Digitizer Technology

Digitizer Technology

Three popular types of digitizer technologies are in use today: resistive, electrostatic, and electromagnetic. The following sections describe these different technologies.

Resistive Digitizers

The least expensive type of digitizer is resistive. Resistive digitizers can have one or two layers. A single-layer resistive digitizer is made with a transparent film of known, predictable resistance (hence the name). Voltages can be applied at the four corners of the film. Because the film introduces even resistance across its surface, applying a voltage on one edge of the film creates a gradient of voltage across the entire film. Figure 2-3 shows how applying voltage at the corners can introduce such a gradient. When a single-layer resistive digitizer is used, a pen connected to the computer reads the voltage when the pen makes contact with the film. By alternating between horizontal and vertical gradients, the computer can pinpoint the pen s position by making two readings on contact: a horizontal gradient voltage reading for the x position and a vertical gradient voltage reading for the y position.

Figure 2-3. A horizontal gradient is applied by raising the voltage of the two left corners. Similarly, a vertical gradient is applied by raising the voltage of the two top corners.

The main disadvantage of a single-layer resistive digitizer is that its pen needs to be connected back to it because voltages are read through the pen when it makes contact with the resistive film. Two-layer designs sidestep this disadvantage by means of a highly conductive film. This second film is placed below the resistive film, and the two are separated by a series of spacers usually set in a layer of polyester. A cross section of this design is shown in Figure 2-4. When a pen presses down on the resistive film, it forces the film to make contact with the conductive film beneath, which then conducts the voltage to the digitizer circuitry. Because the voltage is communicated to the digitizer directly through the conductive film, the pen does not need to be connected to the computer at all. In fact, because contact between the two films is all that s required, a pen isn t even needed! Your finger, or any blunt object for that matter, would serve just as well.

Figure 2-4. A two-layer resistive digitizer shown in cross section. The resistive top layer is kept apart from the conductive bottom layer by a system of spacers.

Despite the innovation of two-layer resistive digitizers, all resistive digitizers still have other disadvantages. Because resistive digitizers need to be placed over the LCD screen, they greatly reduce the LCD s brightness. This reduction can be significant, in many cases over 30 percent. Although you can address the brightness issue by increasing the LCD s backlight, you cannot mitigate the cloudiness that the overlaying films introduce. Another practical concern with resistive digitizers is that they cannot detect hover, pressure, or tilt. This is a direct consequence of their design, which dictates that voltage values be read on contact. Without hover, pressure, or tilt, resistive digitizers are not practical for use in Tablet PCs. Furthermore, two-layer resistive digitizers do not handle users resting their hands on the screen while writing, a perfectly natural thing to do. The mere act of resting a hand on a two-layer resistive digitizer causes the films to meet, generating a false pen impression.

Even though resistive digitizers have many drawbacks, they are by far the dominant technology in handheld devices such as the Pocket PC. These devices use resistive technology because of its low price. The flexibility of being able to detect the user s finger is viewed as a benefit, and the small screen makes hand contact interference less of a problem. Resistive digitizers are not used on Tablet PCs because they do not supply hover and pressure information.

Electrostatic Digitizers

Electrostatic (or capacitive) digitizers are made by bonding a thin conductive film to a sheet of glass. A pen tethered to the digitizer emits a high-frequency signal that is picked up by the conductive film. The film can sense the pen s signal even from a small distance away, giving electrostatic digitizers an important edge over resistive digitizers electrostatic digitizers can sense pens even without direct contact. The distance of the pen to the digitizer can be determined by the relative strength of the detected signal. Because the pen is constantly emitting a signal, the digitizer can extrapolate how far away the pen is hovering by tracking the fluctuations in the signal s amplitude.

Electrostatic digitizers can also determine the tilt of a pen. Figure 2-5 shows the axes of pen movement, including hover and tilt. When the pen is tilted relative to the surface of the digitizer, it distorts the shape of the field generated by its signal. The conductive film detects this distortion, and the digitizer interprets it as a tilt measurement.

