Archive for category Uncategorized
It’s amazing to see everyone getting involved in VR development. This used to be relegated to a small set of academics and foolhardy startup veterans like me.
I especially can’t wait to try the Vive in conjunction with what they’re here calling environment mapping. It’s the closest thing we have to high quality AR right now.
I have a few nits with the article though, based on some 20 years experience with VR. Most of these made it into the Metaverse Roadmap doc.
Here’s an old glossary of VR that summed up the research as 10 years ago (when I wrote it). In general, I think folks are overloading “presence” too much and only just beginning to grok what else is missing.
What’s really going on is a virtuous cycle that starts with Presence, requires Interactivity, such that people can affect the world. The extent to which the world is mutated by your interaction or mere presence is called Reflectivity. This includes just seeing your own body and seeing it move the way you imagined, or vice versa (also called proprioception, which is greatly hindered by having opaque goggles on). If we get this far, we can develop a kind of Resonance (a self-reinforcing signal) that builds with each positive interaction in this cycle. OTOH, every time this cycle is broken, like when you put your hand through a wall, or something doesn’t move when you expect it to, it degrades the resonance and erodes the effect, sometimes severely.
When it works, we get a two-way synchronization between the entirely computational virtual world inside the computer and the entirely mental virtual world in your brain. This mimics what happens naturally, when we model the real world and our models are validated or refined by each interaction.
It’s not much more complicated than that.
But for those working on latency, resolution, etc.., that’s only the first 25% of the job. Physical interaction is also key, eventually haptic. But so is making the virtual world mutable enough for participants to impact it, so their interactions are credible across low and high-level systems in our brains. Sticking to how things work in nature is always a good starting place.
Currently, the most common testing technique (for men) is what’s known as penile plethysmography. This involves placing a ring-style sensor around the offender’s penis, then measuring any changes in its circumference as they’re subjected to a variety of visual or auditory stimuli. One problem with this approach is that subjects can skew the results by diverting their eyes from the images.
Holy Clockwork Orange, Batman. I can understand the wish to determine if sex offenders are likely to offend again, when determining their parole. But what we’re talking about here is effectively using VR to enable unavoidable thought crime, and entrapment thereof. Look away from the virtual child and you can still be guilty, because what we want is not what you did (legally or otherwise) but what you will do, aka your latent intent, even if you don’t consciously know it.
Nicely done, but I can’t figure out how to embed properly:
The only real conclusion one can draw from this article is that marketers are really excited about VR’s ability to attract attention. Here are 7 reasons to think harder:
1. “Movie theaters” full of HMDs are unlikely (even ignoring hygiene & robustness issues). The economics don’t make sense for the equivalent of having 300 people watch the same thing on one expensive big screen.
Even factoring in the cost of a new PC or console, we’d more likely see the equivalent of “internet/game cafes” for those who can’t afford their own VR setup at home (plus rentable airline equivalents), more as a niche and trailing edge.
2. Hollywood + VR movies already exist. They’re called games. Now, game developers generally put substantial effort into making their cinematic intros and cut scenes. But even with higher production values, most people watch these a few times and then skip right to the gameplay. The gameplay must be better than the intro movie or the investment will only succeed on YouTube, if at all.
Take-away: interactivity is key.
3. Physical interactivity in VR is not yet ready for prime time. Reason being: the closer you get to the human body for sensing our movement, the more proprioceptive skill we have and the less tolerant we are of noise and other errors.
So in the short-term, the level of interactivity will range from “almost none” to the equivalent of a gaming controller. Designers have to work around those limitations. A good example is to skip touch entirely and use voice to control things.
4. In UX research, we found that people’s levels of comprehension of things like story and character in VR were very poor, probably due to information overload and not knowing where to look for cues. Movies solve this by leading the horse to the water, so to speak, with expert cinematography and more. So the chances of a subtle cinematic narrative are slim until we develop those muscles in VR over many years.
Think more “TV Soap Opera” than “Gosford Park.” And in terms of Presence, think more “Saving Private Ryan” than “The Man from Earth.”
5. Movies are an inherently social experience, esp. going to the theater (which we said isn’t helped by VR). Perfect, you say, because Face/Rift is a social network. Actually, FB today is more a social experience of last resort. It is most social when you don’t have a better way to interact. Just imagine a group of six friends hanging out, noses down, all browsing FB on their so-called smartphones. I know it happens a lot, and they certainly think they’re being social, but who believes it? It’s at its best when it’s connecting people who won’t otherwise see each other.
6. VR movies will initially be more of solo experience that we can talk about and retroactively construct the social element, like talking about the latest episode of “Lost” or “Game of Thrones” the next day. We can feel like we watched them together for some “social backfill.” I’m guessing that the more presence we feel in the VR experience, the harder it is to later backfill in those missing friends, but the more we’ll want to try, leading to more of the feeling that we’re losing real human connections by going so virtual. Prediction.
7. Someone will therefore add avatars to these immersive VR movies to solve this. Good thinking. If captured with high fidelity, this will be a little closer to the quality of being together in person, and there’s always the cool new immersive milieu to explore together.
But here’s the dilemma. If your movie is interactive, you have to solve the holy grail of immersive interactive 3D storytelling, which the fictional Holodeck didn’t even get right. Tony and Tina’s wedding (the interactive play) is probably our best model, but that’s all about the actors making it work.
