromaticities and Color Gamut to see how they all affect color and gray-scale accuracy. Click this link to read Part II. In Part III we examined the complex world of Display Artifacts and how they affect Image Quality. Click this link to read Part III. Here in Part IV we'll continue with a discussion of specialized artifacts that apply to only one particular display technology and then analyze and assess each display technology in detail and tie together all of the results from Parts I to IV. First we’ll consider some special display technology issues like Display Aging, Ambient Light, Viewing Angles and Direct-View versus Rear Projection. Then we’ll discuss the viewing tests followed by the individual technology assessments, with breakdowns into many categories.

 

 

Special Issues

In previous sections of the article we’ve discussed the photometry, colorimetry and artifacts for the different display technologies. Here we will address some other common issues that are relevant to all of the technologies.

 

Display Aging

One topic that has turned into a hot marketing issue is display aging. While this is definitely a crucial issue for displays used in commercial applications that run continuously 24 hours per day (often with images that don’t change much over time), it’s really not a major issue for the current generation of displays in normal consumer use. However, the public’s perception of the aging issue has been blown way out of proportion, particularly in the case of Plasma displays, where many people are afraid to buy them for this very reason. None-the-less all displays do age so it’s important to understand the issues and specifications involved. Displays age primarily based on the total accumulated hours of use. If you run a display an average of 4 hours per day that amounts to about 1,500 hours per year and 8 hours per day amounts to about 3,000 hours per year.

 

Phosphors are the source of light in CRT, LCD and Plasma displays. The primary issue for these displays is phosphor aging. While you may hear talk about phosphor burn, the phosphors used in modern displays generally don’t burn (which is actual heat damage to the phosphor that can be seen as a discoloration when the display is turned off). All phosphors lose brightness based on cumulative use, which results from Coulomb Aging (due to an electron beam current) or Solarization (due to ultraviolet radiation). The greater the electron current or ultraviolet radiation the faster the aging progresses. The aging specification is generally in hours of use up to when the brightness (luminance) falls to 50 percent of its original value, which is referred to as the Lifetime.

 

The backlight in LCDs is made up of a varying number of fluorescent lamps that have a phosphor coating. Most LCD displays have a Brightness Control that varies the brightness of the lamps, which affects the rate of aging and therefore the time to 50 percent luminance. You’ll see a wide range of lifetime ratings for LCDs, typically from 25,000 to 60,000 hours. That should be 8 to 20 years of use at 8 hours per day before the brightness falls by 50 percent. However, there is a crucial yet subtle and often obscure or unreported factor in these lifetimes: it’s the setting of the backlight intensity for the specified lifetime. Often it’s not for the maximum backlight intensity, but rather for some lower value that ages more slowly, in which case the lifetime will be less and the display will age faster than you anticipate. Be sure to find that out what brightness value the lifetime specification refers to. For LCDs the lamps should all age uniformly and, in principle, they can be replaced when they get too dim. Unless it’s a really expensive display you probably won’t find that worthwhile.

 

For CRT and Plasma displays aging is more complicated because the phosphor aging can be non-uniform over the screen. The biggest issue for home theater users is caused by Letterboxing (watching wide content on a narrower display) with reduced aging at the top and bottom of the screen or Pillarboxing (watching narrow content on a wider screen) with reduced aging on the left and right sides of the screen. Some displays and signal sources introduce gray bars to help balance this form of uneven aging, but this method can be quite distracting, and it may be unnecessary. If you start to detect uneven aging many Plasma displays have user utilities that can help reduce it by running negative images as part of a maintenance function. To test for uneven aging put up uniform full-screen test patterns in white, red, green and blue at 75 percent signal intensity and look for any variations in brightness over the screen.

 

Another form of aging is due to repetitive content like video games or station logos. This is the origin of the term image “burn-in” where (negative) ghost images of the repetitive content appear superimposed upon whatever image is on the screen. Many Plasma displays deal with this potential problem by introducing a very slow orbiting motion to the image on the screen. Negative images can also be used to rebalance the aging if it becomes a problem. Another issue is that the red, green and blue phosphors can age at different rates based on their physical properties and your viewing habits. (If you watch a lot of golf the green phosphor will age faster.) In critical calibrated imaging applications this actually requires the regular replacement of displays, but for most people all of the uneven aging tends to average out over time and it isn’t particularly noticeable. (Note that on some Plasma displays temporary latent images may appear due to a charge build up within the screen, but they typically disappear after a short time, so they are not a form of aging.)

 

We have been describing all of the variations on CRT and Plasma aging; now it’s time to get a handle on the time scales for aging to 50 percent luminance. For CRTs it can vary quite a bit but 20,000 hours is a typical value (this includes the aging of the cathode, which is the “C” in CRT). That’s almost 7 years at 8 hours per day. For the current generation of Plasma displays the spec is typically 60,000 hours, which is about 20 years at 8 hours per day. However, like LCDs, there is a crucial yet subtle and often obscure or unreported factor in these lifetimes: it’s the Average Picture Level APL for those published specifications. For computer CRTs it’s generally a 100 percent APL, but for Plasma displays in video applications it’s a much lower 15 to 25 percent APL, which is the average for most video content. So if you’re watching typical video then it will take 20 years to reach 50 percent luminance (but if your APL is higher it will be proportionally less). While you’re not likely to hold onto a display for nearly that long it’s a tremendous safety net for all of the potential uneven aging that we discussed above, making it much less likely that aging will be a factor during the time you own the display.

 

Projectors are the easiest aging category to evaluate. They have a replaceable bulb with a specified number of hours to 50 percent luminance (which may be different from the replacement time). Some projectors have a high and low lamp setting, which will result in two specified lifetimes. There are many other subtle factors in display and projector aging but they are generally of concern only in commercial signage applications.

 

Ambient Light

While our discussions have dealt primarily with displays that are viewed in a perfectly dark environment, most displays operate with varying degrees of ambient light (often with significant differences between daytime and nighttime). Under these circumstances, the measured screen black-levels will increase, often substantially. The display’s Dynamic Range and Contrast Ratio will then decrease by the same factor. One topic that doesn’t get all of the attention it deserves is the technology used to reduce the ambient light reflected by a display.

