Everything you need to know about Google I/O 2017: Daydream standalone headsets and a lot more

Last week has finished with a big boom for virtual reality: at Google I/O conference, Google has announced its upcoming Standalone headsets. In this post I’ll try to describe all the main information about this and all other AR/VR news they’ve announced… information that are splattered around dozens of posts around the web and that I’m trying to condensate into a single article. From this idea, I’ve come up with the above super-clickbait title 🙂

We all know that some months ago Google has announced Daydream, a platform for mobile virtual reality. It is like GearVR, but with two main differences:

• Its headset is more fluffy: I’ve tried and reviewed it (full review here) and I found it incredibly comfortable;
• The platform is compatible with more phones made by different vendors. GearVR is only Samsung compatible (and every Gear VR model is compatible only with some high-tier Samsung phones), while Daydream will be compatible with all phones that will be Daydream-compatible. This will mean a more open virtual reality platform, perfectly in the style of all Google software.

From last Google I/O conference, Google has made lots of announcements about this platform, both on a software and hardware side. Let’s start from the platform news… I’ll give you the most juicy news at the end of the article (skip there if you’re just interested in Standalone headsets!). As Latins said, dulcis in fundo.

Daydream Apps number

About standard Daydream platform, Google announced that currently there are 150 apps for Daydream. In my opinion the number is quite low and this is surely due to the fact that:

• Daydream is a young platform, it has just born some months ago;
• Google started approving only high-quality apps, blocking all the experiments and low quality experiences;
• There are few Daydream headsets out there, so developers are not that interested in developing for a little niche. This is the same issue we’ve already seen a bazillion times in the VR market and that hopefully will start solving in 2019.

150 is not a great number, to be honest. Gear VR apps are surely at least 5 times that number (but Gear VR is out since a longer time).

Daydream platform improvements

To keep pace with its competitors, Google will improve current Daydream platform. No, wait, it will improve just everything. The upcoming Android OS, Android O, will be designed to keep VR in mind. The operating system will include lots of features and improvements necessary for VR.

Apart from that:

• The Home screen will be revolutionized and will let people to organize contents (VR apps, 360 videos, etc…) in a smarter way. So not just a flood of apps, but categories into which put the content and keep everything organized. I hope this will happen because for example I find Oculus Home too simple… I don’t want just a list of my apps without being able of categorizing them;
• A notification system will be included. As Road To VR says, “2.0 will add a dashboard on top of any VR experience so you can check notifications, change settings, pause experiences, and most importantly switch from app-to-app like on a smartphone. This includes support for 2D system UI”;
• Daydream Cast will allow people to easily stream current VR experience to a nearby Chromecast, so that everyone will be able to see what the user is seeing in VR. This is something that is already possible, so I guess that the news is that it works better and out-of-the-box;
• Daydream Sharing will let people take screenshots of the VR apps they’re playing and easily share them on the social networks. Cool, since at the moment this is a real pain for VR mobile apps;

Google Chrome AR and VR upgrades

Google will finally make its Android-version of Chrome to be completely WebVR compatible. This will mean that opening a WebVR link on Chrome on a Daydream-compatible phone will let us live the immersive experience with just a single click. Chrome will become as current Samsung browser: you will able to use it from within the headset in VR mode or on the standard flat screen of the phone in 2D mode. It will be the same browser, so it will share all history, bookmarks, etc… and switching from one mode to the other will be seamless. Just amazing.

No, wait… there’s something even more amazing! I’m talking about WebAR… we’ve just began praising WebVR that WebAR is coming out! WebAR is a framework to make web-based augmented reality applications and this would be amazing for a-shopping (e-commerce using AR) for instance. Google has already released an experimental build for Chromium supporting WebAR (Chromium is the open source version of Chrome and it is also the browser that supports WebVR on PC if you remember well).

Samsung S8 will be Daydram compatible

Samsung has soon realized that it can’t bet on Gear VR only. While Gear VR is surely an amazing platform and we all praise it for having been the first high-tier mobile VR headset, the fact that is single-producer oriented may mean it will have little market shares in the end. That’s why Samsung decided that it would be a great idea to be compatible with both Gear VR and Daydream platform. And in my opinion this is a smart move.

This summer, Samsung will release an update for its models S8 and S8+ that will make them Daydream-ready. This way Samsung will be the only vendor that will support all the main mobile VR platforms (Cardboard, Daydream, Gear VR) and this could mean a little advantage for the Korean manufacturer in the VR enthusiasts market.

