Google AR: Tango is (probably) dead, long live ARCore

Google has unveiled ARCore, a new software developers’ kit (SDK) and set of tools for building augmented reality (AR) applications for Android. The launch suggests the inevitable demise of Tango, Google’s other AR platform, which labored under hardware requirements with which ARCore appears to be unburdened.

Google’s launch is a response to Apple’s ARKit, which was unveiled back in June, and which was well-received by developers.

Currently, ARCore is only available in a limited preview form, running on the Google Pixel phone and requiring Android 8.0 Oreo or 7.0 Nougat, or on the Nougat-running Galaxy S8. Google says it is working with Asus, Huawei and LG, to add ARCore to their handsets. Google is apparently aiming to have the SDK supported by 100m phones, and has scheduled the preview to run into the winter.

Samsung’s support is very significant, as the smartphone market leader never supported Tango in its designs before. The firm’s Galaxy S8 is likely to be one of the bestselling smartphones over the next year, and while the Oz demonstration app (featuring characters from The Wizard of Oz) might provide a few minutes of entertainment for new buyers, Google is hoping that ARCore goes further, and manages to persuade these users that AR is a valuable all-round feature.

The ARCore functionalities on show look pretty interesting. A developer can rely on the background lighting from the camera to cast the appropriate shadows and glare on their AR model graphics, for realistic lighting, and that sensitivity also lets the developer create elements that react to lighting changes – with a cartoon lion that is afraid of the dark a prominent part of the demo, jumping and clutching his tail when someone turns off a light switch. That same framework will let elements be positioned according to what the camera captures, sitting on tables and standing on pavements.

Meanwhile, Tango hasn’t had much in the way of recent announcements, and remains confined to just two phones – the Lenovo Phab 2 Pro and the Asus Zenfone AR. With Tango’s focus on specific hardware and sensors, including an infrared depth sensing camera design, which would be integrated into the handsets, there was always a question of cost and adoption hanging over the technology’s head.

CNET put the question of Tango’s future to Clay Bavor, Google’s head of VR and AR, asking whether ARCore meant the shuttering of the older brand. “That’s too strong a statement,” replied Bavor, but it’s hard not to agree with CNET’s conclusion that things look grim for Tango.

“Our developer-facing brand and consumer awareness – I think we will shift very much towards ARCore,” he said, and added that Google was still interested in advancing AR hardware. Speaking to The Verge, Bavor said: “I think Tango fades into the background as more an enabling technology that kind of works behind the scenes.”

So Tango’s software components are likely to be bypassed by or merged into ARCore, and if Tango survives, it will likely be as a skunkworks project working on new hardware designs that could be integrated into OEM phone designs. The success of Tango will depend on those potential designs being cheap and easy to install – otherwise the likes of Samsung are unlikely to opt for them. Ars Technica has a more in-depth interview with Bavor, which is worth a read.

Broadly put, ARCore is less capable than Tango, but far more accessible and easier to adopt on a wider range of devices. Three main components drive the experience; motion tracking that uses sensors and the video feed to estimate location; environmental understanding for recognizing and reacting to objects; and the light estimation that helps graphics blend into their backgrounds.

In terms of developer environments, ARCore supports Java, OpenGL, Unity and Unreal – which allow for the creation of the graphics. Also available are two AR-enabled web browsers, one for Android, and one for iOS. In demos, the browsers enable previews of items on shopping websites, such as furniture previews for Wayfair.

As for future uses for ARCore, it could serve as the basis for interactive applications that harness other Google offerings. An example might be using the camera to capture an object about which the user wants information, and having Google’s search engine retrieve that information and overlay it on the camera feed using ARCore. Usage information, or things like purchasing links, could also be included, as a means of monetizing the system – which will likely rely on a hefty amount of cloud-based processing.

It’s also worth mentioning Daydream, Google’s VR-focused developer framework, which also includes the Daydream headset. That’s a platform that seems to compete with Samsung’s Gear VR systems, but Samsung is also involved in VR projects with another web-giant – providing software and integrations for Facebook’s early experiments.

Facebook’s $3bn acquisition of Oculus was a definite play for AR market share, as a means of increasing engagement on the social network. Early demos included cartoonish avatars, which would allow users to experience a shared environment together – an environment that Facebook can sculpt to boost advertising impact, if desired.

Live video streaming and sharing is the other element of Facebook’s interest in VR and AR, although Google has a similar interest via YouTube’s livestreaming options. Both are looking to get eyes glued to a video feed, as viewers experience something very engaging. That feed could be manipulated to insert advertisements into the video itself, or simply have the adverts overlaid on top – which could prove very influential if the brands can rely on viewers having their eyes glued to the VR or AR screens.

Also of note is another popular application. Snapchat’s popularity has done much to make AR more familiar to smartphone users. While it’s filters and AR elements are mostly for taking pictures and capturing video, the interaction with the phone and environment is in the same manner that more focused AR applications will have to rely on.