If your movie is not so interactive but you still add friends and family to the scene, the greater degree of presence ironically makes it more awkward to see them unless they’re transformed into the story, a stark reminder that you’re not actually there. It’d be good for Jurassic time travel, but not so great for Star Wars, where seeing my mother standing next to Darth Vader would change the experience a bit.
Not surprisingly, VR will likely work better for participants who are more physically remote than in the same room — exactly like FB does today. It adds to social interactions where distance makes it harder, but caps it where real proximity would make it easier. It’s no wonder FB likes this view of the future.
For Hollywood, it’s about the business of monetizing attention on one level, and the art of storytelling on another. On second thought, maybe they’re not that different after all.
I have to admit, even with 25+ years experience with computer graphics, on first viewing I thought The Lego Movie was mostly done with stop action photography.
I figured maybe 80% physical and 20% virtual. Turns out it was closer to 99% CGI with some real legos thrown in for good measure. Other than the live action scenes, I couldn’t tell you where the real legos sat.
There were some things, like the water, explosions and more, that looked way too procedural to be done by hand. But still the rendering, shading and animation were so close to perfect, so physically correct down to sub-surface scattering and extreme depth of field, that it was almost impossible to tell.
Amazing job. And especially impressive given how well they could tell the story without relying on the usual tricks of animation and CG, staying true to only what real legos can do.
The real tip-off about the CGI was in the lighting, which allowed for certain legos to emit light or light to come from no actual source. That would be pretty hard to do in reality without a really complex effects pipeline on top.
Here’s a longer video that explains how it was done:
This is a brilliant idea and striking demonstration of the effects of rising sea levels.
It has only one small problem. It doesn’t take altitude into account. It shows the same nicely rendered water level no matter where you go.
I don’t think it would be hard to do a lookup of the altitude of any address and move the water level up or down. But given the very rough depth map from Street View and the apparent lack of an “up” vector, it might be hard to properly intersect the water with the scene.
Here’s a more accurate example, without the nice immersive visuals: http://geology.com/sea-level-rise/
Still better than trying to depict a 1000ft water level near the top of the Oakland Hills. If that happens, we’d be long gone.
For at least 20 years, I’ve been telling this to anyone who might conceivably take this idea and run with it as a business. Form a real-time sphere (or cubemap) of video around an airplane using an array of cameras. Give anyone who wants one a VR HMD and index their chosen POV into said sphere of video. Viola. Invisible airplane, at least to those wearing the HMDs. It would be a much better way to pass several hours in flight than watching movies IMO.
Anyway, alas, if I want to do this today, I’d need to drive a tank. Tak, Norway.
Scoble’s love affair for Google Glass could apparently only last so long. This underscores some of the problems with developing a product out in public, or at least half-way out. Long-lead technical challenges (battery, size, cost) are still hidden below the core design and marketing issues (utility, fashion, desire).
Google at least did a good job of starting the conversation, even if they haven’t yet figured out how to finish it.
Down in the comments, he elaborates on why he doesn’t like them any more and what he thinks Google did wrong:
1. They launched it with WAY too much fanfare that the product simply can’t live up to. Jumping out of a blimp and doing live video on the way down (along with a ton of really well produced videos that promised it would be a great assistant as you walked around) set expectations VERY high and the product hasn’t gotten to that fit and finish yet, even two years later. The product, for instance, still doesn’t do live video (at least not by itself).
2. The team started out very public, with very nice collaborative team members. Then they turned secretive and can’t tell us what’s coming. It’s almost like someone told the team “turn Apple.” That secrecy made developers and influencers feel like they were no longer part of the process of developing this into a real product, which it still is not.
3. Because it was launched at a developer conference (Google IO) expectations were set that this would be mostly for developers, and that a great API would come soon. The API did come, but 18 months later. There still is no real store. The UI is way too simplistic to handle dozens or hundreds of apps. Google refused to answer questions about sensors at the next Google IO.
4. Two years ago the price was announced as $1,500. Today the price is still $1,500. How many tech products stay the same price for two years? Not many. The pricing is now at the point where it seems just wrong. The team even admits it’s artificially high to ensure “only people who really want one get one.” That creates weird distortions in belief about the product too.
5. When people got theirs a year ago we expected a ton of REAL updates to both UI and functionality. The updates we’ve gotten just haven’t met expectations. Compare my drone, which lets me take a photo, see it on my iPhone, AND post it to Facebook and Instagram WHILE IT IS FLYING to Glass. I still can’t post photos to Facebook or Instagram without plugging the Glass in and putting them elsewhere (er, Google+) first.
6. We expected new designs by now and updated electronics. I’m holding out hope that we’ll see a much better design (battery life sucks, it cuts my ears, it is starting to look quite dated, video is poorly compressed, etc etc) by now.
When I say this was launched poorly Google set way too many expectations on this product and it has failed to meet them. It should have launched much quieter and set expectations that it was only for vertical markets. Then as those showed up, they could have expanded expectations. If Google had done that then the early adopters could have said “hey, these are designed for use in very specific places, like surgery rooms.” That would have kept people from freaking out.
The other thing that sucks is that many of the explorer program members I’ve talked with are tired of being asked to demonstrate them. Google is getting US to pay to test, provide PR, AND demonstrate them to the public. THAT is NOT empathetic at all and is actually quite arrogant on Google’s behalf (a company that makes billions of dollars every year).
It’s not shocking to me that most Google employees I know that have them aren’t wearing them around anymore and that many in the community are grumbling behind the scenes (most won’t write about their concerns because they need to have good relationships with Google).
Give it a try. Build with Chrome.