A few years ago, Al Pacino starred in a film about a virtual actress named “Simone” who was so realistic that she fooled the fiction’s Oscar committee. The “virtual” actress in the movie, of course, was played by a real-life woman; but there is some truth behind the fiction. Computers have long been able to create still shots that are indistinguishable from actual photographs. Simulated live-action sequences are almost to this point, while interactive gaming has become both increasingly realistic and increasingly... well, graphic in its depiction of sexuality.
These breakthroughs open the possibility for the creation of virtual pornography that is of such high quality that it can compete with actual filmed sequences. The advantages to porn producers would be enormous, especially when it comes to handling talent.. No more no-shows. No more whining that “I don’t do anal.” No more paychecks. And, of course, no more AIDS. The advantages to porn consumers may be even more enormous. Virtual porn could enable the simulation of virtually any sexual act, involving virtually anybody, taking place anytime and anywhere. The possibilities are literally endless.
How close are we to achieving this level of realism? And how will it change online erotica?
State of the Software
Let’s start with a scenario. A porn consumer clicks on “Build Custom Scene.” This opens up a menu of various sexual positions, e.g. missionary, reverse cowgirl, double penetration, etc. After the consumer selects a sexual act, depending upon the act selected, a series of menus appears for the selection of specific “actors.” The consumer can change the face, build, coloring, outfits, etc. of any and all of the virtual participants. Finally, the consumer clicks on “Build Scene,” and within minutes, the customized porn scene is ready to be viewed. During the scene, the consumer might even control the action and choose the ending “money shot.”
Needless to say, consumers would pay for such a service, particularly if the characters involved could be created from celebrities, pre-defined idealized beauties, or (better yet) an uploaded portrait photo of the consumer’s wife’s hot girlfriend. The technology already exists to provide this kind of service – except that it would be far too expensive to offer to the average consumer. However, that is a temporary limitation that will soon disappear as the result of impending improvements in both hardware and software technology.
There are three software techniques that need to be in place, and available economically, in order to make the above scenario into a commercial reality. The first is the ability to quickly create textures and three-dimensional shapes that are vivid enough to fool the eye into thinking that it’s looking at the real thing. That’s already been achieved, as evidenced by the much-publicized Miss Digital World contest (www.missdigitalworld.com). The “contestants” on that site are so realistic that some actually contain the kind of beauty-enhancing flaws one would expect to find on a real-life model (such as Cindy Crawford’s lip mole).
The second software technique that’s necessary is the ability to rotate and move those three dimensional shapes, so that the virtual camera can “pan” around the scene to show different viewpoints. The software to do this also already exists, as evidenced by many commercially-available video games. However, in these games, the level of realism is limited by the number of shapes, or polygons, that the computer can simultaneously manipulate. Thus, in order to make it move inside the virtual “world,” a “face” in a video game, for example, has to have nose consisting of 10 or 15 polygons. While this is sufficient for combat simulations, it’s unlikely to provide much erotic stimulation.
However, the cartoonishness of today’s computer gaming graphics is simply a limitation in computer power, not in programming technique. From the standpoint of the software, it’s much the same thing to manipulate, say, the dozens of polygons that might make up the ninja in a simple computer game as it is would be to manipulate the thousands of polygons that might make up a virtual figure model’s face.
The third and final required software technique is the ability to make ultra-realistic three-dimensional figures move as if they were real. This turns out to be far more complex than either of the previous two problems, because in order to make such real world movement reproducible in the “virtual” world, it needs to be reduced to a set of algorithms, or mathematical formulae, which define the characteristics of the movement. (This is why computer gaming characters have a limited number of positions or poses.) Many of the algorithms needed for the simulation of sexual activity, such as the realistic bouncing of a D-cup breast, are already in place – as evidenced by the space whore robot character “Six of One” in the sci-fi TV show Tripping the Rift (www.scifi.com/tripping).
Simulating sexual acts means undertaking the kind of preparation work that goes into the high-quality computer graphics simulations that are part of first-run movies. This involves putting multiple sensors on live actors and filming them with special cameras that record the positions of those sensors while the actors moving. Those actions are then reduced to their mathematical equivalents, which provide the raw information to animate the polygons that define the virtual actors.
If all of this sounds a bit complicated, it’s important to remember that software handles most of the busywork. Furthermore, such technology is rapidly becoming within the reach of even a very small business operation. Single individuals can already produce their own computer graphics short films without too much bother. Jeff Lew (www.JeffLew.com), for example, recently created the short movie Killer Bean 2 that, in its own way, is as entertaining as much of what comes out of Hollywood. While Lew’s work is admittedly cartoonish, some of the material coming out of top computer graphics studios is stunning, as shown by the 2004 “Show Reel” available from Soul Pix (www.soulpix.de/frames.html).
To repeat, the software is already in place to create extremely high-quality virtual porn. There’s only one problem. Using today’s technology, making a virtual porn film still costs much more than simply setting up a camera and throwing some rubber-coated woodies at the poon-tang. In order to be cost effective, the software will not only have to become easier to use (which is happening), but computer power will need to be vastly less expensive than it is today.
State of the Hardware
As of this writing, the second element in the virtual porn equation, computer hardware, is comprised at the highest end of PCs that have CPU chips that run at 3.4 Gigahertz, many thousands of times the power of the original IBM PC. While that seems fast, it’s not nearly fast enough to do the kind real-time graphics transformations that will be required in order to make virtual porn cost-effective. However, that’s all about to change.