 

Virtually all direct-view and rear projection displays include one or more layers of screen treatments that are designed to reduce reflections and glare. The most common method is to add a light absorbing layer on the screen that transmits, for example, only 50 percent of the light. While this cuts the display’s luminance by a factor of two, it cuts down the ambient light reflections by a factor of four because they have to travel through the absorbing layer twice if they are reflected back towards the viewer. On the other hand light originating within the display travels through the layer only once. As a result the Dynamic Range (full field contrast) improves by a factor of two if the ambient light reflections are greater than the display’s own black-level (which is often the case). If the room is pitch black then there is no ambient light; in that case the Dynamic Range does not change because both the peak brightness and the display’s own black-level are reduced by the same factor. Still, in our example, the hard earned peak brightness is always reduced by a factor 2, which might seem to be an incredible waste. But the payoff in both cases is a lower black-level, which is even more precious than peak brightness. (Note that if the peak brightness is too high for normal viewing and you reduce it by using the Contrast Control, the black-level generally remains unchanged, so the Dynamic Range (full field contrast) will decrease from the manufacturer’s published specification. So it’s generally better to find a display with an appropriate peak brightness, particularly if it’s obtained with an appropriate absorbing layer.)

 

Another common screen treatment is the addition of a matte surface finish to the front of the screen (for example, by etching a glass screen). This cuts down on specular (mirror-like) reflections from the normally polished screen surfaces. It also causes a small loss of image sharpness due to the additional light scatter. A much more effective and expensive (and less common) approach is to use a multi-layer optical coating like those on high quality camera lenses.

 

Screen treatments are almost always separate from the display device so they will vary significantly from model to model. That’s why manufacturers should be bragging about their special anti-reflection and anti-glare treatments and contrast enhancement factors. There are some inherent differences between display technologies: the phosphors used in CRT and Plasma displays are highly reflective so they need to have a good absorbing layer. (CRTs almost always have the absorbing glass built into the faceplate.) LCDs have an advantage because their polarizers and color filters automatically reduce ambient light reflections, however many LCDs still supplement this with an additional 50 percent light absorbing layer in order to further improve ambient light contrast. For rear projection the treatments are part of a complex and expensive multi-layer screen. The cost of a rear projection screen can vary from $50 to $1000 so expect to see major differences in performance based on how much the manufacturer decided to spend on the screen. (For front projection all of the ambient light treatments are part of a separate projection screen.)

 

A good way to qualitatively compare different displays and display technologies for ambient light effects is to shine a bright flashlight at the screen from a few feet away when the display or projector is turned off. The darker the beam looks on the screen the better.

 

There is a side effect to the contrast enhancing absorbing layer: it reduces the range of viewing angles for the display because light going through the absorbing layer at an angle travels through a greater thickness of material than when going in the normal face-on direction, so it experiences greater absorption. As a result the brightness (luminance) decreases with viewing angle. We consider this topic next.

 

Viewing Angles

Another topic that is fraught with confusion is the viewing angle specifications. Different technologies use different ad hoc “standards” and some of them seem to be designed so as to generate published specifications with values close to the ideal 180 degrees, which is the maximum possible viewing angle.

 

A perfectly diffusing screen (called a Lambertian source) will disperse light so that the luminance doesn’t vary with viewing angle. This is considered the theoretically ideal distribution and direct-view CRT and Plasma displays come very close to producing it. As discussed above, the contrast enhancing absorption layer increases the attenuation as the viewing angle increases, so CRT and Plasma displays will show some luminance variation with angle, but they still produce the widest light distribution that you’re likely to see in a display. The viewing angle that’s quoted for Plasma display specifications is the angle where the luminance drops to 33 percent of the face-on value.

 

One problem with a Lambertian light distribution is that a lot of light is wasted in directions where no one is likely to see it, like the ceiling, for example. So almost all front and rear projection screens intentionally concentrate the light in the directions where there is likely to be an audience. By redistributing the light the image appears brighter than it would with a uniform distribution. This increase is called the Gain of the screen and the higher the Gain the narrower the light distribution. For screens the viewing angle is defined as the angle where the gain or luminance drop to 50 percent of peak, and it’s often different for the vertical and horizontal directions. Note that this definition is different than for Plasmas, so the viewing angles for each cannot be compared.

 

LCDs generally produce the narrowest viewing angles of all the display technologies and much of their R&D is directed towards improving this spec, along with the response time. Many of the major manufacturers have their own proprietary technology for increasing the viewing angles so you’ll see acronyms like IPS (In Plane Switching, Hitachi and NEC), MVA (Multi-domain Vertical Alignment, Fujitsu), PVA (Patterned Vertical Alignment, Samsung), and ASV (Advanced Super View, Sharp) for many panels. The problem is that it’s very hard to compare the true viewing angles for these technologies because for LCDs the viewing angle is currently defined as the angle where the full-on and full-off contrast (Dynamic Range) decreases to a value of 10 (yes, ten, it’s not a typo), which is an incredibly low value that may be meaningful for LCD watches but not for imaging displays. As a result, for computer and video direct-view LCDs you’ll generally see viewing angle specs in the narrow range of 170 to 178 degrees, which isn’t particularly helpful since they don’t produce a satisfactory image at those angles. The best thing to do under these circumstances is to do a visual check yourself. Put up a high quality still photographic image that has a wide range of intensities and colors, then shift your viewing position and decide whether it’s an acceptable variation in image quality for you.

 

Hopefully the entire display industry will come up with a uniform and useful definition of viewing angle in the near future. One definition is the angle where the luminance or contrast decrease to 50 percent of maximum. Another is the percent of maximum luminance or contrast at 45 degrees. Either one or both would be very useful.

 

Direct-View versus Rear Projection

One important but subtle form of competition for the display technologies is between direct-view and rear projection. That’s where most of the market share and therefore money lies. Front projection is a much smaller high-end market, but it shares technology with the rear projection units, so it has an important stake in the outcome. CRT and LCD technology can work with either method. On the other hand Plasma is limited to direct-view while DLP and LCoS are limited to projection because the devices are small microchips.

