Tennant covers SIGGRAPH 2006
Excerpts and Highlights from Dan Tennant's coverage of ACM's SIGGRAPH 2006 conference in Boston:
Realism is, hands down, the holy grail for game developers, and their journey towards photo-realistic gameplay is getting a very nice boost courtesy of ATI. ATI is showcasing a new technology called Parallax Occlusion Mapping at its booth on the show floor. For those of you somewhat familiar with graphical terms, think of POM as a variant of bump mapping or normal mapping, but with actual depth -- not the "simulated" depth other technologies create with fancy lighting effects. To show off the technology to me yesterday, ATI loaded a 3D representation of a cobblestone street in which each stone had a significant height raised well above its setting in the ground. The stones were not, however, shaped individually by a 3D artist; rather, the whole walkway was modeled as a perfectly flat, two-polygon surface (one polygon is a triangle, so just two are needed to make a rectangle). The POM technology gave the walkway actual depth, making it seem as though thousands of polygons were used to create a vision actually modeled with only a couple. To really show it off, an employee at the booth panned the camera down and swung it between two cobblestones -- something utterly impossible with normal mapping's "simulated" depth, which is revealed as truly flat at such ranges and angles. POM looks to be singularly awesome, and it's already hitting the streets on the XBOX 360. Gamers, cue your drooling.Tomorrow's Treasures, Day 2 - SIGGRAPH 2006 - Computerworld Blogs
I watched as users opened multiple instances of the same map, resized and moved the windows to partially overlap each other, set each instance to display the single map differently (one topographically, another politically, etc.), and then effortlessly scroll, rotate, and zoom the map, watching as it dynamically shifted to their easy touch.
To put it short and sweet, this is exactly the kind of technology that would be positively awesome to see in a future kitchen. Forget posting pictures on the refrigerator -- just bring them up on the wall, scale them, crop them, cut them, rotate them, and organize them at your leisure. Stop buying post-its for use in to-do lists; just use the integrated wall-keyboard to tap out animated notes for your loved ones. Need to check a map before heading out? I couldn't imagine an easier-to-master interface.
Tomorrow's Treasures, Day 3 - SIGGRAPH 2006 - Computerworld Blogs
Though graphics and animation are obviously the overriding themes of SIGGRAPH 2006, there are several sub-themes that unite the various showcased technologies into mini-groups. For example, there are numerous innovative touch interfaces throughout the conference, both in the Emerging Technologies hall and on the main floor (the Multi-Touch Wall I mentioned yesterday being one of the stronger contenders in the area). One sub-theme of the conference is, unsurprisingly, game-related technology; considering the fact that computer and video games so often drive the development of video tech, many companies on the show floor are using games to show off their individual products, while others are simply showing off their games.
One of the booths I stopped by yesterday was run by Linden Lab, the creators of the massively multiplayer online community known as Second Life. Jeffrey Ventrella, one of the technical developers for the game, invited me to take a gander at a player-driven animation technology he was working on, and I must admit that I was impressed, despite the fact that I don't play the game. Apparently, players will soon be able to pose their online avatars by moving and manipulating body parts through a very intuitive, point-and-click interface. Not long afterwards, they'll be able to create personalized animations using the technology, saving them for hotkey-use in future sessions.
LucasArts, the gaming arm of George Lucas's development kingdom, is also sharing some new technology at the show. The company showed off a technology yesterday called euphoria (all lowercase, because apparently that makes it look modern) by Natural Motion Ltd. euphoria is being used in LucasArts' upcoming Indiana Jones title, driving much of the animation of the game using a set of predefined rules rather than scripted keyframes. When something smacks our good friend Indy in the back of the head, for instance, euphoria determines how he'll be jarred forward, will attempt to recover his balance, and then either stay up or fall over, creating all the animations on the fly. Thus, almost every time anything animates in the game, the movements are subtly different. It's a step in the right direction for realism.
Realism is, hands down, the holy grail for game developers, and their journey towards photo-realistic gameplay is getting a very nice boost courtesy of ATI. ATI is showcasing a new technology called Parallax Occlusion Mapping at its booth on the show floor. For those of you somewhat familiar with graphical terms, think of POM as a variant of bump mapping or normal mapping, but with actual depth -- not the "simulated" depth other technologies create with fancy lighting effects. To show off the technology to me yesterday, ATI loaded a 3D representation of a cobblestone street in which each stone had a significant height raised well above its setting in the ground. The stones were not, however, shaped individually by a 3D artist; rather, the whole walkway was modeled as a perfectly flat, two-polygon surface (one polygon is a triangle, so just two are needed to make a rectangle). The POM technology gave the walkway actual depth, making it seem as though thousands of polygons were used to create a vision actually modeled with only a couple. To really show it off, an employee at the booth panned the camera down and swung it between two cobblestones -- something utterly impossible with normal mapping's "simulated" depth, which is revealed as truly flat at such ranges and angles. POM looks to be singularly awesome, and it's already hitting the streets on the XBOX 360. Gamers, cue your drooling.
