Saturday, March 29, 2014

The Earthquake Channel: Who's in?

California has had two moderate earthquakes in the last week and many aftershocks from the most recent one that was centered near La Habra. Earthquakes in Southern California have been more frequent recently, which leads residents to consider whether the next earthquake will be the "Big One." Frankly, EVERY earthquake makes Californians worry about the Big One.

Earlier today, someone that I follow on Twitter suggested that earthquakes should be given names, like hurricanes. That got me thinking. The Weather Channel now names storms, not just hurricanes. If we're going to name earthquakes, why not create a cable channel for them: The Earthquake Channel? Spoilsports might suggest that there aren't enough earthquakes, and we still don't know how to predict them in advance. In fact, there are plenty of earthquakes; the USGS reported 42 worldwide in the last 24 hours with a magnitude 2.5 or greater, of which 26 were located in North and Central America. As for not being able to forecast them, that just adds to the excitement.

One of the problems with reporting earthquakes is that it would be nice to have someone like Jim Cantore standing in a vulnerable location (next to an unreinforced masonry wall, for example) when the earthquake hits, to give live shots of the earthquake's impact. Unfortunately, since we can't predict earthquakes with any reasonable level of accuracy, we would have to have people standing all over the country, 24 hours a day. While it would be an interesting exercise, it would be both expensive and extremely boring. So, the next best thing would be a national network of cameras hooked up to seismometers; seismic activity above a certain threshold would start the camera recording, and would also capture the data from the seismometer. In some locations, the camera could be pointed at an animatronic dummy positioned in a vulnerable location (with a robotic voice saying "THIS...IS...JIM...CANTORE...AND...I'M...AARGH!"); in less important locations, the cameras could be pointed at chandeliers or glasses of water.

Some party poopers may raise another objection: How can an earthquake channel have enough content to run 24 hours a day? If you've watched any cable news channel ever, you know that's not a problem. CNN has gotten several weeks of content from Malaysia Airlines Flight 370 with almost no solid information. When things got slow, CNN speculated about "supernatural causes" and black holes. Just imagine what we could do with earthquakes.

Earthquake coverage would cover the following stages:

  1. Fear-mongering: Here's the fault map for your local area. Any of these babies could break loose at any moment and trigger a devastating earthquake. Here's what you should do to prepare. Here's what you should do in the event of an earthquake. Here are the places that you shouldn't live. Is your earthquake insurance paid up? You DO know that your householder's insurance policy doesn't cover earthquakes, don't you? DON'T YOU?
  2. OMG, there's been an earthquake: Massive overcoverage of the earthquake. Interviews of people who felt the earthquake. Interviews of people who know people who were told things by people who felt the earthquake. Experts to discuss where the earthquake occurred, how long it lasted, which faults were involved, and how many aftershocks are likely. Throughout all of this, show every available video of the earthquake you can find. If there aren't any, show cheesy computer-generated simulations of what it might have looked like had there been a camera.
  3. How do they recover: Experts discuss how to repair the damage, and when the next earthquake might occur. Say how sorry we feel for the people hurt by the earthquake, while implying that they're idiots for not properly preparing.

Okay, all three stages involve fear-mongering...and that brings me to my last point: Who would advertise on The Earthquake Channel? There's a surprising number of potential advertisers:

  • Contractors to strengthen building against earthquakes and perform earthquake repairs
  • Hospitals and urgent care centers to provide emergency medical services
  • Moving services to help people relocate away from earthquake country
  • Cars for getting out of Dodge quickly when an earthquake hits
  • Bottled water for when water pipes are broken
  • Freeze-dried, powered, shrink-wrapped, bubble-packed survival foods
  • Small arms, rifles and ammunition for when society breaks down and anarchy reigns supreme
  • Insurance--lots and lots of insurance
So there you have it: The Earthquake Channel. Any takers?

Tuesday, March 25, 2014

Facebook buys Oculus VR: What now?

Earlier today, Facebook announced that it had acquired virtual reality headset maker Oculus VR for $2 billion, in the form of $400 million in cash and 23.1 million shares of Facebook common stock worth $1.6 billion. Oculus VR's shareholders could earn an additional $300 million based on achieving several unspecified performance objectives. Oculus started with a Kickstarter campaign in 2012 that raised $2.4 million; last year, the company closed A and B rounds worth a total of about $93 million.