Google VPS

We all know that Google has been the first one (among major companies) that has created a system for reconstructing and tracking an environment from a handheld device: Tango. Now, using Tango capabilities, plus some Google Maps magic, Google has developed VPS (Visual Positioning Service): an indoor localization system. It basically works this way: the phone scans the environment while you move inside it and is able to reconstruct a complete map of it, so it is able to reconstruct exactly your position inside a certain building. So it is like a GPS for indoor environments. If you’re wondering, accuracy is of few centimers.
So with GPS you can arrive at a certain building and then with VPS you can start orienting yourself inside that building. This can be super-useful for business applications: for example in a hospital you may be guided by VPS on your phone to your target department; or first-day employees may orient inside a huge factory without no one accompanying them (so time can be spared for training new hires). I guess that this will be huge when added to AR glasses, since you’ll be able to see all informations directly in front of your eyes.

Google Lens

Google Lens is Google way of saying “bitch please” to Facebook Camera. It has been defined as

a set of vision-based computing capabilities that can understand what you’re looking at and help you take action based on that information

This means that thanks to Google enormous AI capabilities, the system is able to detect what you’re looking at. So, you point the system towards a flower and it doesn’t only detect you’re looking at a flower, but it also understands which kind of flower you’re looking at. This is pretty impressive (and also quite scarying). I guess that the system is still in development, but they on the stage showed some examples nonetheless.
One is a mix of Google Maps with Google Lens: you’re in a street and point your camera towards a restaurant. The system, from GPS informations and from the image of the restaurant facade is able to detect the restaurant and show to you in AR information about that restaurant, directly on the camera stream. Really impressive.

Google Seurat

https://gfycat.com/NearPepperyHoneyeater

(GIF from Road To VR)

Google has announced Seurat: a new light-field technology that is able to show ultra-complex CGI scenes using cheap hardware like the one of mobile VR headsets.

I mean, you can have a pre-recordered scene taken laser-scanning an environment or you can have constructed it with Blender or such… I’m talking about a complex scene with millions of polygons, with super cool materials. It is impossible to put the user inside it in VR and let the user move inside it at 60FPS. The only possibility is to render the scene as a flat 360-videos and put the user inside that video… but that means a static flat scene projected inside a sphere, without positional tracking and 3D data. Well, it seems that this WAS impossible.

Thanks to Seurat, you just fed the super-mega-assets to Seurat, that transforms them into a very little structure, super-optimized for VR and for the point of view the user will be at. Then when you put the headset on, you’ll just see the scene as you were expecting it, with high-quality on a low-end device. You can also move your head and tilt it and see everything remaining coherent: all in 3D, all with high quality. This is possible thanks to an incredible set of optimizations: since you know what is the expected position of the user, you can cull away all the polygons you’re sure that will never seen by him, for instance; and the same hold for the light, since you know the position of the user, you can compute lighting once for all. There is the possibilities of moving inside the scene, but only in a limited space, so we can estimate that the user is like in a fixed position.

https://gfycat.com/HomelyMajesticGroundbeetle

(GIF from Road To VR)

How this can be exactly accomplished is pure magic: Road to VR has seen this technology and has remained astonished by its performance, its support for positional tracking and the file size of the experience (current light-field technologies require lots of gigabytes to store their data, this one just megabytes). Details on Seurat have still to be released.

Google Standalone headsets

Let’s be honest: this has been the real news, that has been highlighted by all the major VR outlets. It has been also reported by Google’s VP of AR and VR Clay Bavor on the company blog and also on Medium (even if IMHO the post on Medium is super-fluffy and is not worth reading).

Google has announced a new line of Daydream headsets: Daydream standalone headsets. With standalone I mean a headset that includes everything, that is all-in-one: it has the visuals, the processing power, the tracking technology, etc… without requiring any external hardware to work. You just buy the headset, wear it and have virtual reality.

In my opinion, VR standalone headsets are the future and will surely supersede mobile VR headsets in the long term. Tethered headsets will have longer life due to higher processing power of a desktop PC (but they will become wireless… or maybe we will just use standalone headsets connected to a streaming PCs), but as I’ve written in a previous article, mobile VR headsets are destined to die in my opinion. Google has taken its move because it has read my article, for sure 😀 . Jokes apart, Palmer Luckey too has expressed in an interview his vision about the future success of standalone headsets and all companies are betting hard on them, starting from Oculus (Santa Cruz) and Intel (Alloy), to Microsoft (HoloLens).

