Oct 17 2013

Virtual reality is finally coming next year


Virtual Reality has been a joke for years. For all the derision heaped on nonexistent flying cars, jetpacks, and cures for cancer, no seemingly straightforward bit of technology has been vaporware for so long. 1992's film The Lawnmower Man enticed us with the idea of entering a fictional world, a facsimile of Star Trek's Holodeck, without the need for matter/energy conversion or force fields. But the unfeasibility of the technology in 1992 is fairly obvious when you look at the problems: virtual reality headsets, the central component of such a system, would have had to be made of cathode ray tubes back in the time of Lawnmower Man. CRTs are heavy, huge, power hungry, high voltage, glass monstrosities—they therefore make bad apparel. Moreover, the devices needed a way of tracking the orientation of the headset, so that when you turn your head the image in front of your eyes changes accordingly. No good way existed to do this with any real speed, certainly not cheaply. And two of them are needed for stereoscopic (3D) viewing. VR headsets could be produced, but were so bad and expensive that they never really got beyond the lab or high-tech industries.

In the late 90s and early 2000s, head mounted displays using LCDs were released by several manufacturers, but HMD90sthese were normally quite low fidelity and/or obscenely expensive. Large, low density TFT-LCD screens, often backlit with hot lamps, were beginning to be affordable, but nothing small and energy efficient. Consumer models that were true head-tracking displays never materialized in anything above 600 vertical pixel, well below that needed for realism. There wasn't a large market for small high-density screens.

Screens weren't the only problems by a long shot. Lawnmower Man may have shown Pierce Brosnan a terrifying image of his test subject Jeff Fahey, but the graphics were deliberately crude and abstract due to the limitations of computer, and especially graphics, processing, at the time. The technology that would be needed to allow for walking and hand movement would have been no more sophisticated than a joystick; practical motion capture like that used in The Lord of the Rings was quite a bit off, requiring fast, high definition digital photography and computers to parse the data in real time.

All of these problems have been solved. In fact, you've likely already interacted with many of the solutions. And now there are people preparing products that combine the technologies into inexpensive, amazingly lifelike virtual reality. It's coming next year, and a lot of people have already been using it for months. It's going to be incredible, a watershed moment for interactive media.

The solution begins with a device my actual grandmother literally owns: an iPhone. Specifically, the iPhone is a high-pixel-density screen married to a number of micro-electromechanical sensors and a battery with a whopping energy density despite being small and lightweight. It can sense orientation, display a high definition image, and doesn't require a huge power cord or hefty weight to do it; that's practically all a head-mounted display has to do. Put two of these suckers into a piece of plastic, add some optics, and you have the Oculus Rift.

To make a realistic, brain fooling virtual reality setup, you have to get several things right. First, the eye has to focus far away without simply making the screen look far away, which means that nonspherical lenses have to be made. This allows the screens to be physically near the eye but appear to be far from it, and still retain very large field of view. Peripheral vision has to work, and it does with the Rift. The head tracking is incredibly fast, fast enough to be capable of fooling people into thinking the experience is real. Among those people who have ordered and received the Developer's Kit, many report that going inside the Rift is like going into another world; exiting it is disorienting, as you feel you are shifting from one reality to another. If someone taps on your shoulder while you're in the Rift, things get very uncomfortable. Horror games are TOO scary because all of a sudden you really feel like you're there, in danger.

The Rift is also doing what it needs to do to integrate with current game makers, people who have spent the past 20 years pushing for huge amounts of graphical computing power to run their sophisticated 3D game engines. Oculus has made the task of programming a Rift game purportedly easy, something that past head-mounted displays never did. More importantly, they seem to have achieved critical mass, with game makers programming specifically for the Rift and scads of people trying out entirely new types of virtual experiences, such as a game where you experience getting your head cut off. Sony has announced that they are working on a competitor VR display for the Playstation, and such me-too-ism from a major gaming corporation shouldn't be taken lightly.