Ever since the announcement, comments about Oculus's acquisition have been flying around the Internet. The general tone is that Facebook will destroy Oculus's independence and load its software up with links to Facebook content. Some game developers have gone so far as to say that they've either stopped or won't begin software development for the Oculus Rift. My take is that Facebook's acquisition will be good for Oculus and its third-party developers, although a few years down the road, Oculus's shareholders may wish that they had kept ownership of the company.

But first, why did Facebook acquire Oculus? Facebook's not a hardware company, and when it's tried to specify smartphone hardware or design smartphone user interfaces, it's come up short. It'll take years for Oculus to become a meaningful generator of revenues and profits for Facebook, if it ever does. However, Facebook believes that VR is the next big hardware platform after mobile devices. VR is Facebook's "blue ocean," an undeveloped market opportunity that could be the basis of tens of billions of dollars in revenues. And, Facebook is buying the recognized market and technology leader: Even though there have been VR devices for almost two decades, the Oculus Rift is the first device that can truly begin to exploit virtual reality without all the physical discomfort that's come to characterize VR systems. As I've written previously, Oculus is six months to a year ahead of Sony and other VR headset manufacturers. And, Oculus is already generating revenue: Some 75,000 developers kits have been sold, at $300 each. Most other VR developers have only dozens or a few hundred developers working on their platforms.

As for Oculus and its shareholders, Facebook's $2 billion represents a nice return on their investments (Spark Capital and Matrix Partners are said by to both have received a 20-times return on investments made from a year to just a few months ago,) although I believe that if Oculus had remained independent and had become successful, the company's valuation would have easily been ten times what Facebook paid. However, Facebook brings to Oculus enormous financial resources. The company will be able to expand its development facilities and hire more aggressively. It will also have more influence on component manufacturers, because Oculus will be able to commit to much bigger purchase quantities and bigger non-recurring engineering costs than it could before Facebook. This is critical, because Oculus doesn't have the ability to build its own displays, camera and sensors, and some of the components it needs simply aren't yet available as off-the-shelf products.

Oculus's hardware development is likely to dovetail nicely with Facebook's rapid software development philosophy. At Sony's announcement of its "Project Morpheus" at GDC, the company said that it's been working on VR longer than Oculus, but most observers would agree that Oculus is still significantly ahead of Sony. The combination of Oculus and Facebook should be able to iterate hardware and software faster than Sony (and most other big companies) can even imagine. So, Oculus will most likely be able to get its first consumer version of the Oculus Rift out faster, and possibly at lower cost, with Facebook's backing.

But what about Facebook's influence on Oculus? After all, Facebook is hardly a benign presence. In Boing Boing, Dean Putney wrote a scathing takedown of the Facebook-Oculus deal; here's a quote:
The problem isn't that Facebook is going (to) ruin Oculus, by plastering it with ads and making it a pain in the ass like everything else they've shat all over. Although that wouldn't be a surprise. The problem is that this was an opportunity for something different. And it just died.
I don't think that the acquisition is as "soul crushing" as Putney believes; Facebook has said that Oculus will be run independently, and its offices will likely remain in Orange County, at least for a while. In addition, we're unlikely to see the consumer version of the Oculus Rift until next year, and that gives competitors an opportunity to catch up. However, paradoxically, the same funding that makes the Oculus Rift less risky as a development platform increases the risk that Facebook will turn it into a platform that you wouldn't want to develop for. I don't have a good counterargument to give to the people who believe that Facebook will "shit all over" Oculus, other than you should be okay if you develop for multiple headsets, not just the Oculus Rift.

Sunday, March 23, 2014

Games are just the tip of the VR iceberg

At last week's Game Developers' Conference, Virtual Reality (VR) was front and center, led by Oculus VR's second (and likely final) developer's kit, and Sony's prototype Project Morpheus VR headset. At Sony's launch event, a company executive noted that there are a number of applications for VR, but that games are the biggest opportunity. Games are an obvious application for VR; existing 2 1/2 D games, such as first person shooters, can be modified to enable the player to be immersed in the field of play using a VR headset. However, while games may be the first application for VR, in the long run, they're not the biggest application.

