does anyone know the HW specs needed to develop and use apps for Meta2?
UPDATE: HW requirements have been posted!
Intel Iris Pro / NVIDIA GT 650M / AMD Radeon HD7970 (recommend NVIDIA GTX 960 / AMD 280 equivalent or greater) Intel i7-3610MQ equivalent or greater 8 GB RAM HDMI 1.4 video output 1x USB 3.0 ports Windows 8.1 or newer 64 bit Unity 5+ on Windows (for development, not end users) Intel HD compatible sound card USB 3.0.
They make it sound like there's no native SDK, only a Unity SDK. That seems like a mistake if it's true. Unity it's great for prototyping and for certain kinds of applications but as a current VR and possibly future AR developer I want the option of a native SDK, not least for access to other engines like Unreal.
In other words, not a breakthrough in regards to the readability of text. I'm not saying text will be unreadable. I mean, I do run a project that has a significant text-editor component in VR, and I do this on the DK2, which is significantly lower resolution. But I think that also means I'm one of the most qualified people to say that it's not going to be "like unlimited monitor space". For the application support they are touting, this is going to be a bit like the old 800x600 days. Now, we certainly got work done in that time. And for a variety of reasons, the perceived resolution of text is higher than the actual resolution in an HMD, but it's maybe only a factor of 2. You're not going to be replacing your office with this device. We just aren't there yet. I'm working towards that day, and I think the work needs to be done now to be prepared for that day. But we aren't there yet.
Given that VR is already pretty taxing on modern, top-of-the-line hardware, I'm personally waiting until at least the next generation for an AR device.
No, it's because 90 degree FOV is just not that much and the resolution is not high enough for even that FOV to make for invisible pixels. I don't think we need to have completely invisible pixels to be productive in VR text work, but it will definitely be a completely different day when that is available. For now, this is "better than DK2", but not "better than GearVR". It's a step forward, but not objectively great yet.
The fact that all of the 3D acceleration hardware we have today is built for raster graphics, not vector graphics. Filling fields with a vector display is extremely expensive. There's a reason they died out for most uses.
I think VR devices will be sufficient to build the tools for the "office of the future". I'm saying skip AR devices and stick to VR devices this gen, as the added complexity of real-world object recognition is likely too much for current-gen hardware, when the 3D graphics for VR is already taxing and you need that for AR, too.
I loved everything about it until I saw this image and then I literally laughed out loud. I think the "you look ridiculous" factor is even higher than VR... Imagine going to a meeting with this on!
I'm not on a horse. Their website is extremely glossy and it's version 2 apparently; fair enough if you say it's not a consumer device.
I feel like the looking ridiculous factor is a perfectly reasonable fear to have and a big reason why I wouldn't buy it. Downvote away but it's a reasonable stance to take IMO and this clearly suffers from it more than VR.
Development Kit, not "release". This is the second prototype aimed at developers. And a glossy site doesn't mean the product is released :|.
Besides, while it is a fair point to make, people using cell phones looked ridiculous back in the days. And yet everyone has a cell phone now. I don't think the ridiculous-factor will be a major blocker if the benefits of the device outbalance it.
I guess you can hold out for contacts or an eyeglasses form-factor, but I can't really wrap my head around a technology enthusiast thinking "I'd love for my sensory perception of reality to transcend the physical world, but I don't want to look like a dang nerd."
And even if that's the case, it looks pretty sleek & sci-fi IMO
The interaction with the objects looks interesting and is something I haven't seen in similar product videos. Good content is obviously key but there should be some immediate use cases for business settings where you can sell development of the needed content (consumer, not so sure the content "needs to be there").
Ships Q3, I'll see if I can convince someone to get at least one for our HCI labs.
I used to work at Meta while they were located in Portola Valley.
Meta "applications" are highly customized Unity 3D scenes with a generous amount of custom GUI elements.
I would take everything in this video with a grain of salt.
Also, Meta has a terrible habit of copying logos. Their old logo was a blatant copy of MIT's "M" and their latest is eerily similar to Mazda's "M" series logo
> Also, Meta has a terrible habit of copying logos. Their old logo was a blatant copy of MIT's "M" and their latest is eerily similar to Mazda's "M" series logo
Not really disagreeing since you have more experience with them than I do but if you asked me to write a cursive M, it would probably come out pretty much the same as both. It's like the laziest logo.