 

Let’s compare the relative advantages and disadvantages of each: the major marketing advantage for direct-view LCD and Plasma displays is that they are very thin, typically 3 to 5 inches deep, so they have a much greater set of placement options and can be hung on a wall or conveniently put on top of a piece of furniture. On the other hand, rear projection units are typically 13 to 25 inches deep and so it’s much harder to find an appropriate location for them. The extra depth is required by the optical system in order to project the image from inside the unit. In a recent major development InFocus has developed a proprietary rear projection “Light Engine” that produces a 61 inch screen size in a cabinet that is only 6.5 inches deep, so its placement options are similar to LCD and Plasma. It’s likely to be a major factor in the balance of power between the display technologies.

 

In terms of image quality direct-view is generally sharper because the pixels are generated right at the screen, while the optics and screen reduce sharpness in projection units (more so in rear projection units because they typically use a multi-layer transmissive screen rather than the reflective screens used in front projection). On the other hand, direct-view displays show much greater pixelation because the red, green and blue elements are tiled side-by-side on the screen. So, for example, when only a single primary color is on, the fill factor can’t be any higher than 33 percent because the other two primary colors are black. This means that the Screen Door Effect is much greater for direct-view than front or rear projection. This doesn’t apply to direct-view CRTs because the phosphor elements are generally much smaller than the pixel size. In a similar fashion, as the screen resolution increases the visibility of the Screen Door Effect decreases because the eye is unable to resolve the pixel structure at normal viewing distances.

 

 

Display Assessments

The assessments for each individual display technology include highlights and summaries of the most important issues that have been discussed in Parts I to IV. Please refer back to the appropriate sections and articles for details, discussions and background information. The assessment categories include Primary Specs and Measurements (from Parts I and II), Notable Variations (interesting technology from other manufacturers), Recent Developments (new technology announcements), Special Issues (topics and artifacts unique to the technology that have not been discussed elsewhere), Strongest Points, Weakest Points, Other Artifacts (not mentioned under Weakest Points, see Part III), Computer Application Viewing Tests, Video Application Viewing Tests, Future Trends, and Commentary. Note that many of the Primary Specs entries show a Rank of 1 to 4 for each technology based on the lab measurements. This specifies their pecking order, with 1 for the best performance and 4 for the lowest performance. When two Ranks are listed, it means that there was a range of values for one or more of the technologies that affected the ordering.

 

Viewing Tests

The central concept for this article was to carefully set up, test and evaluate all of the display technologies at the same time under identical conditions and procedures, using advanced instrumentation where appropriate. All of the displays were set up side-by-side for simultaneous comparative viewing in a completely dark lab. This allowed us to detect subtle differences between the displays. We used computer and video-based test patterns, plus DVD, television, and computer applications, including a wide selection of test patterns at the HD resolutions of 1920 x 1080 and 1280 x 720 from our own DisplayMate for Windows Multimedia Edition (see www.displaymate.com). See Part I for an in-depth discussion of the test methods and instrumentation that were used.

 

Evaluation DVDs and Photographs

The DVDs that we selected were chosen primarily as diagnostic tools to challenge and stress the display technologies and are different from typical review criteria. The following were the most important evaluation DVDs that we used. We needed one really high quality traditional cinematic source and we chose Seabiscuit. It includes a wide variety of interior, exterior and landscape scenes under widely differing lighting conditions. There are a lot of rich and colorful scenes including many full face close-ups that have lots of subtle facial tones and textures. The Matrix includes a subtle green caste for all scenes that are not part of the “real world.” We looked for the accurate reproduction of this delicate feature (which is particularly sensitive to Gamma) in a wide variety of lighting conditions from the very dark to broad daylight. The Matrix also has a lot of very challenging dark material. The other DVD selections were chosen primarily from older science fiction movies such as 2001: a Space Odyssey, Alien and Blade Runner. They include a lot of dark, high contrast scenes, some with significant grain and noise. Spacecraft scenes generally include very bright objects on very dark near-black backgrounds with very dark large-scale intensity gradients. For photographic stills we used a selection from the 1994 to 2002 editions of the InfoComm Projection ShootOut (see www.infocomm.org), which were originally chosen to challenge projector image quality. The key to using all of this material was the simultaneous viewing and comparison of all the displays against one another and against a reference standard, which was a direct-view professional High Definition CRT studio monitor. By simultaneously comparing the relative image quality degradations between displays we were able to differentiate display performance and artifacts from factors that were due to source image quality, encoding and compression artifacts.

 

CRT Assessment

Our reference CRT was a Sony PVM-20L5, which is a direct-view professional High Definition studio monitor and was chosen as the reference standard for image and picture quality because it delivers virtually perfect performance. At 19 inches, it is a standard size studio monitor designed to fit in equipment consoles and racks. Although it’s much smaller than the other displays, we viewed the monitor from a much closer distance in order to help balance the size differentials.

 

 

 

 

CRT Notable Variations: Consumer direct-view color CRTs currently go up to a 40 inch diagonal size for 4:3 aspect ratio screens and up to a 36 inch diagonal for the wide 16:9 aspect ratio screens ● Most consumer TVs use some digital signal processing, which can add digital artifacts to the analog CRT.

 

CRT Recent Developments: Some CRTs are now coming equipped with digital DVI inputs. Since almost all analog signals are actually generated from digital sources, using DVI can actually improve the CRT’s analog image quality substantially because the digital-to-analog D/A conversion occurs near the end of the signal path instead of at the beginning inside the signal source (such as a DVD player or computer graphics board). As a result there is much less degradation and the manufacturer can fine-tune the D/A to the performance characteristics of the display.

