Oculus SDK 0.6.0 and Oculus Rift in Action

Oculus publicly released version 0.6 their SDK today.  This version represents the biggest change to the API since the first release of the C API in version 0.3, slightly over a year ago.  The new version simplifies the API in a number of ways:

• Removed the distinction between direct and extended modes from the perspective of the developer  
• Removes support for client-side distortion
• Moves the management of the textures used for offscreen rendering from the client application to the SDK
• Fixes a number of OpenGL limitations
• etc, etc...

So those who have purchased or intend to purchase our book may be asking themselves what version we'll be covering exactly.

The short answer is that currently the book covers version 0.5 of the SDK (which itself is almost identical to version 0.4.4).  That isn't likely to change before print.  At this point we're well into the final stages of publication, doing final editorial passes for proofreading.  If we were to update to 0.6 we would have to go back and revise quite a few chapters and start much of that work over again.  That said, when we found out what a big change 0.6 was, there was much discussion, consideration and consternation.

However, with the public release of the SDK, there came a more compelling reason for us to stay with the version we're currently targeting.  In a recent blog post on the Oculus website, Atman Binstock stated...

Our development for OS X and Linux has been paused in order to focus on delivering a high quality consumer-level VR experience at launch across hardware, software, and content on Windows. We want to get back to development for OS X and Linux but we don’t have a timeline.

Since we started writing the book, we've felt very strongly about keeping a cross-platform focus.   While it's likely that most of the initial consumer demand for the Rift will be driven by gamers, we the authors feel that VR has a promising future beyond just the realms of gaming.  Two of the core components that make the Rift possible are powerful consumer graphics processors and cheap, accurate inertial measurement units (IMUs).  Much of the initial consumer demand for both of these technologies was driven by gaming, but in both cases the use of the technologies is spreading far beyond that.

We believe that VR will also grow beyond the confines of gaming, probably in ways we can't even imagine right now.  However, to do that, we need to make sure as many people as possible have the opportunity to try new ideas.  And while Macs and Linux machines may hold an almost insignificant market share when it comes to gaming, we believe that new ideas can be found anywhere and that innovators probably aren't divided proportionately in the same way that operating systems are.

So we were faced with a choice.  We could stick with the 0.5 SDK, which still functions with the new Oculus runtime (at least for the DK1 and DK2).  We could abandon Linux and Mac developers and switch to the 0.6 SDK.  Or we could try to cover both 0.6 and 0.5 in the book.  We chose the first option.

Now, if you're a Windows developer and you want to move to 0.6, we still want to support you.  To that end, we will have a branch of our Github examples that will keep up to date with the latest version of the SDK.

Additionally, we will try to cover the gaps between what the book teaches, and and the current version of the SDK with articles on this blog.  I'm not talking about errata, but something more like whole updated chapter sections.

We will also be working to create examples focused on other mechanisms of working with VR devices, such as the Razer OSVR and the HTC Vive, using the appropriate SDKs.

We also feel that there is a great deal of value in the book as it stands now that is outside of the low level SDK details.  In fact, to be honest, the bulk of the changes to the book would probably be in chapters 4 and 5, and there are a dozen chapters.

The point is, whatever platform you're on, whatever hardware you're working with, if you want to create a world, we want to help.  The forthcoming edition of Oculus Rift in Action is only the first step.

Book Update!

We are heading into the final stretch with our book Oculus Rift in Action.  All chapters are now complete and we are just about half-way through the final editing process. The book is available for pre-order from Amazon.  If you can't wait to get your hands on it, you can order directly from Manning Publications and get access to all chapters now through the Manning Early Access Program.

Unity + Leap: Hand Selection UI Prototype

Immersion is definitely affected by how closely your avatar's hand looks like your own.  In the demo I am working on I want the user to be able to select the hands they have in the game before entering the game.  A prototype in-game UI for hand selection is seen in the video below.

To create this UI, I created a world space canvas and added buttons for each of the available hands. To each button, I added a box collider as a child object. A script attached to the box collider detects when a hand has collided with it.  To detect a hand, I used the Leap libraries and then checked to see if the collision object is a Leap HandModel.

In this prototype UI, I am using large buttons for two reasons. First, reading small text in the Rift can be difficult, and second, while using the Leap allows me to see my hands, in my experience, it does not track finger motion well enough for detailed interactions to be effective. In several of the tests I ran, the user's hand was generally in the right place but the fingers more often than not were at different angles than the user's actual hand. The effect was that my users seemed to have the fine motor skills of a toddler - they could reach out and touch everything but they didn't have a lot of control. On the positive side, when the user has hands in the game, it appears to be very natural for users to try to touch items with their hands. Even when users don't have visible hands in the scene, you'll often see them reaching out to try to touch things. While I have the start button say "Touch to Start," once users know to use their hand to affect the scene they get it right away and don't need prompting or other instruction.

Leap Motion has just released a "Best Practices Guide" and I'll be looking at incorporating many of the ideas documented there in future prototypes.