Indeed about the logo. Is this by direction or just uncreative logo artist?
Saw the 'cribs' thing about them a while ago and I thought the setup it was interesting but my Wife commented on the 'cult' vibe. What was it like living and working there?
I visited Meta (they were trying to recruit me) about 2 weeks after they moved in there... I won't comment on what I saw there, or my thoughts on Meta, but I will say I take all augmented reality claims with a huge grain of salt. I would say that Meron purposefully is trying to form a cult of personality, and while it may be respected as a Steve Jobs-esque thing in the Valley, it is personally a turn off for me.
Funny thing is that the Portola Valley property is now rented out to a friend of mine, which is subleasing it out to tenants... The TechCrunch video really doesn't do it justice in terms of size, and sadly all but one of the tanks are gone. http://sfbay.craigslist.org/pen/roo/5464646558.html
Oh, so I remembered correctly that it was Meta whose CEO did that creepy tour of their shared villa, explaining how they all work all the time, because the location was too remote to do anything else. There was a very strong cult-like vibe there for sure.
Looks better than HoloLens and props for showing the actual footage trough the headset and not just some misleading illustration like Microsoft did, but the tracking needs some improvement because it looks annoyingly laggy and I cannot imagine using it for anything serious for a longer time.
Yeah I don't much care for that either. I'm guessing it's a byproduct of the screen being basically a piece of transparent glass as opposed to an actual "screen" or whatever.
This, of course, is the biggest problem with AR tech - you can only "add" light, not "subtract" it - meaning you can only make pixels brighter than they are, not darker. Essentially this will be unusable in bright daylight. Notice how all the promo videos are filmed in fairly dark rooms.
I was impressed that the footage was from iPhone pressed against the headset to get that video. I wonder if some of the jumpiness or supposed tracking is due to handling the headset/iPhone, and maybe some scan rate 'moire' patterns or other artifacts due to a video-of-a-video?
I just got approved for a HoloLens dev kit (like everyone else and their dog who applied) at $3000. This looks like a very similar product for 1/3rd of the price. I wonder what the differences are.
One is backed by one of the largest companies in the world and another is a startup. You'll definitely get your HoloLens, but you may not get your Meta 2 (scheduled in Q3 2016, but you know how that goes).
Competition is great though, hopefully they will push each other.
Yes, seems like a tough field to be in. You're up against not just HoloLens, but Magic Leap, which appears to be even scarier competition. Good luck to them, though.
What I'm really looking forward to is a virtual desktop, where I can work on a tropical beach with a very large high-res monitor in front of me (virtually).
But it seems that the resolution of the current VR goggles is not sufficient :/ This seems a bit weird given that "retina" displays are already quite mainstream.
>The high resolution needs to be only at the center of the field of vision. Outside the center, low res will do fine.
It sounds like you're describing foveated rendering[1] which requires that the headset also track where your eyes are looking, not just the direction of your head. It's a very exciting possible future direction for VR/AR but I think it's still a bit off (and likely still requires very high resolution miniaturised display for the foreseeable future)
> The high resolution needs to be only at the center of the field of vision. Outside the center, low res will do fine.
The problem with that idea for goggles is that the "field of vision" doesn't have a fixed center relative to the display, because eyes can move independently of the head (and do in normal human interaction with the environment.)
Obviously, the device needs to track pupils (well, lens, to be exact), not head. Sort of what ophthalmic auto-refractometers do, although not exactly (well, if it'd measure focal length it can solve the issue of lacking true DoF that modern 3D suffers from, yay!)
(On a slightly unrelated note: Is there any headsets with VRDs with deformable mirrors that do this? The idea's decades old, but I haven't ever heard about any devices like that on the market.)
Nah, that's not usable with screens. You surely can't swing a display matrix, trying to keep up withe eye. :)
But with projector (so-called VRDs, virtual retinal displays) and a deformable mirrors it's possible. At least this was discussed in theoretical literature, articles and various patents - I've read about this stuff a bit, some years ago.
Giving you the impression of a very large high-res monitor with VR goggles is quite a challenge, you'd need the displays in the headset to be significantly more pixel-dense than they are (remember that while "retina" quality displays are quite mainstream VR requires them to be very small and light and the black gap between (sub)pixels to be as thin as possible to avoid the screen door effect).