 

CRT Special Issues: Direct-view color CRTs come with either a Shadow Mask or Aperture Grille. The Shadow Mask tubes have a matrix of round phosphor dots and are better for reproducing fine image detail. The Aperture Grille tubes (such as the Sony Trinitron™) have phosphor stripes and are better for reproducing photographic images ● Moiré interference patterns, which are wispy waves that appear superimposed on fine image detail, degrade image quality on direct-view color CRTs. Moirés are intensity beat patterns between the screen phosphor elements and the image pixel structure. They are generally only visible on finely focused high resolution displays ● Getting the red, green and blue primary color pixels to line up accurately on top of one another across the entire screen is called color registration or convergence. For CRTs it’s very hard to achieve without advanced adjustment circuitry because minor magnetic anomalies affect each of the beams in a slightly different way. Poor convergence or registration leads to color fringing and to a loss of sharpness ● All CRTs are susceptible to magnetic interference and distortion from the earth’s magnetic field, and from other nearby CRTs, transformers, audio speakers, building steel and furniture, etc. All of these can affect color registration, uniformity and purity, Moiré patterns, and geometric distortion.

 

CRT Strongest Points: Best black-level and Dynamic Range of all the display technologies ● Highest color and gray-scale accuracy ● Most accurate Gamma ● Perfectly smooth gray-scale with no false contouring ● Excellent accuracy and low noise at the dark-end of the gray-scale ● Supports a wide range of resolutions ● Image rescaling not necessary ● No motion artifacts ● Widest viewing angles ● Least artifacts of all the display technologies ● Gaussian beam profile produces a very smooth image.

 

CRT Weakest Points: Gaussian beam profile reduces image sharpness ● Lowest contrast for fine text and graphics ● May produce some image flicker for refresh rates below 75 Hz, particularly with interlacing, which may result in visual fatigue for some people ● Largest direct-view screen size is only 40 inches ● Analog adjustments and calibration are more complex than other display technologies ● Drifts more over time than other display technologies, both short-term (hours) and long-term (weeks). Periodic adjustments and recalibrations are generally necessary for critical applications ● Lowest peak brightness of all the display technologies ● Strong internal reflections within the faceplate ● Bulky and heavy.

 

Other CRT Artifacts: Imperfect color registration reduces sharpness and produces color fringe artifacts ● Moiré interference patterns for direct-view color displays ● Variations in gain and frequency response between red, green and blue channels lead to color artifacts ● Susceptible to geometric image distortion ● Imperfect focus reduces sharpness ● Raster line structure may be visible, particularly at lower resolutions ● High voltage screen regulation issues can cause geometric distortion and gray-scale shifts (not apparent in this high-end Sony model).

 

CRT Computer Application Viewing Tests: Although this particular model is not intended as a computer monitor it performed very well when we supplied a component video signal from an ATI Radeon computer graphics board running Windows XP at 1920 x 1080i and 1280 x 720p. DisplayMate test patterns were reproduced extremely well. With some images and test patterns flicker was apparent at 1080i. The InfoComm ShootOut photographic images were rendered beautifully.

 

CRT Video Application Viewing Tests: For video the picture quality was absolutely stunning. The lowest black-levels, the fewest artifacts, the lowest picture noise, the best color and gray-scale make the CRT the clear winner in image quality (true even if we hadn’t appointed it the reference standard). Very bright and very dark scenes were rendered beautifully.

 

CRT Future Trends: At the high-end there will be fewer and fewer choices for CRT direct-view monitors, front projectors and rear projection units as they will become specialty items for purists and collectors – just like their vacuum tube amplifier cousins. High-end direct-view and projection CRTs should survive in the long-term. Screen size and image quality are likely to continue to increase, but not as fast as their prices. Rear projection CRT units are no match for the image quality of rear projection LCD, DLP and LCoS technologies so they are likely to disappear first, once their price advantage erodes.

 

CRT Commentary: This is impressive performance for a 75 year old technology that produces an image with just a single moving pixel. If overall image and picture quality is your absolute criterion and you can live with a 36 inch maximum widescreen size then the direct-view CRT remains the undisputed king of displays, especially for video. It’s still very good for computer applications, but the sharpness and contrast aren’t as good as the flat panels operating at their native resolution. In spite of this excellent assessment, the CRT is an endangered species. While shipments of inexpensive CRT displays are still increasing the sales of high-end units are falling rapidly because fewer people are willing to pay a premium price for a CRT. It’s the cheapest and lowest performance CRTs that are surviving due to their price advantage over the flat panels. CRTs will be collector’s items soon. Buy now and make a killing on eBay in the future.

 

 

LCD Assessment

Our LCD selection was the NEC LCD4000, which at the time of our testing was the world’s largest production direct-view LCD. It’s marketed as a commercial computer display, but it's an outstanding large-screen LCD panel and will perform extremely well with video when interfaced with the appropriate front-end electronics. This panel had the finest LCD performance characteristics that we had measured up to that time. LCD technology has been evolving rapidly, so the size, resolution, brightness, contrast, viewing angles and response times are improving with each new generation. By early 2005 NEC-Mitsubishi is expected to be introducing displays in this product line up to 55+ inches with resolutions up to 1920 x 1080p.

 

 

 

 

LCD Notable Variations and Recent Developments: The largest LCD panel currently shipping is 46 inches with a native resolution of 1920 x 1080p (Samsung). The largest size LCD prototype that has been shown is 65 inches 1920 x 1080p (Sharp). LG.Philips will be shipping a 55 inch 1920 x 1080p LCD panel in the fourth quarter of 2004 ● The highest resolution LCD panel currently available is 3840 x 2400 (a 22 inch computer display) ● Rear projection LCD displays are available up through a 70 inch screen size. Note that projection LCDs use a small poly-silicon based LCD chip rather than a large amorphous silicon based panel.

 

LCD Special issues: For LCDs the native Transfer Characteristic (the brightness or luminance for a given signal voltage applied to the panel) has an irregular “S” shape. This means that the brightness changes slowly with intensity at the ends of the gray-scale near black or peak white but changes rapidly in the middle of the gray-scale. (The graph looks like a stylized S with long legs that turn almost horizontal near the top and bottom.) The signal processing electronics has to reconfigure this behavior into a straight logarithmic (power-law) gray-scale relationship (see Part II). To do it well takes 12 or more bits of look-up tables and digital-to-analog converters. Many displays are unable to do this and that leads to a compression of the gray-scale at the bright and dark ends and to other irregularities at the dark-end ● Every pixel in an LCD has control electronics on the inside of the panel that produces dark gaps between pixels. This accentuates the appearance of individual pixels and is referred to as the Screen Door Effect because of the similarity to looking through the mesh screen on a storm door. The fill factor or aperture ratio of the light emitting portion of the pixel depends on the particular LCD technology and the pixel pitch and is generally between 50 and 70 percent.