VR, which places the participant in an immersive environment, has the potential to recreate real-world environments much more successfully than technologies such as CAVEs that project an environment on multiple walls of a cube in which the participant stands. VR can be used for a variety of simulations; for example:
  • Military and police training, where recruits learn how to differentiate friend from foe and both how and when to engage with an enemy or suspect.
  • Medical training, where medical students can perform tests and try out surgical techniques.
  • Flight training for pilots and trainees.
There are also many applications in education. Students can be transported into ancient Rome or Greece and participate in activities as the people who lived there did. They can run science experiments that would be dangerous or deadly if they were to do them in real life, and they can observe processes at scales that would otherwise be impossible (from sub-atomic to galactic.) Students can do these things today with PCs and tablets, but the addition of VR makes them much more visceral, which should help students to learn more and retain more of what they learn.

There's also an opportunity for VR in television and movies. We think of these as passive media, but VR has the potential to immerse the viewer in the action to a much greater degree than has been possible with 3D. However, there are many questions that need to be answered, especially for live (non-computer-animated) content:
  • Can the viewer participate in the action, or are they merely an observer?
  • How much freedom does the viewer have in participating--can they move freely, or are their movements limited?
  • How do the characters in the production interact with the viewer? Can the viewer's actions change the production's course of action? (For example, in an action/adventure production, could the viewer's character kill the hero? What happens next--roll the credits?)
I believe that ultimately, these applications with be much bigger than gaming. However, gaming technology will be very important for creating these applications, and game developers will be in an excellent position to be early developers of these applications. Gaming engines that have been adapted to support VR will be the basis of the first generation of new applications, while production and development systems built specifically for VR will come along later. In short, VR's first big opportunity will be gaming, but the application set and market opportunities will ultimately be much bigger than gaming alone.

Wednesday, March 19, 2014

Oculus VR releases DK2, its second VR headset/software development kit

Just in time for GDC, Oculus VR has released its second Oculus Rift development kit, called DK2. Engadget got a look at the new hardware, and it's a mild improvement over the Crystal Cove prototype shown at CES in January. Here's a summary:

  • Oculus now supplies its own user-facing IR camera to track the headset's position.
  • It's made the IR-reflective dots on the headset used for position tracking, which were visible on Crystal Cove prototype, invisible and harder to damage.
  • There are USB 2.0 and headset ports above the user's left eye.
  • A single cord breaks out to connect the headset to a computer's HDMI and USB (power) connections.
  • The display resolution is 1920 x 1080 (1080 x 960 for each eye.)
Oculus has priced DK2 at $350, $50 more than the original development kit. The company notes that the the final consumer version of the Oculus Rift will have higher resolution displays, a faster refresh rate and lower latency, and will be physically lighter than the DK2, but it's significantly closer to the final product than the original development kit was.

Even with the announcements from Sony and several other VR headset vendors at GDC, it's likely that Oculus is still significantly ahead of the pack. However, the company's one potential weakness is that it's dependent on third-party manufacturers for much of its technology--displays, sensors and cameras. For example, DK2's display has a 72 Hz refresh rate, which is as fast as Oculus could find in a 1920 x 1080 OLED display, but it needs at least a 90 Hz refresh rate to eliminate visual artifacts that can cause motion sickness. Sony manufactures image sensors and many of its own displays, so it may be in a better position to source the necessary components. 

Tuesday, March 18, 2014

Sony's "Project Morpheus": VR for the PS4: Good start, but far from prime time

Last night, Sony announced "Project Morpheus" at an event at the Game Developers Conference. It consists of the PS4 console, the PS4 camera, Sony's existing Move controller and a new VR headset. The presentation was spun by Sony as a major technology announcement, but in fact, it was largely a slimmed-down version of presentations over the last few months by Oculus Rift and Valve. While both companies have been quite specific about what they've found to be necessary in a VR system (not just the headset, but the sensors and content as well,) Sony stayed away from hard numbers in its presentation. For example, when asked what Project Morpheus's refresh rate is, a Sony executive replied "As fast as possible."