Figure 2-5. Hover, tilt, rotation, and pressure are the four axes of pen movement.

Conventional electrostatic digitizers require the pen to be connected to the digitizer because the pen itself needs to emit a signal. This requirement makes such digitizers less desirable for use in Tablet PCs.

Touch Pads

A variant of electrostatic digitizers is used in touch pads that come with laptop computers. Touch pads were once built on resistive technology, which had the drawback that a small amount of pressure had to be exerted on top of the pad for it to detect a touch. Repeated applications of this pressure would eventually cause the pad to wear out. Modern touch pads use a two-layer electrostatic technology that relies on mutual capacitance between the layers. Fingers, which have different dielectric properties from air, change the mutual capacitance when they approach and touch the pad. Interacting with the touch pad requires little pressure, making electrostatic touch pads more durable than their resistive counterparts.

Electromagnetic Digitizers

Electromagnetic digitizers work in a manner similar to their electrostatic cousins. However, they require much less power because their pens need to emit only a low-power RF (radio frequency) signal. The reduced power consumption makes it possible for electromagnetic digitizers to use pens that are not tethered to the digitizer. Instead, the pen can contain all the necessary circuitry, along with a little battery, to generate the RF signal by itself.

In addition to having pens that need not be attached, electromagnetic digitizers can be installed below the LCD screen. This flexibility is quite a benefit, and one that distinguishes electromagnetic digitizers from other types of digitizers. By installing below the screen, electromagnetic digitizers do not affect the screen s brightness or clarity. Because the screen s brightness is not affected, electromagnetic digitizers have lower power requirements for an equivalently bright screen. Like electrostatic digitizers, electromagnetic digitizers can detect hovering and tilt. With some additional smarts, some electromagnetic digitizers can also detect rotation. A cross section of an electromagnetic digitizer is shown in Figure 2-6.

Figure 2-6. An electromagnetic digitizer in cross section.

Electromagnetic digitizers are the best of the existing pen technology. Low power requirements, stand-alone pens, and behind-the-screen mounting make them the clear choice for Tablet PCs.

Digitizer Distortion

All electronic devices generate electromagnetic fields when operating. Within the confines of a computer, some devices can generate large electromagnetic fields. For instance, the hard drive is usually a significant source of such fields, as is the power supply. Other devices within a computer, such as video cards and memory chips, also generate electromagnetic fields (albeit much smaller ones). Together, the electronics within a computer create a halo of electromagnetic interference. The closer you get to the computer, the stronger the interference you will experience.

Electromagnetic digitizers usually have a back plane under them that is designed to shield them from electromagnetic interference. Unfortunately, these back planes aren t perfect. They might not always be able to isolate the digitizer from interference originating from below. In the case of Tablet PCs, the compact size of the computer exacerbates the problem of electromagnetic interference leaking through the protective backplane. Ultraslim tablets (which can be less than three quarters of an inch thick) have digitizers that are literally millimeters away from underlying hardware. It becomes impossible at this proximity for the backplane to do its job perfectly, so electromagnetic interference is introduced into the digitizer s readings.

What does this interference mean in practical terms for the digitizer? Figure 2-7 shows the distortion lines on a typical Tablet PC. These lines started as a series of straight, parallel lines drawn across the surface of the tablet (say, with the aid of a ruler). However, the digitizer s readings tell a completely different story. As you can see, the digitizer hardly thinks that the lines were straight. It was misled by electromagnetic signals arising from the hardware below it. These signals commingle with the signals the pen generates, resulting in an inaccurate reading by the digitizer.

Figure 2-7. A Tablet PC s digitizer readings of what began as straight lines drawn on its surface. Notice the strong correlation between certain parts of the underlying hardware and the visible distortion.