 

LCD Strongest Points: Direct-view LCDs produce exceptionally sharp, high contrast images, including fine text and graphics ● Brightest of all the display technologies ● Highest resolution of all the flat panels (but the LCD4000 is only 1280 x 768) ● LCD panel intensity is controlled by an analog signal, which allows it to produce a smooth intensity-scale that is free of dithering noise and artifacts, especially at the dark-end of the scale (but most current digital signal processing implementations don’t take advantage of this due to insufficient bit-depth) ● Image noise resulting from poor quality video signals was less apparent due to the slower pixel response times ● Low reflection of ambient light due to the panel’s polarizers and color filters ● The thinnest displays available and also not very heavy ● Perfectly quiet for normal viewing (but on some models fans turn on at the brightest backlight settings).

 

LCD Weakest Points: Relatively bright black-level ● Brightness and color saturation generally decrease as the viewing angle increases. Same effect also produces hue errors that increase with viewing angle ● Black-level generally increases with viewing angle ● Slowest response time of all the displays leads to motion flicker, smear and artifacts. Note that most response time specs are misleading (see Part III) ● Lowest pixel fill factor or aperture ratio of the technologies, which often results in visible pixelation and the Screen Door Effect due to visible gaps between pixels. Less noticeable at higher resolutions and greater viewing distances ● Possible uneven light distribution from the backlight ● Fixed native resolution. Rescaling required for other resolution formats.

 

Other LCD Artifacts: S shaped Transfer Characteristic often leads to gray-scale compression and saturation near peak white and a poor quality gray-scale near black (but not seen in the NEC LCD4000) ● Variations in screen brightness and color uniformity and a slightly mottled background ● Variations in the panel’s analog signal response can lead to color tracking errors ● Irregularities at the dark-end of the gray-scale due to insufficient signal processing bit-depth.

 

LCD Computer Application Viewing tests: Image and picture quality was absolutely stunning for computer applications. Images were very sharp and had the highest contrast for fine text and graphics. InfoComm ShootOut photographic images were rendered accurately when viewed face on. As the viewing angle increases brightness and color saturation decrease noticeably. This effect is much less noticeable with business graphics and text.

 

LCD Video Application Viewing Tests: The LCD4000 did a relatively poor job of displaying video, in part because important user and calibration controls were missing from this particular commercial model. The image had a strong blue caste with component video, but in S-Video we were able to squeeze out a tolerably good image. An external video processor would have produced excellent video image quality, inline with the computer application results discussed above.

 

LCD Future Trends and Commentary: LCDs are the dominant flat panel technology for computer applications. There is now a major push to try to accomplish the same thing with video. The critical factors being screen size versus cost. While LCDs are still considerably smaller and more expensive than Plasma displays that gap is closing, with many analysts predicting that both will eventually turn in the LCD’s favor. This is due in large part to the much larger economies of scale for LCD manufacturing, research and development. In order to capture the high-end of the video market LCDs will need to continue improving their black-levels and response times and reducing their viewing angle artifacts, which become more obvious with the spread out audiences that watch the larger screens.

 

 

Plasma Assessment

Our plasma selection was the 61 inch NEC 61XM2, which at the time of our testing was the world’s largest production plasma panel. This same panel is found in many other high-end plasma displays sold by other manufacturers. In fact, since it’s the only panel of this size being manufactured, all 61 inch plasma displays regardless of the brand name use this NEC panel. (The NEC factory was recently purchased by Pioneer and officially changed hands on October 1, 2004. NEC will continue selling the same plasma displays, only now they will be manufactured by Pioneer for NEC.) Most 61 inch panels also use the NEC/Pioneer panel electronics but some use their own proprietary implementations. The latest version of this NEC panel is 61XM3.

 

 

 

 

Plasma Notable Variations: A number of manufacturers (Fujitsu-Hitachi, Panasonic, Samsung) have panels with significantly darker black-levels than the NEC panel, resulting in a Dynamic Range (full field Contrast) manufacturer’s spec of 3000:1 or more, but that applies only to the highest peak intensity at a very low 1 percent APL value. Although black-level is very important the NEC 61XM2 has fewer overall artifacts than other panels, which is why we chose it as the reference plasma display ● Pioneer has a panel that runs at 72 Hz and provides 3:3 Pulldown, which eliminates the judder found in 3:2 Pulldown displays (see Motion Artifacts in Part III).

 

Plasma Recent Developments: The largest shipping Plasma panel is 71 inches by LG with a resolution of 1920 x 1080p. The largest prototype is 80 inches by Samsung (also with 1920 x 1080p) ● Many panels are now advertising a 60,000 hour phosphor lifetime (see Display Aging).

 

Plasma Special issues: The Brightness spec listed by many Plasma manufacturers is the peak brightness (luminance) of the bare panel without the contrast enhancing light absorbing layer, which typically reduces the brightness by about 50 percent. It’s also measured for an Average Picture Level APL of only 1 percent, so the actual viewable peak brightness for typical video with 15 to 25 percent APL may be considerably less. Note that the values listed above under Primary Specs are the ones we measured for the display ● The peak brightness listed for many Plasma displays (500 to 1000 cd/m²) is excessively bright for most subdued ambient light viewing conditions. If you turn down the peak brightness by a factor of 2 or 3 then the Dynamic Range (full field contrast) will be reduced by the same factor. Artifacts will also increase because the display is operating at a lower duty cycle. Finding a display with a lower peak brightness should then deliver better performance (less is more). One way to accomplish this would be with a darker absorbing layer. That would maintain the specified Dynamic Range and deliver a better black-level at the same time ● The power consumption of a Plasma display depends on the Average Picture Level APL of the image because the average current drawn by a pixel depends on its brightness. For low APL the power consumption of a Plasma display can fall by more than 50 percent from its peak value at high APL, and may be less than a comparable size LCD panel (because its power consumption doesn’t vary with APL).