Say you had 4K per eye (I understand it's accepted that 8K per eye is more realistic), that's 8K or 16K driven at 90Hz, you would need some serious graphics horsepower to drive that.
Maybe they're running the leanest of lean operations, but given the challenge of actually making hardware, I'm not certain anyone will see a working version of this until the second Hololens is out.
He definitely is. However, I wouldn't consider it a plus or a minus. It is in my view neutral. His research doesn't involve the specific problems that Meta would need to solve nor is he a product design expert.
I'm not knocking the guy. I just don't see his involvement as adding much to the project beyond the name recognition.
Doing the math to compare resolution to the Hololens...
Hololens has "2.5K radiants," where a radiant is a "light point per radian." [1] A radian is ~57.3 degrees, so 2500/57.3 = 44 pixels per degree.
Meta has a "2560x1440 display with a 90 degree field of view." Horizontally, that's 2560/90 = 28.4 pixels per degree.
Having used the Hololens for 30 minutes at Build last year, the field of view was definitely less than 90 degrees. However, the tracking was perfect, there was no lag, and generally speaking the illusion was complete.
Watching the video at the top of metavision.com, it's hard to get a good impression-- some shots look like they pointed a video camera through the Meta glasses!? The frame rate is noticeably low and the gestures look like they... sort of work.
Both headsets' problems seem fixable-- Meta will presumably fix the lag and gesture accuracy (assuming those weaknesses in the video are real). And Hololens will presumably get a larger field of view with later hardware revs. To me it'd be a mistake to dismiss these systems as not being ready. It probably won't be a long wait.
I went straight to the video and about 1 minute in I was sure it was some kind of college humor parody. I guess startups are under a lot of pressure to conform to type expectations and PR companies will make them look a certain way in these promo videos, but seriously, it's so extreme now that it's just laughable. The head of engineering was the only normal person in the video. To make something like this there must be a core group of hardcore engineers, and I can't help but imagine them collectively eye-rolling in the face of all the vice presidents with their perfectly tousled hairstyles and oh-so emphatic TED-talk styled speech.
I watched a Video interview by Robert Scoble with the Meta CEO and they talked to a developer who actually used this as a monitor replacement to write code. Pretty interesting, but i still doubt the resolution is enough to really replace monitors yet, but the premise is very interesting. From what i understand their implementation looks similar to an effect called "Pepper's Ghost" just in a mobile form. I have seen this done using smartphones and some plexiglass but in a static form.
No, Meta Vision is a YC startup doing the headsets you see on their site. Metaio was doing AR for phones and things (and I think were bought by Apple).
Cool to see another company in the AR space! However, their videos that say "actual footage" make the holograms look jumpy and the worry is that could lead to fatigue or nausea. Has anyone tried one of these?
88 comments
[ 4.7 ms ] story [ 165 ms ] threadUPDATE: HW requirements have been posted!
Intel Iris Pro / NVIDIA GT 650M / AMD Radeon HD7970 (recommend NVIDIA GTX 960 / AMD 280 equivalent or greater) Intel i7-3610MQ equivalent or greater 8 GB RAM HDMI 1.4 video output 1x USB 3.0 ports Windows 8.1 or newer 64 bit Unity 5+ on Windows (for development, not end users) Intel HD compatible sound card USB 3.0.
SOURCE: https://www.metavision.com/faq
https://www.metavision.com/devcenter
Given that VR is already pretty taxing on modern, top-of-the-line hardware, I'm personally waiting until at least the next generation for an AR device.
What would stop them using rgb lasers to display vector graphics?
The fact that all of the 3D acceleration hardware we have today is built for raster graphics, not vector graphics. Filling fields with a vector display is extremely expensive. There's a reason they died out for most uses.
https://d3mzncm6cj8c3y.cloudfront.net/assets/pages/home/side...
I feel like the looking ridiculous factor is a perfectly reasonable fear to have and a big reason why I wouldn't buy it. Downvote away but it's a reasonable stance to take IMO and this clearly suffers from it more than VR.
Besides, while it is a fair point to make, people using cell phones looked ridiculous back in the days. And yet everyone has a cell phone now. I don't think the ridiculous-factor will be a major blocker if the benefits of the device outbalance it.