Oculus is still targeting a $200 launch price. That is highly affordable, much cheaper than previous head-mounted displays. It seems to tick every box that would make an ideal visual experience. And it's so big that it's pushing ahead inventors to work out solutions to the rest of virtual reality.Virtuix

Take walking around. How would that work in a typical home? Voila, a company called Virtuix has developed a way of moving within games by actually walking or running with your actual feet. The shoes slip on a slippery surface, and the motion is digitized by sensors on your ankles, while the support structure holds your body upright. Ben Kuchara said it felt "pretty fucking great." This is a prototype of the first generation of such a thing, and it works well with any game.

What about your hands? The Wii, Kinect, and Playstation Move were the trifecta of camera-based motion tracking of the current generation of game consoles. They all work by using a digital camera and a reference object. The Wii has the camera in the controller which looks at the "sensor bar"; the Kinect and Move both use a stationary camera looking at the person. The Move tracks the size of spherical lights on a controller to determine location quite accurately. The Kinect works on the same principle but instead looks for a skeletal image in frame. A company called Sixense has extended this idea and build 3D tracking hardware that will watch your hands and render them in the game. The sensors can be attached to any part of your body to add more degrees of freedom to the virtual reality experience:

I have the full range of my normal movements, so everyone makes sure I have plenty of space to swing my arms and explore. I pick up a barrel, wind up, and throw it over the balcony, and watch it sail off into the distance.

Walking, seeing, hearing, manipulating are all possible in a 3D stereoscopic, high definition environment. What's left? The other senses. Could haptic feedback provide tactile sensory information? This group made a bunch of touchable buttons in mid-air using an array of ultrasonic oscillators to create focal points of touch in the air. We already have a small measure of force feedback in controllers through rumble. Chairs like this one, when used with the Rift, reportedly provided a full sensory experience much like being in a car, though they are currently quite expensive. But if the Rift takes off like I expect it to, peripherals for the virtual reality experience are going to finally undergo economies of scale.

Now, yes, there are going to be some hiccups. For one thing, if you only have a Rift, playing games using a controller reportedly causes nausea in many people. The sooner technology like Virtuix's walking/running setup catches on, the better, because by all accounts this is due to the disconnect between what the brain thinks is happening (movement) and what the body is saying (no movement). Also, you're going to need a pretty good computer to use the Rift. Many currently recommend a Geforce GTX 660 video card, which is about $175. In a world where many people's computers are Dells with cheap Intel chipsets, that isn't going to fly. Laptops are also basically a non-starter.

But for those of us already prepared for the device, it can't come soon enough. Developers won't just be developing their games to use the Rift, they'll be developing games that are FOR virtual reality. That means that games that would previously be "boring", where you aren't constantly shooting someone or jumping out of helicopters, can be a real interesting thing. A game like Myst wasn't very compelling, but what if you were actually walking around it, manipulating things? Who wouldn't want to do that?

I also think this is a boon for home video. The current technology for 3D movies is passive polarizer glasses, both in theaters and on 3D TVs, and each of these lenses cuts 50% of the light out. The Rift already has two screens, one for each eye, so it can run at the full brightness all the time. Moreover, on an airplane you can watch a movie on a virtual screen in a virtual movie theater---abstracting away the stranger you're sitting next to seems like a great idea to me (that is, until he pokes you in the arm to ask to go to the bathroom).

The time is ripe for VR. The Rift ships next year. Hopefully lots of other people want it as much as I do.

Oct 03 2013

Jeopardy contestants don't know good game theory

Two things demonstrate to me that Jeopardy contestants don't understand rudimentary game theory.

Case 1: In many cases you have a score where one person is way ahead, and the other two trail by an amount so that it's a "run-away", meaning that even if the second place contestant bet the maximum amount in final Jeopardy, it's not possible for him to catch up. Example:

Contestant 1 Contestant 2 Contestant 3
$12000 $5000 $3000

Suppose there only a few clues left with no Daily Doubles available. What is the best strategy for Contestant 3? If there aren't enough clues in that category for him to get to $6000, Contestant 3 should stop buzzing in. Contestant 1 would be foolish to bet a single dollar in Final Jeopardy if he has twice his nearest competitor. Contestant 2, on the other hand, can easily earn enough to get to $6000. Therefore. By continuing to buzz in, Contestant 3 does nothing but take potential money away from Contestant 2, and increase the likelihood of a run-away. This hurts himself, not just Contestant 2, because in the event that both Contestant 1 and 2 miss the question, Contestant 3 can then win.