 

Plasma Strongest Points: Direct-view plasma displays produce exceptionally sharp, high contrast images, including fine text and graphics ● Excellent color saturation ● Widest viewing angle of all the flat panels ● Some models have a very dark black-level ● Very fast pixel response time ● Largest direct-view display technology available ● The thinnest displays available.

 

Plasma Weakest Points: Peak brightness and Dynamic Range (full field contrast) decreases substantially with the Average Picture Level (Part I). Generally not an issue for video that has low APLs of 15 to 25 percent ● Spatial and temporal dithering produce noise and false contouring in dark images. These artifacts were more noticeable on Plasma displays than on DLP displays ● Pixelated image with Screen Door Effect due to noticeable gaps between pixels. Less noticeable at higher resolutions and greater viewing distances ● Fixed native resolution. Rescaling required for other resolution formats ● Fan Noise ● Very heavy.

 

Other Plasma Artifacts: Possibility of long-term uneven phosphor aging ● Reflects more ambient light than other technologies ● When viewed from an angle, internal reflections within the panel can produce noticeable ghost images when there is a dark background ● Temporary latent images may appear on some units due to charge build up, but disappear after a short time ● Irregularities at the dark-end of the gray-scale due to insufficient bit-depth in signal processing.

 

Plasma Computer Application Viewing tests: Image and picture quality was excellent for computer applications. The variation of brightness with Average picture Level APL sometimes reduced brightness to well below that of the CRT (see Part I). When there is a switch between images it can take a noticeable fraction of a second for the display to adjust itself to the new APL. The InfoComm ShootOut photographic images were rendered accurately.

 

Plasma Video Application Viewing Tests: For video the picture quality was excellent. The colors were vibrant and saturated and were a good, but not a perfect match to the CRT reference, in part because Gamma and the green primary were quite different from the standard (see Part II). The black-level was occasionally quite noticeable on the NEC panel. Very bright scenes were rendered beautifully, but in dark scenes noise and contouring were quite noticeable (due in part to the low value of Gamma for the display). Performance with poor quality and noisy content was not as good as with the other display technologies.

 

Plasma Future Trends and Commentary: With screen sizes up to 80 inches and aggressive pricing plasmas have captured a significant share of the non-CRT video market. Resolutions were until recently mostly below High Definition, but there are now many panels in the 1365 x 768 through 1920 x 1080 range. Plasmas are a type of digital CRT so it’s not surprising that they have the look and feel of a CRT. Performance has been steadily improving with size and brightness going up and black-levels and artifacts going down. The most important image quality issue is reducing image noise through improved spatial and temporal dithering algorithms and signal processing. The real question is how Plasma displays will hold up to the challenge from direct-view LCD panels.

 

 

DLP Assessment

Our DLP selection was the Optoma RD50, which is the only one of the displays in our roundup that is marketed as a home theater display. This unit is based on the Texas Instruments HD2 DMD 1280 x 720 chip with a 6 segment color wheel. TI has since introduced an HD2+ chip that produces a darker black-level and can be used with a new 7 segment color wheel (see below). The new Optoma model with the HD2+ chip and 7 segment color wheel is called the RD50A, which Optoma says has 20% better contrast than the RD50 and includes our recommended Gamma of 2.20, which will improve its already outstanding image quality. A newly announced Sovereign series offers an even higher performance version of this display with enhanced factory tuning and ISFccc lockable presets for professional calibration.

 

 

 

 

DLP Notable Variations: Although this article has been examining direct-view and rear projection units most DLPs to date are actually in front projectors, however the fastest growth is now in rear projection ● DLP implementations include single chip versions with a color wheel, which is compact and offers perfect color registration, and 3-chip versions, which are considerably more expensive and offer greater brightness, dynamic range and gray-scale bit-depth ● TI has a special high resolution 2048 x 1080 DMD chip that is used in 3-chip digital cinema movie theater projectors ● Color wheels come with a varying number of segments: 3 (RGB), 4 (RGB and White for maximizing brightness at the expense of some color saturation, for computer applications only), 6 (two sets of RGB) and 7 (two sets of RGB and dark green). The color wheels spin at 7,200 (called 4X), 9,600 (called 5X) and 10,800 RPM (called 6X). The faster the wheel and the greater the number of segments the less likely that rainbow artifacts will be seen (see below).

 

DLP Recent Developments: TI recently introduced an enhanced version of their HD2 1280 x 720 chip, called HD2+, which produces a darker black-level, primarily by reducing the gap between mirrors and the dimple where each mirror is attached to its post. TI calls this enhancement DarkChip2™● A new 7 segment color wheel adds a dark green segment that improves performance at the dark end of the gray-scale, reducing contouring and dithering noise by effectively providing a 10-bit intensity-scale at the dark-end ● TI’s SmoothPicture™ technology uses a time-varying mirror actuator to shift the image by half a pixel in order to give the image a smoother appearance ● TI’s new HD3 chip has a 640 x 720 matrix of mirrors that works with TI’s SmoothPicture mirror actuator to produce 1280 x 720 addressable pixels onscreen. The mirrors are oriented at 45 degrees in a diamond configuration in order to work with SmoothPicture to eliminate all visible pixel structure without sacrificing resolution. Its Dynamic Range (full field contrast) is lower than the HD2+ because of its smaller size ● TI also has a new 1400 x 1050 DMD chip, which has a 4:3 aspect ratio and is designed primarily for the computer display market.

 

DLP Special issues: Most DLP projectors use only a single DMD chip together with a high-speed rotating color wheel in order to sequentially generate the primary colors needed to produce a full color picture. A similar color wheel concept was used for the first color television broadcasts in 1951 and for the color television broadcasts from the Moon in 1971. The color wheel offers a number of major advantages: much lower projector cost and size and perfect color registration. The color wheel also has some disadvantages: lower light efficiency because only one primary color is in use at a time so light for the other two is wasted, a reduced number of digital gray-scale levels because the Pulse Width Modulation cycle is time-shared by all three primary colors, and the most curious effect of all are rainbow artifacts that are occasionally seen by some people.