Engadget has gotten a chance to try out Sony's VR headset at GDC, and they found it to be not bad for a first try, but far from Sony's own goals. For example, in Sony's presentation, the company said that comfort is one of its most important objectives, but the Project Morpheus headset requires multiple manual adjustments, and when it's in the correct position, it's so tightly sealed that sweat fogs up the lenses. Screen resolution is lower than Oculus VR's new DK2 developer kit (see next post.) Vision blur is a significant problem, which suggests that Sony isn't using a low-persistence display. In addition, Sony's headset offers a 90 degree diagonal field of view vs. Oculus Rift's 110 degrees. On the other hand, one advantage that the Sony headset has over Oculus Rift is full-body tracking, but there are a lot of companies working on that.

My take is that this is Sony climbing on the VR bandwagon with a very incomplete solution. The expectation has become that if you're serious about gaming, you have to be working on VR, so Sony checked off that box. However, from Oculus's release of the DK2 earlier today, I'd say that Sony is six months to a year behind with its headset. As for audio, Sony says that it makes good headphones, but so do many other companies, and headphones aren't the problem. Will Sony's existing Move controllers and PS4 camera really work well in a VR system? I doubt it. Those devices weren't designed for VR, and I doubt that Sony put in the costs necessary to make them work at the speeds and with the accuracy needed for VR.

What we're likely to end up with is a suite of devices that Sony purpose-builds for VR--and it's not at all certain that even the PS4 will have the refresh rates necessary to drive Sony's final VR headset.

Wednesday, March 12, 2014

It's tough out here for a camera (maker)

Over the last seven years, we've seen a number of companies jump into the professional cinema camera market. Some have had an easier time of it than others. Red launched its first camera, the Red One, at NAB 2006, and a small number shipped to customers more than a year later. Red announced its Scarlet camera in late 2008 but didn't ship a much different (and more expensive) Scarlet model until about three years later. Canon's first cinema camera, the EOS 300, shipped in January 2012, and the EOS 100 and 500 followed in the fall of 2012. Blackmagic Design launched its Cinema Camera at NAB 2012 but didn't start shipping in quantity for almost a year; Blackmagic's Pocket Cinema Camera was announced at NAB 2013 and started shipping small quantities in August 2013. Its 4K Production Camera, also announced at NAB 2013, has only just recently begun shipping in small quantities. Finally, there's the Digital Bolex D16, which started as a Kickstarter project in April, 2012 and started shipping to early supporters at the end of 2013.

Of Red, Canon, Blackmagic Design and Digital Bolex, only Canon shipped its cameras close to when they said they'd ship. All four companies shipped cameras with problems, but Canon's were minor. By comparison, Red spent two years debugging the firmware in its Red One and subsequent cameras, and Blackmagic's cameras are missing essential features, such as audio meters, an indicator of how much storage is available and the ability to format storage in the camera, that should have been there when they shipped. Digital Bolex's D16 shipped late, but so far, it seems to be working reasonably well.

It's interesting, but not surprising, that Canon was the only company that shipped its cinema cameras on time and with relatively few bugs. After all, Canon shipped its first camera in 1936, its first camcorder in 1981, and its first digital DSLR in 1995. They knew how to make or integrate everything needed for a Cinema Camera when they launched the EOS 300 in 2012. (Some of you may be wondering why I didn't mention Arri and its Alexa cameras, but the story is much the same as Canon: Arri launched its first film camera in 1924, and its first digital camera, the Arriflex D-20, in 2005.) Red and Digital Bolex, on the other hand, were founded specifically to make cinema cameras, while Blackmagic Design made an extensive line of video hardware and software before gearing up to make cameras.

My point, which I admit it's taken a long time to get to, is that it's hard to make a good cinema camera. You can't simply go through parts catalogs and find all the pieces you need to build a camera; in some cases, you can't even depend on a parts manufacturer to build the parts you need to your specifications. Blackmagic ran into problems with CMOS imagers on both its original Cinema Camera and 4K Production Camera. Both Blackmagic and Red ran into integration problems--getting all the parts to work together as designed--and both struggled with firmware needed to implement promised features. Blackmagic's cameras use too much power--third-party tests show that its 4K Production Camera can burn through its internal battery in 20 minutes.