Case 2: In any of the two-day contests, such as the finals of the Tournament of Champions, wagering a huge amount in Final Jeopardy of the first game is extremely imprudent. Instead, they should treat that final like any other Daily Double. Although the score is reset at the beginning of the next day, the dollar figure that matters is the total of both days. It's therefore insane to bet all your money, unless you would normally bet all your money on a Daily Double (hint: whenever somebody bets a large amount on a Daily Double there are gasps in the audience).

Aug 21 2013

Let's Play Mass Effect - Episode 1

Jessica and I embark on a journey to snark all over a great video game.


Part 2:

Jul 26 2013

Hollywood's box office debacle this summer is not hard to understand

Via Slashdot:

In the weeks since Spielberg’s prediction, six wannabe blockbusters have cratered at the North American box office: “R.I.P.D.,” “Turbo,” “After Earth,” “White House Down,” “Pacific Rim,” and “The Lone Ranger.” These films featured big stars, bigger explosions, and top-notch special effects—exactly the sort of summer spectacle that ordinarily assures a solid run at the box office. Yet for whatever reason, all of them failed to draw in the massive audiences needed to earn back their gargantuan budgets

For whatever reason? There's something lacking among these 6 movies that most summer blockbusters: none of them has the number "2" or higher after it. Sequels drive show business during the summer months, and have for many years.

I mined Box Office Mojo's data of the top 100 grossing films from years 2000-2012 and picked out the summer movies from it. The results are not particularly surprising. The biggest hits by gross profit are listed below (dollar values in millions):

Film Gross Cost Release Profit P/cost ratio Sequel?
Marvel's The Avengers 623 220 5/4/2012 403 1.83 Yes
The Dark Knight 533 185 7/18/2008 348 1.88 Yes
Star Wars: Episode III - Revenge of the Sith 380 113 5/19/2005 267 2.37 Yes
Spider-Man 403 139 5/3/2002 264 1.90 No
Finding Nemo 339 940 5/30/2003 245 2.61 No
The Hangover 277 350 6/5/2009 242 6.92 No
The Twilight Saga: Eclipse 300 680 6/30/2010 232 3.42 Yes
Toy Story 3 415 200 6/18/2010 215 1.08 Yes
Shrek 267 600 5/16/2001 207 3.46 No
Transformers: Revenge of the Fallen 402 200 6/24/2009 202 1.01 Yes

60% of them are sequels, which is even more significant when you realize that only 33% of summer movies are sequels.

Here are the biggest summer bombs:

Film Gross Cost Release Profit P/cost frac Sequel
Battleship 65 209 5/18/2012 -143 -0.69 No
Prince of Persia: The Sands of Time 90 200 5/28/2010 -109 -0.55 No
Final Fantasy: The Spirits Within 32 137 7/11/2001 -104 -0.77 No
Stealth 32 135 7/29/2005 -102 -0.76 No
Poseidon 60 160 5/12/2006 -99 -0.62 No
Robin Hood 105 200 5/14/2010 -94 -0.47 No
The Island 35 126 7/22/2005 -90 -0.72 No
The Sorcerer's Apprentice 63 150 7/14/2010 -86 -0.58 No
Around the World in 80 Days 24 110 6/16/2004 -85 -0.78 No
Green Lantern 116 200 6/17/2011 -83 -0.42 No

Not a sequel among them. Being a sequel is fairly well correlated with success at the box office


The obvious question is: why the hell aren't they releasing more sequels this summer? The answer seems to be at least in part that they've wrung most of these franchises dry. Taking a sample and looking at how the franchises are doing shows that many are flagging:


Some of these franchises aren't even turning a profit anymore (Pirates of the Caribbean and X-Men, for instance, are in negative territory). Others just aren't turning much of a profit compared to their budgets.