And even if that's the case, it looks pretty sleek & sci-fi IMO
As they were, actually.
https://en.wikipedia.org/wiki/Virtuality_(gaming)
Ships Q3, I'll see if I can convince someone to get at least one for our HCI labs.
Meta "applications" are highly customized Unity 3D scenes with a generous amount of custom GUI elements.
I would take everything in this video with a grain of salt.
Also, Meta has a terrible habit of copying logos. Their old logo was a blatant copy of MIT's "M" and their latest is eerily similar to Mazda's "M" series logo
http://www.miata.net/faq/mfield_files/image004.jpg
Not really disagreeing since you have more experience with them than I do but if you asked me to write a cursive M, it would probably come out pretty much the same as both. It's like the laziest logo.
Saw the 'cribs' thing about them a while ago and I thought the setup it was interesting but my Wife commented on the 'cult' vibe. What was it like living and working there?
I visited Meta (they were trying to recruit me) about 2 weeks after they moved in there... I won't comment on what I saw there, or my thoughts on Meta, but I will say I take all augmented reality claims with a huge grain of salt. I would say that Meron purposefully is trying to form a cult of personality, and while it may be respected as a Steve Jobs-esque thing in the Valley, it is personally a turn off for me.
Funny thing is that the Portola Valley property is now rented out to a friend of mine, which is subleasing it out to tenants... The TechCrunch video really doesn't do it justice in terms of size, and sadly all but one of the tanks are gone. http://sfbay.craigslist.org/pen/roo/5464646558.html
(Though Magic Leap's videos seem to indicate they've solved it. I'm still skeptical though.)
edit: and the FOV looks way bigger on the meta than the hololens
Competition is great though, hopefully they will push each other.
But it seems that the resolution of the current VR goggles is not sufficient :/ This seems a bit weird given that "retina" displays are already quite mainstream.
It sounds like you're describing foveated rendering[1] which requires that the headset also track where your eyes are looking, not just the direction of your head. It's a very exciting possible future direction for VR/AR but I think it's still a bit off (and likely still requires very high resolution miniaturised display for the foreseeable future)
[1] http://www.roadtovr.com/hands-on-smi-proves-that-foveated-re...
The problem with that idea for goggles is that the "field of vision" doesn't have a fixed center relative to the display, because eyes can move independently of the head (and do in normal human interaction with the environment.)
(On a slightly unrelated note: Is there any headsets with VRDs with deformable mirrors that do this? The idea's decades old, but I haven't ever heard about any devices like that on the market.)
But with projector (so-called VRDs, virtual retinal displays) and a deformable mirrors it's possible. At least this was discussed in theoretical literature, articles and various patents - I've read about this stuff a bit, some years ago.
Say you had 4K per eye (I understand it's accepted that 8K per eye is more realistic), that's 8K or 16K driven at 90Hz, you would need some serious graphics horsepower to drive that.
Maybe they're running the leanest of lean operations, but given the challenge of actually making hardware, I'm not certain anyone will see a working version of this until the second Hololens is out.
I'm not knocking the guy. I just don't see his involvement as adding much to the project beyond the name recognition.
Hololens has "2.5K radiants," where a radiant is a "light point per radian." [1] A radian is ~57.3 degrees, so 2500/57.3 = 44 pixels per degree.
Meta has a "2560x1440 display with a 90 degree field of view." Horizontally, that's 2560/90 = 28.4 pixels per degree.
Having used the Hololens for 30 minutes at Build last year, the field of view was definitely less than 90 degrees. However, the tracking was perfect, there was no lag, and generally speaking the illusion was complete.
Watching the video at the top of metavision.com, it's hard to get a good impression-- some shots look like they pointed a video camera through the Meta glasses!? The frame rate is noticeably low and the gestures look like they... sort of work.
Both headsets' problems seem fixable-- Meta will presumably fix the lag and gesture accuracy (assuming those weaknesses in the video are real). And Hololens will presumably get a larger field of view with later hardware revs. To me it'd be a mistake to dismiss these systems as not being ready. It probably won't be a long wait.
[1] https://dev.windows.com/en-us/holographic/hardware_details
Also found on their site:
"direct hand interaction with holograms"
"grab and move holograms just like physical objects"
"Holograms"? Really?!
Marketing BS
Huge difference in how it feels to the user, but yes, pepper's ghost.