I'm not saying that any of these companies shouldn't have entered the cinema camera market; Red changed users' expectations for price and imaging quality, and Blackmagic pushed prices far below where they'd been for comparable cameras from competitors. However, they (and their customers) have learned how hard it is to build a good cinema camera. Given what's happened since 2006, I'd say that other than lenses, cinema cameras are the hardest broadcast or cinema products to design and build.

As we prepare for NAB 2014, we're likely to see a new round of camera introductions. When a company with no previous camera design or manufacturing experience announces a new cinema camera, my recommendation is to not be one of the first customers in line with a preorder. To be safe, you should add at least a year to the company's announced release date, and probably a year after that for the camera's firmware to be stabilized and for all key features to be implemented. The frustration you save will be your own.

Thursday, March 06, 2014

GoPro: Succeeding by sticking to its knitting

Let me introduce you to Bigbird, a juvenile pelican that was orphaned and abandoned by its flock, and washed up on the beach at the Greystoke Mahale resort in Tanzania. The resort's staff took care of Bigbird--which included teaching the pelican how to fly. The resort's team of "pelican trainers" mounted a GoPro sport camera on Bigbird's beak, and the camera captured the bird's process of learning how to flap its wings while simultaneously running down the beach. Most importantly, it captured Bigbird's first flight.

That video, which was taken by the Greystoke Mahale's staff, was edited by GoPro, and you can watch it on YouTube. You can also watch two deer be rescued from a frozen lake in Minnesota, or a kitten rescued from a burning building and resuscitated by a Fresno, CA fireman. If you can keep from crying after watching any of these videos, you're stronger than I am. If you want something a little more exciting, you can watch people snow skiing, snowboarding, water skiing, surfboarding, skateboarding, mountain biking, racing superbikes, racing cars, and much more, all from a first-person perspective. In fact, if you can name a sport or outdoor activity, it's probably been captured by a GoPro camera and can be found by searching on YouTube.

GoPro is arguably the most successful camera company in the world--they've certainly done more with fewer models than anyone else. The company exclusively makes sports cameras--small, rugged, inexpensive video cameras that nevertheless produce high-quality images. At any one time, GoPro only sells three models; currently, they sell two versions of their top-of-the-line Hero 3+ cameras, and they sell one model of their previous-generation Hero 3 line as their entry-level model. Prices range from $199 to $399. In addition, GoPro sells a range of accessories that allow the cameras to be mounted just about anywhere, and to be protected from just about any conditions (short of placing them in an oven.) The accessories for the Hero 3 and 3+ series are identical, so customers who purchased Hero 3 models can upgrade to a Hero 3+ and use all of their existing accessories.

By comparison, GoPro's biggest competitors--Sony and JVC--sell sport cameras as a sideline. For example, Sony sells full lines of point & shoot cameras, DSLRs, ILCs and camcorders, not to mention its line of professional camcorders and cinema cameras. And, in there somewhere, it's got two sport cameras. JVC is in a similar situation; its two-model ADIXXION line of sport cameras are a small part of a much bigger line of camcorders. GoPro has been able to keep ahead of its competitors, both big and small, with its laser-like focus on features that are of value to its target market of active sportspeople. Sony and JVC may each have a single product manager focusing on that market, while GoPro's entire company is focused there.

One way that GoPro keeps on top of its customers' needs is to watch videos that they make--and it gets thousands of them. Most of the videos featured on GoPro's website and YouTube site were produced not by GoPro but by its customers. So many people are producing videos with their GoPro cameras, and the videos are so popular (especially after being edited and enhanced with music by the GoPro team) that the company is launching its own branded media channels. It provides an in-flight video channel to Virgin America, and plans to launch a video channel on Microsoft's Xbox One and Xbox 360 game consoles this summer. A GoPro-branded media channel separate from YouTube is likely in the near future; such a channel would enable GoPro to fully monetize its videos, and cement customer loyalty by sharing advertising revenues with video submitters.

Last month, GoPro made a private IPO filing with the U.S. Securities and Exchange Commission. We don't yet know how much money the company plans to raise or how much of the company it plans to sell, but most observers believe that the IPO proceeds will, at least in part, be used to fund new products and the development of GoPro's media channels. By focusing on a small slice of the camera market, GoPro has built a big business and a cadre of loyal customers that's producing the content for a whole new business opportunity.