Anyone who can see that many of the popular franchises have been depleted, saw a slate of new IP movies coming, might have predicted a bad summer. The standard deviation on profit/cost ratio for sequels is only 0.83, but for new IP it's 7.7! There's a lot of uncertainty when you release a non-sequel, and that means you can have a big, unexpected bomb. And it's happening.

Jul 10 2013

Netflix streaming still a crap deal compared to disc plan

Zap2It is out with a press release that Netflix streaming will now be expanding its vast collection to include ... a major network show that's been running for almost 9 years.

Netflix, Inc. today announced an extension of its multiyear licensing agreement for select CBS library content streamed instantly to Netflix customers. Under the revised deal, new titles such as “L.A. Complex,” “4400,” and “CSI: NY” will be available to Netflix subscribers.


The fact that Netflix, five years into its streaming foray still has to fight tooth-and-nail for TV series which are naturally and simultaneously available on its disc service shows just how little progress the company has made. If I look at the list of shows I've enjoyed that are available on the disc service, cross referenced with IMDB's list of shows available on streaming, a little less than 39% of them are actually available (list below). Now, I don't know if that represents the entirety of all TV shows, but since that data isn't available (that I can see), this will have to do. And it's dismal.

One major problem with the streaming service is that HBO and Showtime refuse to participate, and a lot of great shows in this golden age of TV were or are on those services.

As bad as the figure for TV shows is, for movies it's far worse. Forbes did a quick look at the top 100 movies from 2010 and found only 8 that were streaming, with a similar number for 2011. After Netflix's deal with Starz expired, and MGM pulled a large number of titles recently, it's not hard to see why.

But 100% of the movies from that 2010 list are available on disc. So are 100% of my TV show list.

Then you consider the quality of the stream. Netflix uses 475 kB/s encoding for their movies. For a Blu-ray disc the number is more like 4000 kB/s, a factor of 10 better than Netflix's stream. Not that I blame them: Netflix already accounts for a third of all US internet traffic during the evenings. But you're paying for quality, and the disc is far better. Even DVD is 2 to 3 times as much data per second (although it uses an inferior compression algorithm, so this may mitigate the effect).

Netflix allows me to have a disc-only plan, and it's pretty cheap. I don't have streaming because, well, it's a shitty deal. I hope they fix it, but it's not in their control really.


Show Available on
24 Yes
Angel Yes
Arrested Development Yes
Babylon 5 No
Battlestar Galactica Yes
Beavis and Butthead No
Big Bang Theory No
Breaking Bad Yes
Brisco County Jr No
Buffy Yes
Bullshit No
Burn Notice Yes
Californication No
Castle No
Chuck No
Dark Angel No
Dead Like Me No
Dexter No
Elvis Costello Spectacle No
Extras No
Farscape Yes
Felicity Yes
Firefly Yes
Foyle's War Yes
Freaks and Geeks Yes
Fringe No
Game of Thrones No
Holmes on Holmes No
Homicide Life on the Streets No
Human Target No
In Treatment No
Inspector Lynley No
Invader Zim No
Jericho Yes
Keen Eddie No
Lost Yes
Louie Yes
Luther Yes
Mad Men Yes
Miracles No
Monk Yes
Monty Python's Flying Circus No
Moonlighting No
Murder One No
My So-called Life Yes
Nowhere Man No
Prime Suspect No
Prison Break Yes
Psych Yes
Reaper Yes
Red Dwarf No
Remington Steele No
Reno 911 Yes
Sex and the City No
Sherlock Yes
Six Feet Under No
Star Trek No
The Larry Sanders Show No
The Mentalist No
The Prisoner No
The Singing Detective No
The Sopranos No
The Wire No
Twin Peaks Yes
Undeclared Yes
Veronica Mars No
Wonderfalls No

Older posts «

» Newer posts