 

The rainbows arise because the red, green and blue primary color images are drawn in sequence at slightly different times. If there is any rapid eye or head motion the color sequences will appear in slightly different locations on the retina. That produces a temporarily mis-registered triple image, which is ordinarily not very noticeable in most photographic style images. However, when the image contains bright-white compact objects on a dark background it then appears as a red, green and blue triplet that looks like a prism or rainbow image of the object.

 

Most people generally aren’t aware of these rainbow artifacts, but I believe that most people will see them under the circumstances mentioned above when viewing the screen up close in a completely dark room, which seems to increase the incidence of rapid eye movements because there aren’t any visual points of reference in the dark. I see rainbows almost constantly when I’m working with test patterns up close to the screen and in the dark, but I see them only rarely when I’m viewing normal video content nine feet back in a dimly lit room. Spacecraft scenes in science fiction movies (like 2001: A Space Odyssey) are the “best” place to determine if you may be sensitive to rainbows. If you are, try a DLP projector with the highest speed color wheel available, which is currently 10,800 RPM.

 

DLP Strongest Points: Darkest black-level and highest Dynamic Range of all the flat-panels ● Closest match to CRT Gamma and primary colors ● Perfect color registration for units with a color wheel ● High pixel fill factor of 90 percent produces a smooth yet sharp image with no apparent pixelation except close up to the screen ● Pixel intensities generated by the digital DMD chip are digitally precise, stable and reproducible ● Very fast pixel response times and few motion artifacts ● Native ATSC Mode 1280 x 720 for the HD2 chips means some HD content doesn’t require rescaling and also allows scaling by other video components that can generate 1280 x 720 ● Very little aging effects other than lamp dimming and replacement.

 

DLP Weakest Points: Spatial and temporal dithering produce some noise and false contouring in dark images ● Color wheel rainbow artifacts ● Possible visual fatigue due to temporal dithering and rainbow artifacts. Some people report significant discomfort but most people don’t appear to be affected ● Direct-view not available – projection only ● Fixed native resolution. Rescaling required for other resolution formats ● Noise from the color wheel and cooling fans.

 

Other DLP Artifacts: Pixels are a bit softer than direct-view displays due to the rear projection optics and screen ● Irregularities and dithering noise at the dark-end of the gray-scale due to insufficient bit-depth in signal processing ● Intentional variation in brightness with viewing angle produced by the rear projection screen so as to maximize the luminance at normal viewing angles.

 

DLP Computer Application Viewing tests: Image and picture quality was excellent for computer applications. DisplayMate test patterns and InfoComm ShootOut photographic images were rendered accurately. Pixels are a bit softer than the direct-view flat panels because of the rear projection optics and screen. Contrast for fine text and graphics was also lower than the direct-view flat panels for the same reason. A slight overscan (when the image is larger than the screen size) resulted in the loss of 1 percent of the image pixels on each side, which can be a problem for some computer applications. (Note that the Optoma’s 1 percent overscan is the smallest of any rear projection display.)

 

DLP Video Application Viewing Tests: For video the picture quality was outstanding, with an excellent match to the reference CRT monitor, but on a much larger screen, an impressive achievement. Dithering noise was occasionally noticeable in dark scenes. These effects should be much less noticeable on the new higher resolution displays. As with most rear projection units there was a significant variation in brightness with viewing angle, which is intentionally introduced by the projection screen in order to maximize the luminance at normal audience viewing angles. However, there is no variation of Dynamic Range, hue or saturation with viewing angle.

 

DLP Future Trends: The new DLP 1920 x 1080p high resolution rear projection displays should begin appearing in early 2005. They use TI’s new xHD3 chip, which has a 960 x 1080 matrix of mirrors that works together with TI’s SmoothPicture moving mirror actuator to produce 1920 x 1080 addressable pixels onscreen. The mirrors are oriented at 45 degrees in a diamond configuration in order to work with SmoothPicture to eliminate all visible pixel structure without sacrificing resolution. (The xHD3 is simply a higher resolution version of the existing 640 x 720 HD3 chip mentioned above.) Note that the xHD3 chips will be available only for rear projection units because they represent a much higher volume market than front projectors, so TI has concentrated its system engineering and development efforts for that market. Hopefully sometime soon TI will announce a 1920 x 1080p product for front projection ● Faster 7 and 8 segment color wheels will be available ● New optics and electronics configurations will result in less expensive 3-chip DLP projectors ● TI’s soon to be introduced DynamicBlack™ technology will improve the darkest black-levels and the Dynamic Range (peak full field contrast) by as much as a factor of 5 by modulating the output light intensity on a frame-by-frame basis after analyzing the scene content ● A scrolling color wheel made up of red, green and blue Archimedes spirals could significantly reduce rainbow artifacts and improve light efficiency because red, green and blue segments are always active simultaneously. This technology has been around for a while but has yet to be implemented in a consumer product.

 

DLP Commentary: In a span of 8 years since its commercial introduction, DLP has steadily improved performance to the point where it has now become a dominant player in projection technology for both computers and video. High Definition Television is the “killer application” for DLP and the 1280 x 720 chips have already captured a major share of the high-end home theater market by delivering outstanding picture quality. The long and eagerly awaited DLP 1920 x 1080 displays will begin arriving in stores in early 2005, the last of the major technologies to introduce a consumer product at this optimum HD native resolution. This higher resolution will significantly increase sharpness and detail while reducing the visibility of spatial and temporal dithering artifacts. However, the big issue for DLP is not further reducing black-levels or artifacts or the price of 3-chip configurations, but rather packaging the DLP units so they have a form factor and footprint closer to the direct-view LCD and Plasma units. The InFocus Light Engine is a gigantic and exciting step in that direction.

 

 

Closing

In this four-part series we have analyzed and compared the major display technologies that are currently available as either direct-view or rear projection units. Competition between the technologies has always been strong but it has been intensifying with the anticipated high demand for High Definition Television displays.