The reason for this is something called “Moore’s Law,” an observation by one of the pioneers of computer hardware 40 years ago that the number of transistors on a chip doubles every two years. What’s interesting about Moore’s Law is that it’s a geometric progression, and thus operates like compound interest, accelerating over time. The more components that you put on a chip, the cheaper it is per component, which means that each generation of CPU power gets geometrically less expensive to manufacture and geometrically more powerful. This sequence is on the verge of reaching threshold, which will result in the ability to build supercomputers at the price point of one of today’s PCs.
Some of the computer architectures on the drawing board are almost beyond belief. IBM Research recently built molecular-level circuits so small that 190 billion would fit on a 7mm pencil eraser. Mapped onto a 300x300mm chip wafer, that’s 349 trillion circuits. Cram these wafers into a 20x60x60cm desktop tower PC case, and you’re dealing with 279 quintillion circuits. Assemble just 20 of these boxes, and you end up with enough circuitry to store and process all the words ever spoken by all the human beings who have ever lived on the planet. With this kind of computer power, handling hundreds of three-dimensional, high-resolution, high-quality virtual actors would be trivial, which is good news for anybody who might want to simulate orgy scenes involving simulacra of favorite pro-football cheerleading squads.
Such mega-architectures are a bit down the road, but the benefits of the “compound interest” of Moore’s Law is about to impact the general marketplace. IBM, Toshiba and Sony are working on a production CPU chip called the “Cell,” which IBM bills as a “supercomputer on a chip.” The intent of this project is to leapfrog current Intel CPU technology and create an alternative CPU architecture which, according to industry sources, is roughly 100 times more powerful than today’s Pentium 4 chips. More importantly, because the Cell is likely to be the CPU for the Playstation 3, it will be manufactured in volumes high enough to make the per-chip price extremely low. At the same time, Intel is scrambling to create new technology that will compete with the Cell, raising the possibility of a price war for CPU hardware.
In other words, CPU power is about to become almost unbelievably inexpensive. So inexpensive, in fact, that some high-tech analysts are concerned that the chip firms may end up creating so much computer power that consumers will have to find new applications in order to chew it all up. If such a “application vacuum” occurs, the ability to create and manipulate virtual porn is an excellent candidate to fulfill that need.
Improvements in the average computing environment will create market conditions where virtual porn will be not only economical, but greatly in demand, especially by computer vendors, who will want to encourage applications that justify the purchasing of new computer hardware. In short, if Bill Gates wants to keep making money, he’s going to have to make certain that the software exists to make virtual porn easy to create and consume.
State of the Culture
Of course, not every porn consumer will need a super-CPU in order to enjoy virtual porn. All that will be necessary is for porn Webmasters to provide the backend processing power as a computer-intensive task on their servers. Under these conditions, high-resolution, customized porn could be implemented within the next two or three years, providing that somebody in the porn industry has the foresight to make it happen.
Some questions remain, though. First, while there’s no doubt that some porn consumers will love the ability to place their favorite celebrities, ex-girlfriends, etc. in compromising positions (as it were), there would appear to be a large portion of the porn-consuming public for which the primary turn-on is that the porn is “real” and that what was filmed actually happened. There are sites, mostly amateur and semi-amateur, that encourage audience participation – even to the point of having site members participate in the filming, as with the notorious comeonherface.com.
It’s also no great secret that many porn stars double as escorts and exotic dancers, and generate quite a following in both roles, drawing largely from the audience of their X-rated performances. This fact suggests that some porn consumers may not be as interested in porn stars who don’t actually exist in the real world. However, the most common form of “human” interaction between porn stars and porn consumers is through video chat, an application where a virtual porn star (who never tires and will do anything) could possibly be more effective than a real one.
In fact, a virtual porn chat room operation could be backended with Artificial Intelligence, of the kind found in “Ramona,” an interactive chat-bot prototyped on AI pioneer Ray Kurzweil’s Website kurzweilai.net. While the richness of conversing with Ramona leaves much to be desired for its intended application (customer support), the limitations of the program might be perfect for the repetitive nature of sexually oriented chat requests. And if an AI interface proves too unrealistic, it would always be possible to have a human provide the back-end intelligence. That means that chat site operators, rather than having to cope with legions of unreliable and cranky better-than-average ex-mall chicks, could put up gorgeous front ends, with voice and intelligence provided by the same overweight housewives who staff 976 lines.
There are also legal issues. First and foremost, virtual porn would make it possible, indeed trivial, to simulate acts that are illegal. Law enforcement officials who rightly want to prevent exploitation might find that more difficult if they had to spend extra time and effort to trace sexual acts that never actually took place. Less importantly, there may be legal ramifications concerning the use of celebrity faces and bodies or, more accurately, virtual faces and bodies that closely resemble celebrity faces and bodies.
For example, if a porn site built a virtual character that combined the features of, say, Jenna Jameson and Pamela Anderson (silicon meets silicone?), would that constitute a violation of their individual trademarks? The issue would become even more unclear if the resulting composite simulacra could be edited to the consumer’s tastes. (“Uh, could you make those boobies just a leeeetle bit larger?”)
Legal problems aside, there’s little question that virtual porn will be a reality within the next five years. Predictably, there will be an outcry; but it’s unlikely that objectors will be able to generate as much heat and umbrage when the only “victim” is a computer program. One thing is certain, though. Despite the technical and legal challenges, somebody is going to make a truckload of money on this perfect intersection of sex and technology. The only question is: Who?