Reason is: standalone devices are very handy and user-friendly (you just put them on and you have VR) and can be hyper-optimized for VR (if you put a smartphone in a VR headset, you’re using a device that is meant to be used with a touch interface in 2D to perform virtual reality tasks; with standalone headsets, you just make hardware and software all optimized for VR).

Standalone reference design

As it is current trend, Daydream will not be a model, but will be a reference design for future standalone headsets that Google partners may produce. Google made the smart choice of not creating a standalone headset reference from scratch, but started from the promising design of the Snapdragon 835 headset. This design included positional tracking, eye tracking and also hands tracking thanks to Leap Motion integration. It is not clear if the latter two will be integrated into the headset, but the positional tracking will surely be.

First producers entering this standalone waves have been Lenovo and, surprisingly, Vive. Yes, you’ve read well: there will be a Vive standalone headset. This news shocked me a bit, but makes sense in the end: due to the uncertain evolution of the VR market and the fact that Vive was the only one not having a standalone headset (they have only a wireless add-on), this was a natural choice. They had to enter the standalone headsets market and Google, that has the same opensource philosophy of SteamVR, was a natural choice. If you go to Vive website, there’s already a page for the upcoming standalone headset, that should be released before the end of this year. So we’ll continue seeing a Oculus vs Vive war even on standalone devices: fanboys will be super-happy.

Worldsense positional tracking

How does Daydream standalone headset perform positional tracking? Well, exactly as all other HMDs with inside-out tracking: using two RGB cameras that are in front of the device (and no depth ones). Thanks to Tango technology (they have developed it for too many years… we VR enthusiasts were wondering what all of this was going to be used to), the system is able to analyze the data coming from the cameras and reconstruct the exact 3D position of the user, that can dodge, walk and jump freely. Since in the VR world people loves giving technology super-marketing-fluffy names, this technology’s name is Worldsense.

Some selected journalists have been able to try the device and both Road To VR and UploadVR ones report that the tracking was wonderful. The system is also robust to occlusions of one of the two cameras and can continue doing a good tracking even in monoscopic mode! Due to the demo setup, there are two things that are still not clear:

• If the system can track only “rooms” or if can track whole environments. I mean, due to the above demo of VPS, we start to wonder if these standalone headsets are able to track all our houses and let us walk freely from one room to the other with the HMD on or if we can just translate inside a single room;
• If the system is precise enough and how it deals with real objects: the demo was carefully studied so that there were no real and virtual objects near the journalists; furthermore it was quite short. It would be interesting to understand if the system drifts over time; if the tracking is precise enough; how the system behaves if I’ve virtual objects near me (so maybe it could warn me about the fact I’m goint to stumble across a chair) and how much this virtual avoidance feature is precise.

3DOF Controller

Do you remember when, some lines ago, I was wandering if hands tracking would have been included in this reference design? Well, it seems that at the beginning it will be not. The controller for the first headsets will be the Daydream 3DOF controller, with 3DOF meaning it would be able to detect only its rotation and not also its position. This seems to be a reasonable choice to start with, first of all because I think that for a lot of VR applications the remote is the best controller ever and then because giving the fact that there are no external cameras, it is very hard to have 6DOF controllers. We’ve seen how Microsoft has had to make big compromises to have 6DOF controllers for its cheap headsets. I think that big players will need time to solve this issue.

Software platform

Daydream headsets will just run Android O and all apps developed for Daydream mobile headset will work with Daydream standalone headset. Don’t know how much time this will last, but at the moment this is surely a good news. Especially for developers, that will be able to target two platforms with just one app. I’m wondering if Vive will push to make SteamVR compatible with Daydream headsets, too.

Price

We still don’t know, but Google has revealed that it will be similar to the one of Oculus and Vive. So we can expect a price around $700-$900, a very reasonable price for an all-in-one device, considering that HoloLens costs something like $3000. In my opinion the real issue in the short term will be convincing people to spend such money for a device able to do only VR stuff: with Gear VR you spend that money, but you obtain a high-tier smartphone that you can use every day to do what you want.

So I guess that at the beginning this device will be sold to innovators and to the business market: arcades, marketing agencies for VR booths, training companies, etc… These companies are able to spend such money and are very happy to obtain a great device that work out of the box without any hassles.

—

And that’s it! Hope to have not forgot anything! And remember, if you don’t want to miss VR related news, subscribe to my newsletter using the form on the right sidebar! ( Please… 😉 )

(Header Image by Google)