Khronos Group – Spime Wrangler

I’ve been involved and observed technology standards for nearly 20 years. I’ve seen the boom that came about because of the W3C's work and the Web standards that were put in place early. The standards for HTTP and HTML made content publishing for a wider audience much more attractive to the owners and developers of content than having to format their content for each individual end user application.

I’ve also seen standards introduced in emerging industries too early. For example, the ITU H.320 standards in the late 1980s were too limiting and stifled innovation in the videoconferencing industry a decade later. Even though there was an effort to correct the problem in the mid-1990s with H.323, the architectures were too limiting and eventually much of the world went to SIP (IETF Session Initiation Protocol). But even SIP has only had limited impact when compared with Skype for the adoption of video calling. So, this is an example where although there are good standards available, they are implemented by large companies and the mass market just wants things that work, first time and every time. AR is a much larger opportunity and probably closer to the Web than video conferencing or video calling.

With AR, there’s more than just a terminal and a network entity or two terminals talking to one another. As I wrote in my recent post about the AR Standards work, AR is starved for content and without widespread adoption of standards, publishers are not going to bother with making their content available. In addition to it being just too difficult to reach audiences on fragmented platforms, there’s not a clear business model. If, however, we have easy ways to publish to massive audiences, traditional business models such as premium content subscription and Pay to watch or experience, are viable.

I don’t anticipate that mass market AR can happen without open AR content publishing and management as part of other enterprise platforms. The systems have to be open and to interoperate at many levels. That's why in late 2009 I began working with other advocates of open AR to bring experts in different fields together. We gained momentum in 2011 when the Open Geospatial Consortium and the Khronos Group recognized our potential to help. These two standards development organizations see AR as very central to what they provide. The use of AR drives the adoption of faster, high performance processors (which members of the Khronos Group provide) and location-based information.

There are other organizations very consistently participating and making valuable contributions to each of our meetings. In terms of other SDOs, in addition to OGC and Khronos, the W3C, two sub committees from ISO/IEC, Open Mobile Alliance, Web3D Consortium and Society of Information Display are reporting regularly about what they’re doing. The commercial and research organizations that attend include, for example, the Fraunhofer IGD, Layar, Wikitude, Canon, Opera Software, Sony Ericsson, ST Ericsson and Qualcomm. We also really value the dozens of independent AR developers who come and contribute their experience as well. Mostly they’re from Europe but at the meeting in Austin we expect to have a new crop of US-based AR developers showing up.

Each meeting is different and always very valuable. I'm very much looking forward to next week!

In preparation for the upcoming AR Standards Community Meeting March 19-20 in Austin, Texas, I’ve conducted a few interviews with experts. See here my interview with Marius Preda. Today’s special guest is Neil Trevett.

Neil Trevett is VP of Mobile Content at NVIDIA and President of the Khronos Group, where he created and chaired the OpenGL working group, which has defined the industry standard for 3D graphics on embedded devices. Trevett also chairs the OpenCL working group at Khronos defining an open standard for heterogeneous computing.

Spime Wrangler: When did you begin working on standards and open specifications that are or will become relevant to Augmented Reality?

NT: It’s difficult to say because so many different standards are enabling ubiquitous computing and AR is used in so many different ways. We can point to graphics standards, geo-spatial standards, formatting, and other fundamental domains. [editor’s note: Here’s a page that gives an overview of existing standards used in AR.]

The lines between computer vision, 3D, graphics acceleration and use are not clearly drawn. And, depending on what type of AR you’re talking about, these may be useful, or totally irrelevant.

But, to answer your question, I’ve been pushing standards and working on the development of open APIs in this area for nearly 20 years. I first assumed a leadership role in 1997 as President of the Web3D Consortium (until 2005). In the Web3D Consortium, we worked on standards to bring real-time 3D on the Internet and many of the core enablers for 3D in AR have their roots in that work.

Spime Wrangler: You are one of the few people who has attended all previous meetings of the International AR Standards Community. Why?

NT: The AR Standards Community brings together people and domains that otherwise don’t have opportunities to meet. So, getting to know the folks who are conducting research in AR, designing AR, implementing core enabling technologies, even artists and developers was a first goal. I need to know those people in order to understand their requirements. Without requirements, we don’t have useful standards. I’ve been taking what I learn during the AR Standards community meeting and working some of that knowledge into the Khronos Group.

The second incentive for attending the meetings is to hear what the other standards development organizations are working on that is relevant to AR. Each SDO has its own focus and we already have so much to do that we have very few opportunities to get an in depth report on what’s going on within other SDOs, to understand the stage of development and to see points for collaboration.

Finally, the AR Standards Community meetings permit the Khronos Group to share with the participants in the community what we’re working on and to receive direct feedback from experts in AR. Not only are the requirements important to us, but also the level of interest a particular new activity receives. If, during the community meeting I detect a lot of interest and value, I can be pretty sure that there will be customers for these open APIs down the road.

Spime Wrangler: Can you please describe the evolution you’ve seen in the substance of the meetings over the past 18 months?

NT: The evolution of this space has been rapid, by standards development standards! This is probably because a lot of folks have thought about the potential of AR as just another way of interfacing with the world. There’s also been decades of research in this area. Proprietary silos are just not going to be able to cover all the use cases and platforms on which AR could be useful.

In Seoul, it wasn’t a blank slate. We were picking up on and continuing the work begun in prior meetings of the Korean AR Standards community that had taken place earlier in 2010. And the W3C POI Working Group had just been approved as an outcome of the W3C Workshop on AR and the Web.

Over the course of 2011 we were able to bring in more of the SDOs. For example, the OGC and Web3D Consortium started presenting their activities during the Second community meeting. The OMA Mob AR Enabler work item presented and ISO SC24 WG 9 chair, Gerry Kim, participated in the Third Meeting in conjunction with the Open Geospatial Consortium’s meeting in Taiwan.

We’ve also established and been moving forward with several community resources. I’d say the initiation of work on an AR Reference Architecture is an important milestone.

There’s a really committed group of people who form the core, but many others are joining and observing at different levels.

Spime Wrangler: What are your goals for the meeting in Austin?

NT: During the next community meeting, the Khronos Group expects to share the progress made in the newly formed StreamInput WG. We’re just beginning this work but there’s great contributions and we know that the AR community needs these APIs.

I also want to contribute to the ongoing work on the AR Reference Architecture. This will be the first meeting in which MPEG will join us and Marius Preda will be making a presentation about what they have been doing as well as initiating new work on 3D Transmission standards using past MPEG standards.

It’s going to be an exciting meeting and I’m looking forward to participating!