 

All of the high-end displays that we examined produced excellent image quality. There was no across-the-board single display technology winner because these are complex multi-parameter devices and there are a wide range of applications and personal preferences. Price is also a major factor and there are considerable price differences between the display technologies.

 

If you can live with a bulky unit and a 40 inch or smaller screen size then the CRT is the clear image quality winner. For thin direct-view displays the Plasma is currently better for video but the LCD is better for (non-gaming) computer applications. (For computer gaming stick with a CRT.) For large screen image quality the DLP currently produces the best overall image quality, however, there are new challengers in JVC’s HD-ILA and Sony’s SXRD (see below). LCDs are rapidly approaching a 60 inch screen size in a major challenge to Plasma technology. Rear projection units are also trying to become almost as thin as the direct-view LCD and Plasma units so that they too can be placed almost anywhere, including a wall (a major selling point). The most impressive such technology to date is the Infocus Light Engine, which can produce a 61 inch screen size in a cabinet that is a mere 6.5 inches deep, and it can be hung up on a wall. The battle between direct-view and rear-projection units will be the most interesting of all mainstream developments. At the high-end front projection is in a class by itself, but it shares its technology and future with the high-volume rear projection units. Another big question is how well CRT front projectors will weather the competition over time.

 

Future image quality improvements will come in three ways: (1) correcting and improving display parameters like Gamma and the primary chromaticity coordinates, and settings like color temperature so they are a closer match to the industry standard values. This is by far the easiest, fastest and cheapest way to improve a display’s performance. Many manufacturers have been customizing their Gamma curves in order to give their products a special look. We’ve shown that this just produces artifacts and errors in brightness, contrast, hue and saturation. Similarly, many manufacturers have been enlarging the display color gamut beyond the SMPTE C or ITU-R BT.709 standards, primarily for bragging rights in their promotional materials. Unfortunately all this really accomplishes is to increase color reproduction errors because the primaries are different from those found on professional displays that are used to color balance most source material. While it’s possible to electronically transform these non-standard primaries back into the standard values, that is seldom done and then it also undercuts the very purpose of wide gamut primaries (which are currently only useful for specialized applications).

 

(2) Investing to develop and improve the existing signal processing technologies within a display is another comparatively easy and inexpensive approach to obtain improved display performance. Many of the artifacts that we’ve discussed in Parts III and IV could be reduced or eliminated through improved processing electronics. For example moving up to complete start-to-end 12-bit digital signal processing with a complete set of properly implemented functional controls and Gamma look-up tables would make a big difference in image quality, and the development costs are relatively small compared to the development costs of the display devices themselves. In Part II we suggested that the next generation of display signal processing should be done entirely in luminance and chromaticity coordinates Lu'v' in order to substantially increase color and gray-scale accuracy.

 

(3) Improving the display device is where most of the development efforts and funds generally go because that’s what makes the technology proprietary and determines its fundamental performance capabilities. The R&D is very expensive and the improvements generally come slowly. But what really matters is the total system performance and often the signal processing, display standards and calibration don’t get as much attention as they should from the manufacturers.

 

The current renaissance in display technology is likely to continue into the near future. In the last few years we’ve seen the introduction of several new display technologies and also major improvements in display performance. For example, by early 2005 all of the major display technologies will have displays or projectors that run at 1920 x 1080 or above, and we’re starting to see prototypes for as high as 4096 x 2160. 1920 x 1080 is an important benchmark, because it’s the threshold for very high quality home cinema. Black-levels, contrast ratios, resolution and screen sizes are all improving rapidly as well. In fact, lately they’re improving even faster than Moore’s Law for semiconductors, which is a doubling of performance every two years. (Use total pixels and screen area rather than linear pixels and size in the comparisons.) But this renaissance can’t go on forever: development costs are sky rocketing, limiting the number of new players that can afford to launch new display technologies that can challenge the already high image quality of the existing technologies. So enjoy the ride and excitement while it lasts…

 

 

Acknowledgements

Many people read early versions of the manuscript; special thanks to George H. Claborn, Dr. Edward K. Kelley, Gary Reber and Greg Rogers for many important suggestions. Special thanks to Dr. Edward K. Kelley of the NIST (National Institute of Standards and Technology) for many interesting discussions and for generously sharing his expertise. Many manufacturers provided technical information on their technologies: special thanks to Bruce I. Berkoff (LG.Philips LCD), Wing Chung (Optoma Technology), Todd M. Fender (NEC/Mitsubishi), Alan Keil (Ikegami Electronics), Peter F. van Kessel (Texas Instruments), Tom Kwon (Konica Minolta), Masaaki Nishio (NEC Plasma Display Corp, now with Pioneer), John P. Pytlak (Eastman Kodak), Craig Verbeck (Pixelworks) and Dr. Larry F. Weber (plasma technology). Special thanks to the Konica Minolta Instrument Systems Division for providing editorial loaner instruments whenever and wherever they have been needed and for providing the CS-1000 Spectroradiometer on a long-term loan for this project.

 

 

About the Author

Dr. Raymond Soneira is President of DisplayMate Technologies Corporation of Amherst, New Hampshire. He is a research scientist with a career that spans physics, computer science, and television system design. Dr. Soneira obtained his Ph.D. in Theoretical Physics from Princeton University, spent 5 years as a Long-Term Member of the world famous Institute for Advanced Study in Princeton, another 5 years as a Principal Investigator in the Computer Systems Research Laboratory at AT&T Bell Laboratories, and has also designed, tested, and installed color television broadcast equipment for the CBS Television Network Engineering and Development Department. He has authored over 35 research articles in scientific journals in physics and computer science, including Scientific American. If you have any comments or questions about the article, you can contact him at dtso@displaymate.com.

 

 

Article Links

Series Overview

Part I: The Primary Specs

Part II: Gray-Scale and Color Accuracy

Part III: Display Artifacts and Image Quality

Part IV: Display Technology Assessments

 

 

Copyright © 1990-2005 by DisplayMate Technologies Corporation. All Rights Reserved.
This article, or any part thereof, may not be copied, reproduced, mirrored, distributed or incorporated
into any other work without the prior written permission of DisplayMate Technologies Corporation