Your browser is not supported. Please update it.

10 July 2014

Upgraded Kinect for Windows announced to developer delight

Microsoft is releasing the second generation of the Kinect for Windows platform on July 15th for $199. This is big news in a lot of small ways, as the Kinect has found a home away from its birthplace as an add-on for the Xbox, largely in R&D projects at the bleeding edge of technological progress, in robotics in particular.

The new Kinect for Windows is the same Kinect that is included with the Xbox One. With an infra-red depth sensing camera and a 1080p camera for image capture, with a 60% wider field of vision, the Kinect v2 can track 6 people’s movements at once. Its IR camera could also track player’s heartbeats, and the camera is able to track precise hand movements noting the orientation of up to 25 joints in the hand. An integrated microphone array provides voice control.

This is a significant upgrade over the original Kinect, which offered the Xbox 360 movement, voice and facial recognition capabilities. While the Kinect hasn’t received much attention from the gaming community and its developers, it has proved wildly popular as a tool for PC developers and has turned up in applications as diverse as simple file management to hands-free operating theater cameras. The Kinect v2 promises to boost these projects with its greater image fidelity and recognition technology.

The module has been embraced by the developer community because it manages to combine a number of functions inside a very affordable package, relative to the cost of going out and buying or building an equivalent system from scratch. Thanks to the scale of production, it was fairly straightforward and cost effective for an engineer to go out and buy a Kinect and then program it to interface with the project at hand.

But the limitations of the v1’s camera meant that tracking minute movements of small objects such as fingers was not feasible. Now there is a large amount of optimism among the developer community that the Kinect and the public SDK will enable more progress to be made on projects using the Kinect as their eyes to the world.

When it was initially launched, and before it was adopted by the developer community at large, the Xbox Kinect simply looked like the next-generation of gaming webcams, which were typically used for games involving dancing and children’s games. The Kinect would push an image to the TV’s display, which allowed the console to overlay its graphics on top of the ‘real life’ image. This meant that dance moves could be scored and video game characters swatted away ‘ all with the novelty of seeing yourself in the game you are playing.

The Kinect was the evolution of something like the PlayStation 2’s EyeToy, which was essentially a 320×240 pixel webcam with an integrated microphone. The EyeToy had sold 10.5 million units as of the end of 2008 after 5 years in stores, compared to the total global sales of 155 million PlayStation 2s.

For the PlayStation 3, launched in November 2006, Sony followed up on the EyeToy with the PlayStation Eye after a year of PS3 sales. Very similar in function to its predecessor, the Eye has a 640×480 pixel camera at 60Hz, but can halve the resolution to double the refresh rate for certain applications.

For the Xbox 360, launched in November 2005, Microsoft developed the Xbox Live Vision webcam and released it in October 2006. Very similar to the EyeToy, it had a 640×480 pixel camera at 30Hz, and could also halve the resolution to double the refresh rate.

In November 2010, after 5 years of sales, Microsoft developed and released the Kinect. It included 2 cameras, one a 640×480 RG B camera at 30Hz, and the other a 640×480 infrared depth-finding camera also at 30Hz. The IR functionality meant that the Kinect could track people in space, instead of simply tracking motions on a screen using a basic video camera. This allowed the Kinect to provide full body 3D motion capture, gesture recognition, facial recognition and voice recognition thanks to the integrated microphone array.

But the Kinect had a troubled life as gaming peripheral. It cost $149 at launch, compared to the $199 of the base model Xbox 360. Justifying that investment was made harder by the fact that very few serious game developers got on board with the Kinect in terms of incorporating it into the flagship titles. Instead, the Kinect was saddled with a reputation as something of a gimmick, for kids, fitness programs and casual gamers ‘ while the serious gamers stuck with the physical controller.

But the Kinect’s hardware capabilities were undeniably impressive, just not very appealing to one of these serious gamers at the $149 price point. To this end, Microsoft decided to follow up on the Kinect by including it in the purchase price of the Xbox 360’s successor ‘ the nonsensically named Xbox One, launched in November 2013. The aim was to ensure that developers would work on incorporating the Kinect into their projects, as every Xbox One could be guaranteed to have the accompanying sensor. With the Xbox 360, many developers opted not to work with the Kinect due to its poor penetration. The idea was that the Kinect would be integral to the Xbox One experience.

But Microsoft set itself up for some major back-pedaling when it announced that the Xbox One would only work when the Kinect was plugged in. There was a huge consumer backlash to the news, with most citing the privacy concerns inherent in a device as powerful as the new Kinect ‘ among other issues with the Xbox One that many felt were anti-consumer (an always-on internet connection to validate a user’s license and games library and lock the console if invalid, being the standout criticism) and more media-focused than gaming-focused as part of Microsoft’s plan to control the living room and not just the console.

The Kinect’s microphone would always be listening for commands, such as the now infamous ‘Xbox on,’ and ‘Xbox off,’ which have been exploited by prank-happy friends and most recently a TV commercial from Microsoft that would turn on any Xbox One in the TV’s vicinity ‘ not the original intention of the advert.

The mob was incensed when patent applications were discovered that showed the Kinect being used to track viewing behaviors, serve targeted advertising and enforce DRM. Nightmare scenarios were imagined that included the Xbox powering off unless the room purchased a multi-user license to watch the Super Bowl, or game licenses being tied to faces and only playable by that specific user. Microsoft backtracked on the mandatory Kinect requirement, and made its use optional – although it still sold a Kinect with every Xbox One.

That was until it became clear that Sony’s PlayStation 4 was winning the console sales race, largely due to being $100 cheaper than the Microsoft offering and being completely free of all the bad press involved with the Kinect. After it emerged that Sony had sold 7 million PS4s while Microsoft had only shipped 5 million units to retailers with 3 million confirmed sales, the decision was made in May to stop bundling the Kinect with the Xbox One. $100 was taken off of the retail price ‘ conveniently aligning it perfectly with the PS4 at roughly $399.

Nonetheless, the Kinect family has sold 24 million units as of February 2013, compared with 83 million Xbox 360s sold. Sales figures for the latest iteration are not yet available, but we expect them to be under the 5 million figure that inspired Microsoft’s unbundling decision. The cynic in us thinks this is an attempt to sell off a few warehouses full of Kinects that Microsoft has left over from the Xbox debacle.