Tomorrow's Treasures, Day 4 - SIGGRAPH 2006 - Computerworld Blogs
One of the more frequently-used technologies in the realm of VR here at SIGGRAPH is a subsonic noise-based tracking system developed by a company called InterSense. Though several booths were showing off giant VR screens for various marketing purposes, they all used InterSense tracking systems to allow users to interact with their virtual wares. The tech works by implementing a set of three tracking bars on the top and sides of a given VR screen (itself little more than a rear-projection screen about 10 feet diagonally). Each bar both emits and receives sonic waves to and from a pair of wireless devices: a glasses-mounted sensor to track eye position (and thus modify a projected VR image on the fly to align it perfectly with the user's eyes) and a handheld sensor attached to an interacting pointer device. In order for the pointer to emit a virtual beam on the screen, giving it a presence in the virtual world, InterSense's sensors must track its location and orientation so that its virtual reflection can be displayed to match.
Thus, when a user pops on a pair of stereo glasses and grabs the pointer, a virtual beam of light is "emitted" from its tip, perfectly aligned no matter where the user is standing or the device is held. Given a virtual representation of a human heart, for instance, a user could hold the pointer horizontally, casting its beam in front of the heart, and then literally step forward, swing the beam around, and cast it through or behind the heart. The effect is startlingly convincing, as though you're moving a handheld pointer around and behind an actual three-dimensional object, pointing at (or modifying at will) different parts of it.
Different VR projectors are provided by different companies and paired with different kinds of eyewear, even though InterSense's technology is used for tracking purposes almost universally. There are two different kinds of stereo projection techniques: passive and active. Passive stereo is what you're probably used to: two projectors simultaneously emit images to the screen, but each projector emits polarized light at different angles. A user wears a pair of polarized glasses to view the image, so that each eye is fed a view from a different projector. Assuming that the distance and angle of the two projected images are calibrated correctly to each other and the viewer (helped along under optimal conditions by one of InterSense's glasses-mounted trackers), you end up with a very decent 3D image. IMAX uses such technology to full effect every day; here at SIGGRAPH, Purdue University is using passive projectors to drive a VR flythrough of its entire campus.
Active stereo, however, takes things a step further than their counterparts. A single projector emits a flip-flopping image to the VR screen, toggling back and forth between the view designed for one eye and the view designed for the other. The user wears special glasses that have tiny shutters in the lenses, opening and closing in time with the projector (which, as you can imagine, is far too fast for the eye to notice). The left shutter opens to view the left image exactly when the projector displays it, and then the process shifts to the right eye. Back and forth, dozens of times a second, and you end up with a 3D image that doesn't suffer from the slight hazing that passive stereo is so often plagued with. InterSense itself is using active stereo at its booth here at SIGGRAPH, as is Barco -- though the latter company has taken things a step further than anyone else by providing six-sided VR cubes for highly immersive experiences.
Tomorrow's Treasures, Day 5 - SIGGRAPH 2006 - Computerworld Blogs
As the final day of ACM's SIGGRAPH 2006 graphics and animation conference rolls around, I wanted to sit down and examine three different technologies showcased at the Boston convention that, given just a bit more development time, could theoretically be used together as the veritable Voltron of virtual reality experiences.
In order to create that experience, we would need technology that could fool the senses into suspending disbelief. Dismissing the olfactory and gustatory systems (smell and taste, respectively), we'd still need to replicate sound (which is easy), sight (which is pretty darned difficult to do convincingly), and touch (which is nigh impossible to accurately simulate). If we wanted to immerse someone -- let's call him Jimmy -- in an environment, we'd need to give him the ability to hear it, see it, feel it, and finally, traverse it.
Hearing, as I said, is a piece of cake; give Jimmy a pair of headphones and a decent sound card and we're set. Seeing is a little harder, though, which is where the good folks at Sensics Inc. come in. Sensics is demonstrating their piSight head-mounted display at SIGGRAPH, which boasts an incredible 150-degree viewing angle -- just an imperceptible smidgen beneath that of human eyesight. The piSight uses multiple OLED-powered displays focused into the eyes with twelve lenses per eye; combine it with InterSense's tracking technology, which I mentioned yesterday, and you get a pretty convincing, fully-3D visual. The level of detail and quality could certainly use a bit of work, but it's unquestionably the best head-mounted display I've ever heard of, much less used.
Immersion is the company responsible for force-feedback technology, so if you've ever felt the rumble of an XBOX controller in response to a game or felt a joystick buck in your hands, that's them talking to you. They produce a product line based on their CyberGlove, which can monitor hand and finger motion. Combined with an exoskeleton they develop (it's ugly as heck, but Jimmy will have his headset on, so he won't notice), the CyberGlove can deliver actual feedback and pressure to the hand, meaning Jimmy could be stopped by a wall in front of him, could pick up and feel a can of soda, could feel the weight of a brick or a suitcase, or could trace the outline of a sculpture with his finger. Let's slip those on Jimmy: not only can he see everything around him, but now he can interact with it all too.
Finally, we need to give Jimmy the ability to walk around in his environment. I'm talking about actual foot motion -- forget about using buttons on the hands to float through virtual space, the only way to be immersed in a VR experience is to be able to walk through it at your own speed and leisure. For that, we have to turn to Hiroo Iwata at the University of Tsukuba in Japan, and his dully-named Powered Shoes, which he presented at SIGGRAPH's Emerging Technology exhibition. The Powered Shoes are literally a form of roller skates whose wheels are actuated by motors. If we were to slip them on Jimmy and he were to walk, the wheels would counter his movements and roll just enough to keep him in one spot. A lot of work still needs to go into the shoes -- changing direction is a little awkward, and you can't run yet -- but the concept is sound, and it works well enough already; it just needs a bit more time to reach perfection.