These distortions are present in the readings of every tablet s digitizer. If they are left uncorrected, the user who tries to draw a straight line on the tablet will instead get a curve. Tapping on a particular spot becomes a challenge because the digitizer rarely agrees with the user about where the pen is pointing. Fortunately, this rather serious problem is not left uncorrected! Tablet PC manufacturers carefully calibrate their digitizers by first taking a series of readings like the ones shown in Figure 2-7. Armed with these distortion lines, they then program each digitizer s device driver to account for the distortions. The digitizer s device driver acts as a filter, adjusting every reading to correct known distortions. The result is that the user typically does not perceive any problem.

Calibration to account for electromagnetic interference is never perfect, however. The main reason for varying interference patterns is that every Tablet PC has a slightly different electromagnetic halo. Even if all the hardware a particular manufacturer uses is exactly the same for each tablet it produces, minute differences still exist between each individual unit. A little spacing difference here, a little twisted cable there, and pretty soon you ve got a unique electromagnetic halo.

Another reason for different interference patterns is entirely out of the manufacturer s control: the real bummer with electromagnetic interference is that it varies with the surroundings. All sorts of external factors change a tablet s interference pattern: temperature, humidity, a nearby Rolex, the user s pacemaker, and solar flares. (OK, maybe we re pushing it too far, but you get the idea.) The most you can hope for is that, under normal conditions, these external sources are somewhat random and transient, often canceling each other out. The net effect should be that the pen seems accurate most of the time. Every once in a while, though, you might find that the pen does not seem to go exactly where you point it. It is precisely at those times that you should seriously consider living farther away from the neighborhood hydroelectric power plant!

Parallax

Even if a tablet is calibrated correctly to offset the effects of electromagnetic interference, there might still be a difference between where a user thinks the pen is pointed versus where the system thinks it s pointed. This difference is due to a phenomenon called parallax. Parallax is the difference between the apparent locations of an object when viewed from two angles. Although the term might sound esoteric, parallax is something you encounter in everyday life. Anyone who has looked into an aquarium through its thick glass walls understands the effects of parallax. As you change the position from which you look into the aquarium, you notice strange distortions in the appearance of what s inside. These distortions are refraction-induced parallax effects: the thick glass refracts light, which alters the apparent location of contents within the aquarium when viewed from different angles. Refraction-induced parallax is diagrammed in Figure 2-8.

Figure 2-8. Refraction creates parallax, but even without refraction, a screen s thickness alone can create parallax.

Being an alert reader, you ve by this point probably begun to suspect that the glass covering a tablet s LCD screen doesn t create enough refraction to account for all the parallax. And you re right refraction-induced parallax is only part of the problem on a tablet. Even if there were no refraction, tablets would still have parallax problems. The second source of parallax, also pictured in Figure 2-8, comes from the distance between the surface of the LCD screen s image and the surface of the tablet. All LCD screens come with a layer of protective glass (or other material). To this, tablet manufacturers add a layer, usually made of plastic, to protect the screen from the constant rubbing of the pen s tip. These protective layers together introduce several millimeters between the surface of the tablet and the underlying screen image. When viewing the tip of a pen and its corresponding screen cursor from an angle, parallax results. Parallax affects users nearly all the time because they rarely look at the pen from directly over it. It s much more natural to look at the pen from a slight angle when writing.

Parallax is unavoidable, for the reasons just mentioned. It can lead to user confusion if not dealt with because the user may end up pointing to and tapping on the wrong things. You should take precautions to reduce the negative effects of parallax in your applications. One precaution is to make sure that controls are big enough to be easily targeted. Another equally important improvement is to offer clearly visible cursors that track pen movement so that the user can reference the cursor s position on screen as the authoritative point of interaction even if parallax is present. Without a reasonably sized cursor that can t be blocked by the pen itself, it s difficult for the user to know what to expect when tapping.

NOTE
You can address the problem of parallax by making controls big enough and providing clearly visible cursors.

The good news is that users, after prolonged use, tend to adapt to parallax. As they come to anticipate a difference between where they think the pen points and where the system thinks it points, they become more adept at using the pen. In the meantime, it s helpful to have onscreen cursor feedback to mitigate this problem.

