Aris Silzars Display Technology Consulting


Rows and Columns...June 2002

There is a fundamental difference between CRTs and all flat panel displays -- yes, besides the obvious one that no one has yet come up with a commercially successful flat CRT. What I am thinking about is that all flat panels that exist today require row-and-column addressing, while CRTs do not.

Perhaps an imperfect analogy would be to think of CRT addressing as similar to an artist creating a painting, while a flat panel may be more like the weaving of a tapestry. The painter stands some distance back and uses a brush or palette knife with complete freedom of when and where to place the paint onto the canvas. The weaver can also create beautiful images but is constrained by the rows and columns of the threads on the loom. And, I suppose, similarly to a CRT versus a flat panel, the painter has a more difficult challenge if precise location of objects is required compared to the weaver who can count to the appropriate number of rows and columns.

Thus, all flat panels are "fixed-format" displays while the CRT can be addressed in any format up to and even exceeding the resolution capabilities of the beam itself. But, you may ask, "What about the screen? Doesn't it have a pixel and sub-pixel structure?" Yes, but typically that is chosen to be compatible with the resolution capabilities of the writing beam and (other than some care that must be taken with moire effects) does not really restrict how the information is scanned onto the screen. In fact, intentional electronic image manipulation can be introduced to compensate for effects such as keystone distortion in projection systems. Ultimately, the screen resolution is limited only by the manufacturing process selected.

Those of you with extensive CRT backgrounds will, of course, instantly note that a typical television raster scan is not the only way to put information onto a CRT screen. Some displays, such as those for analog oscilloscopes, operate more like a graph being drawn in Cartesian or polar coordinates. Other CRT displays operate in a "stroke writing" mode in which the beam traces out the information just as one would do with a paintbrush or a pen. It was only when flat-panel displays came along that we ended up with either segment addressing or having to use a pre-specified number of rows and columns as the information content exceeded what could be realized by making a connection to each individual pixel.

A few days ago, a colleague asked me why television was developed with a scanning technique that requires blanking and a retrace period. The specific question was, "Why do we write the image only from left to right? Why not scan back and forth with a triangle waveform?" I didn't have a really compelling answer. Do any of you know if there is anything more than the convention of how we read a page of text that caused television to develop as it did? I'm pretty sure that the decision was not made because the flyback provides a convenient way to generate a high voltage. Is there an interesting story here from the early days of television? Maybe some of you can help to educate the rest of us.

Over the years, many have recognized the simplicity and versatility of the "paint-brush" approach to addressing a viewing surface. Thus, there have been many efforts made to reduce or eliminate the depth and volume needed to do this method of addressing. Many electron-beam-bending and beam-channeling schemes have been tried, but none has resulted in a successful product. The other major approach has been to try to use some form of tiling. I suppose we could consider this as dividing up our large canvas into many areas and then having many artists each do a section with smaller paintbrushes. Unfortunately, the electronic results are not too different than what one would expect from a group of artists with differing talents and skills. These tiled displays always end up looking as if -- well, as if they have been tiled. It turns out that our eyes are highly sensitive to boundaries (edges) and repetitive patterns. In a typical television set, the display may be 20% dimmer in the corners than in the center. But, we don't notice it. However, if we take these same television sets and create a video wall with them, they now look like eyeballs staring at us with images that have an obvious hot spot in the center. Also, if we see any kind of an edge transition, or an image mismatch of even a fraction of a percent, we will immediately reject this as a display for anything other than possibly non-critical advertising applications. Thus, while it is perhaps possible to demonstrate tiling in a laboratory environment, so far it has proven to be too difficult and too expensive for commercial implementation.

The lack of success by CRT developers to find an elegant solution to the weight-and- volume problem has encouraged flat-panel manufacturers to enter markets that were previously dominated by CRT-based products. But the challenges for flat-panel technologies are also non-trivial. As panel sizes and resolution increase, row-and-column addressing becomes ever more difficult. For every doubling of linear resolution, the number of drivers increases by a factor of four. Thus, improving yield and reducing manufacturing cost become major driving forces.

Recently, we have seen the introduction of LCDs as large as 40 inches and plasma panels that range in sizes from just over 30 inches to more than 60 inches. The need to keep product costs at a reasonable level is helping to answer the resolution question of "when is it good enough?" Upper-end products are settling in at the 700-line level. This indeed turns out to be "good enough" since it matches the imaging capabilities of most movie films as viewed under even the best of theater settings. Therefore, we can expect that for the next few years we will need to work hard to reduce the manufacturing cost of these million-pixel displays -- or alternatively, two-thousand row-and-column-driver displays. The HDTV standard of 1080 interlaced lines will likely be seen only in theaters showing movies using electronic projectors.

The trend toward larger flat-panel entertainment and business displays will also stimulate continued development of projection technologies. It is interesting that certain projection displays can take advantage of the versatility of CRT-like addressing. Row-and-column addressing is required only if the light engine itself is inherently a fixed-format display, such as a TFT-LC, LCOS, or DLP. Projectors based on CRT-like technologies, MEMS-based image-scanning techniques, and a few other new approaches will be able to take advantage of the greater image-input flexibility.

Large-screen displays for entertainment and business applications -- whether direct view or projection -- will be an exciting area for new display product development over the next decade. For LCDs and plasma displays, the greatest emphasis will be on driving down manufacturing cost. Projection technologies, on the other hand, may have opportunities for new and innovative approaches, and these may also have inherent advantages for lower manufacturing cost. We can expect to see vigorous competition among all approaches, and it is likely that no one approach will dominate in market acceptance.

Is there a large-screen display in your immediate future? If you have made, or are about to make a selection, I would enjoy hearing from you. What did you chose and why? You may contact me by e-mail at, by telephone at 425-557-8850, by FAX at 425-557-8983, or by mail at 22513 SE 47th Place, Sammamish, WA 98075.