Issue No. 03 - May-June (2012 vol. 32)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MCG.2012.56
Determining how computer graphics will evolve requires a good understanding of both software and hardware advances. As a software person, I tend to pay less attention to hardware than I should. I've finally realized that all the software I've written over the past few decades relies on hardware. And how my software behaves changes dramatically when my hardware changes. This issue's column is dedicated to computer hardware advances. Some of it might become reality and further advance our field of computer graphics.
Computer graphics algorithms and hardware have been based on pixel frame buffers and raster scanned display screens since the early 1980s. The dominant technologies before then were based on redrawing all vectors each frame from processor memory on a cathode-ray tube or storage tube. Having been involved in both, I can say that display strategies were clearly much different between them.
A Foldable Display
One of the most fascinating trends over the past five years is the rise of mobile technology. Even as good as displays are on tablets and smart phones, they're still more cumbersome than paper. Paper can be folded, spindled, and mutilated and still provide a viewable surface.
University of Toronto researchers have discovered a better way to make flat-panel displays. The discovery could one day lead to computer screens you roll up like a newspaper and wallpaper that lights your living room. The research team has applied organic light-emitting diodes (OLEDs) to flexible plastic. The discovery means a less costly, more efficient, and environmentally friendly way to build brighter flat-panel displays on a thinner, more durable, and flexible surface.
When cleaning sheets of indium tin oxide, a material used in all flat-panel displays, two PhD students noticed that devices incorporating their cleaned sheets had become much more efficient than expected and used less energy to achieve much higher brightness. Chlorine picked up from the cleaning solvent turned out to be the cause. The discovery is akin to dipping a clean sheet of paper in a highly chlorinated swimming pool. Check the video at http://vimeo.com/32678094.
Another Flexible Display
Rather than using OLEDs, HP Labs researchers are testing a flexible, full-color display that saves power by reflecting ambient light instead of using a backlight. The prototype display's pixels are controlled by fast-switching silicon transistors printed on plastic. If the technology can be commercialized, the display will compete with liquid-crystal screens and other low-power color flexible displays in the works. HP is collaborating with Phicot, a Powerfilm subsidiary ( www.powerfilmsolar.com), which prints high-performance transistors on plastic. HP plans to target both the e-reader and tablet PC markets.
Both the Toronto and HP efforts seek to reduce the power consumption and rigidity inherent in e-ink and LCDs. This should be a fascinating race. And the demand for flexibility is trickling into the general population through an April 2012 USA Today article.
Oily (Not Late) Pixels
One challenge of electronic paper based on e-ink and similar technologies is displaying video and other dynamic content. Erasing and redrawing the screen take too long. Today's systems require backlit displays for liquid crystals to display dynamic content. University of Cincinnati researchers experimentally demonstrated a display consisting of a vertical stack of three layers of oil dyed red, green, and blue. The layers are separated by two layers of water. These five layers, along with a hydrophobic insulating layer and hydrophilic grid, are sandwiched between electrodes. Aligning the layers of colored oil creates separate pixels.
To change the color on the display, low voltage is applied to the water layer next to one of the oil layers. When the same voltage is applied to a different color of oil in the vertical stack, the new color exposes the insulator's surface and changes the pixel's color. Applying a voltage to all three layers of the stack creates a white background.
The research team believes the vertical-stack display offers higher resolution and smaller, brighter pixels than other electrowetting displays. A video overview of the technology is at www.liquavista.com/downloads/lqvOverviewPresentation.aspx. Researchers continue finding better and better ways to create pixels.
Literally Flying Pixels
People who know me understand that I'm fascinated with various forms of 3D viewing technology that extend the user experience beyond flat screen. Although a huge amount of work still must be done to solve providing input and to overcome the limitations of stereo, "stepping into" a virtual environment for viewing can be breathtaking. Two MIT teams are developing Flyfire, a unique truly 3D display with no stereo involved.
With Flyfire, a flock of tiny aircraft carrying multicolored LEDs will hover in front of viewers to form an image. As pixels that can move through space, the free-flying LEDs could form a shape-shifting 3D display. Besides displaying moving images like a normal screen, they'll be able to fly to change their position and add real depth.
Experimentation has begun with 10 quad-rotor helicopters. Altitude is controllable to 3 cm, but horizontal positioning is controllable to only 10 cm. Inertial measurement units have similar horizontal positioning issues. Swarm technology is truly evolving, as you can see at http://senseable.mit.edu/flyfire.
Researchers have been working on increasingly larger displays. Issues with large-screen displays include cost, consumption of real estate, and portability. Displaying a truly large panoramic image generally requires multiple projectors that must be manually calibrated. The fewer projectors, the less time calibration takes. Calibration time also inhibits portability.
University of California, Irvine researchers developed a system that combines a camera and software to overcome both cost and portability issues. The camera captures a sample projected image; the software compensates for color differences and projector overlap. NEC/Alienware now offers a curved multipanel display ( www.desktopreview.com/default.asp?newsID=868) that relies on the UC Irvine technology. This technology has fully realized the notion of scaling up the display size in a configuration that is far more portable. For more detail, check http://graphics.ics.uci.edu/drupal/node/13.
Assisting those with disabilities is gaining in importance in the computing community. Although not the size of the commodity market, advances for the disabled (for example, eye tracking) often find more general use. The two projects here have fascinating implications for the visually impaired through the use of haptics.
A Sighted Wheelchair
Luleå University is testing an electric wheelchair that can sense its environment and transmit information to a visually impaired person. The wheelchair has a joystick for steering. A haptic robot acts as a virtual white cane. Basically, a laser scanner creates a simple 3D map of the surroundings. This map is transferred to the haptic robot so that a visually impaired wheelchair driver can sense and avoid obstacles (an open door, oncoming people, animals, and so on).
The first test driver, a visually impaired student, successfully navigated crowded college hallways. Look for more as the research team improves the sweep of input to accommodate hanging and low-lying objects. A video is at www.youtube.com/watch?v=eXMWpa4zYRY.
Touch to See
Visualization is clearly one of the powerful ways we have of gaining insight into complex geometric and nongeometric data. Realistically, without visualization, neither the field of computer graphics nor this publication would exist.
One basic form of visualizing nongeometric data is a graph. Ruth White, a University of Reading student, developed a haptics system that lets the visually impaired perceive a graph in real time. Audio provides more information about the graph's y-axis. Check out a video about the research at www.youtube.com/watch?v=3dIIuebNm0I.