Still Motion

One of the user s most immediate challenges when using a Tablet PC is the difficulty of holding the pen steady. Muscles in a human body are not wired to hold absolutely still. Instead, when you try to hold a muscle steady, in a fixed position, your brain causes alternate muscle bundles within that muscle to tense up in order to give each bundle the chance to relax. This taking turns within a muscle causes the muscle to vibrate just a little around that fixed position.

Make a dot on a sheet of paper and try to hover the tip of your pen about two millimeters above that dot. If you look at the tip of the pen carefully, you will see that it moves around quite noticeably (especially after that three-shot espresso). In fact, the harder you try to hold the pen still, the more it will fluctuate. Relaxing your hand helps, but it is hard to get it to stop quivering. This phenomenon carries over to when a user tries to use a tablet with a pen.

The quivering-pen syndrome does not exist with conventional pointing devices, such as a mouse or a trackball. These devices stay where you park them, making pointing and hovering easy. If you want to keep the mouse over a certain toolbar button, simply make minor adjustments to its position until it rests in the desired location. Clicking on a point is similarly easy just make sure the mouse is positioned correctly, and then click.

Pens, however, do not work like mice. As a result of the quivering-pen syndrome, several big challenges arise, including the following:

• Precise targeting

  A tablet user cannot be expected to target too small an area on screen with a pen. The fact that the pen is always moving around a little makes it hard to land in any particular location. Parallax exacerbates the problem. These two factors make some actions a bit difficult on a Tablet PC. For instance, the typical resizing area of a window border is about 3 pixels wide. That is way too small for a pen to easily target. Some controls on toolbars or in dialog boxes are also too close together, making accidental taps inevitable. An example of controls placed too closely together is given in Figure 2-9.

  

  Figure 2-9. These toolbar controls in Microsoft Word are too small and too close for a Tablet PC user to reliably target.

• Double-Clicking

  It s hard enough for the user to tap once on a desired location. Because the pen is constantly shifting position, tapping twice in the same place is even more difficult. The operating system s default settings for double-click detection require that both taps fall on the same pixel. If the two taps fail to do this, the operating system does not register a double click. In the case of a pen, those two taps are rarely on the same pixel, making double-clicking quite a challenge for the user.

• Tap without drag

  It s easy to click a button on a mouse without moving its position on screen. Tapping on a fixed position with a pen is a lot harder. The problem is that when the tip of the pen touches the screen with some force, it inevitably tends to move a little on the screen. This tendency is exacerbated if the screen is covered with protective material that is too smooth. Imagine using a ballpoint pen to make a somewhat forceful dot on a sheet of glass. If you hold the pen at a slight angle (which is natural in the course of writing), the pen tends to slide on the glass when it makes contact, leaving a streak instead of a dot behind. The same thing happens on tablets sometimes users inadvertently drag something they intended to tap.

• Hovering

  Hovering, in the sense that a mouse cursor can hover over a single pixel, is impossible on a tablet device using a pen. At first, this revelation might seem insignificant. But what about features like ToolTips? ToolTips are the little yellow boxes that pop up with useful descriptions when you hover the cursor over a toolbar icon long enough. In general, ToolTips require that the cursor be absolutely still for a few seconds before they will appear. With a pen, this requirement simply can t be fulfilled. Hovering also triggers similar features in other applications.

You might be wondering what you can do to mitigate these challenges. Fortunately, help is available! Windows XP Tablet PC Edition contains some modifications that specifically address a few of these challenges. The first modification is in double-click detection. Instead of requiring the two taps to be on the same pixel, Windows XP Tablet PC Edition relaxes the requirement to a general area of sensitivity. As long as the two taps land sufficiently close to the same pixel, the system will register a double click. This modification will affect all applications that use the standard system facilities for detecting a double click (WM_LBUTTONDBLCLK and its family of messages), so most applications will automatically receive the benefit. Applications that do not use the standard system provisions for double click but instead implement their own double-click detection will not benefit from this change. Unfortunately, a surprising number of applications continue to implement their own double-click detection, which makes addressing this problem in a centralized manner difficult.

The second modification in Windows XP Tablet PC Edition is in toolbar hover detection for the purposes of displaying a ToolTip. Microsoft has modified the toolbar control in the common controls library (Comctl32.dll) in Windows XP Tablet PC Edition so that hovering anywhere within a toolbar button for the required amount of time will yield a ToolTip. This behavior stands in stark contrast with the unmodified version of the toolbar control, which displays a ToolTip only if the cursor does not move at all. Now, at least, the cursor can move within the area of a toolbar button and still bring up its ToolTip. As with double-click detection, applications will benefit from this modification only if they use the standard system toolbars. You ll find that some applications don t, once again diluting the benefits for the user.

You can also address these challenges when designing your applications. Perhaps most important is to make sure that controls are large enough and separated from each other by enough space so that they are easy to use on a tablet. The simplest way to know whether you ve got it right is to try your user interface on a tablet device with a pen. In general, you ll find that with focus and determination, most users can stay within about a 5-by-5-millimeter area on a tablet screen. This area varies with the individual user, so try to leave generous margins for error. The second modification you can make for users is eliminating features that are accessible only through hovering. Although it is OK to have hover-triggered features, pen-based users will appreciate alternative methods of triggering the same feature. Another option is to make the hover tolerance more lenient so that hovering within a general area, rather than just over one pixel, is sufficient.

Finally, we encourage you to consider using standard system UI controls when possible. In the current version of Windows XP Tablet PC Edition, Microsoft has modified many controls for the benefit of pen users. Future versions of Windows XP Tablet PC Edition might have other enhancements to standard controls to make them even easier to use with a pen. This means that as new versions of Windows XP Tablet PC Edition are introduced, your application will automatically benefit from those enhancements. If, on the other hand, you re creating your own controls, we suggest that you take into account the issues we ve discussed in this chapter concerning the design of those controls.

Handedness

The second major user-introduced set of challenges comes from the user s hand. When writing with a pen, the user s hand ends up blocking a good portion of the screen. The key here is which portion of the screen gets blocked. The screen is blocked differently depending on whether the user is right-handed or left-handed. For an example of how serious the blocking effect can be, see Figure 2-10.

Figure 2-10. The majority of a menu is blocked by the user s hand.

Usually, blocking of the screen by the user s hand is not that big of a deal because the user can just move his hand to see what s obscured. In some cases, however, moving the hand is not possible or desirable. Consider the case of menus and ToolTips. By default, they both appear to the lower right of the cursor. This positioning is a problem when you consider that most people are right-handed. A right-handed user s hand is (surprise, surprise) to the lower right of the cursor as well, so when he hovers over a toolbar, the ToolTip comes up under his hand. In moving his hand to see the ToolTip, the cursor moves, thereby making the ToolTip disappear!

A partial resolution to this problem, available in Windows XP Tablet PC Edition, is that menus and toolbars respect a handedness setting. By handedness, we mean the hand with which a user writes. In the Tablet PC Control Panel item, you can tell the system this information. In the case of right-handed users, the system paints menus and ToolTips to the lower left of the cursor, making them clearly visible. An example of this behavior is shown in Figure 2-11.

Figure 2-11. The same menu as shown in Figure 2-10, this time shown for a right-handed user.

The handedness setting affects only the standard system menus and ToolTips, so applications that paint their own menus or implement their own toolbars will most likely not benefit from this enhancement. As long as you use the standard menus and toolbars, your application will behave according to the user s preferences on Windows XP Tablet PC Edition.

NOTE
If you do implement your own menus or toolbars, you can query the user s handedness setting by calling SystemParametersInfo to check the SPI_GETMENUDROPALIGNMENT flag.