Oculus Executive Calls For 3D Equivalent Of JPEG To Build The Metaverse

by Ian Hamilton • July 22nd, 2016

A new standard for 3D scenes is gaining momentum with support from graphics industry leaders, potentially laying the groundwork for science fiction’s “metaverse” to be realized.

The GL Transmission Format (glTF) from The Khronos Group, a computer graphics industry standards body, could also put magnitudes more 3D content on the Internet. The Khronos Group is responsible for a variety of technologies critical to how computers show visuals. Standards include Vulkan, OpenGL, WebGL and others. One of the latest is glTF, designed to streamline the way 3D content is transmitted and loaded across any device. JPEG helped lead to an explosion in the way people make and use images and glTF could do that for 3D scenes.

Valve and Oculus are members of the Khronos Group, among an extensive list that includes Microsoft, Adobe and Amazon. Experts at Microsoft, Adobe, Box and OTOY directly voiced support for the glTF standard in an announcement today, indicating some industry momentum for the format. A quote from Oculus Chief Technology Officer John Carmack was also included, describing the need for such a format.

“The world has long needed an efficient, usable standard for 3D scenes that sits at the level of common image, audio, video, and text formats,” Carmack is quoted as saying. “Not an authoring format, or necessarily a format you would use for a hyper optimized platform specific application, but something at home on the internet, capable of being directly created and consumed by many different applications.”

Being a member of the standards group and offering a quote are not the same as deeply integrating a technology, but Carmack’s comment and those from important figures at other companies involved in VR points to significant interest in the standard. I reached out to Carmack on Twitter and asked him directly what he thought glTF could do for VR, and he wrote back “I think most people hope that the metaverse won’t be built on proprietary media formats.”  In case you are unfamiliar, the “metaverse” is an idea that would essentially extend the Internet into 3D so countless spaces can be linked up and explored with mixed reality.

The royalty-free glTF specification is meant to minimize the size of 3D scenes and models while providing the tech industry with a “common publishing format for 3D content tools and services, analogous to the JPEG format for images,” according to the Khronos Group.

“Ultimately, the metaverse is going to be built on open standards,” said Tony Parisi, VP of web and open technologies at WEVR and co-editor of the specification. “glTF is the first of those standards, and it’s going to lead to a proliferation of 3D content.”

The announcement from the Khronos Group is part of a wave of information being released by researchers, companies and industry groups ahead of the SIGGRAPH computer graphics conference next week in Anaheim, California. Upload will be at the conference and we’ll keep an eye out for more robust indications of glTF adoption, as well as what it could mean for the VR industry.

Tagged with: , , , ,

  • Steve Biegun

    I had always assumed that .obj was the 3D equivalent to .jpg. Today I learned!

    • corysama

      FBX, COLLADA, OBJ, STL. These are all very well publicized formats. But, it is extremely important to realize that they are explicitly not designed for delivering content to consumers. They are designed to move data between editors.

      If you could, would you replace all of your web site’s PNGs and JPGs with Photoshop PSDs? Of course not! PSD is a terrible final format for delivery over the network. FBX is as bad as PSD for that purpose. Please, Please, Please don’t ship editor formats in your apps! Especially not in WebVR! If you don’t want to invent a format, use glTF. Please!

      • Grzesiek

        >Please don’t ship editor formats
        Some people still publish .doc files instead of PDFs…

    • clawjelly

      To add to the other answer, OBJ as a 3D format is more what TGA is to textures. You just store the very basic data (vertex, normals, faces, uv) in there. No bones, no deformers, no lights, no cameras, no animations and no compression of the data. It’s actually an ASCII-format, so you can edit it with a text editor.

      FBX is compareable to PSD, as it also started as the base format for an application (“Filmbox”, the predecessor of Motionbuilder). As such it stores a lot of unnecessary data, if you would to merely create a viewable scene.

      What we need is a format with the complexity somewhere betweeen FBX and OBJ, which only stores the data necessary to view the scene and with good compression.

  • Graham J

    I wonder what existing metaverses like High Fidelity are using…

    • Neil Trevett

      High Fidelity currently uses FBX and .OBJ models – and see the insightful comments above from corysama and clawjelly as to how ‘delivery-optimized’ glTF can complement those ‘authoring-optimized’ formats.

      If we are to truly build the ‘metaverse’ we need to avoid silo’d data – where different servers, apps and sites don’t understand each others assets. Imagine the chaos on the Web if every browser used a different image format and users couldn’t easily send a picture from one app to another. glTF is working hard to be a widely available, royalty-free, efficient transmission format to help us avoid that dystopian situation for 3D.

  • Julian Gomez

    X3D has been around for 15 years as an ISO standard, and *is* designed for end delivery. All the other formats are industry practice, not standards.

  • galopin

    This is not practical. I understand the need at capturing 3d scene to replay and navigate inside in a generic way. But there is as many way to render such a scene than there is stars in the universe. Lighting, color grading, geometry animation, procedural material, shadow filtering, global illumination, etc. There is no single graphic engine capable to reproduce all the crazy way we invent to improve graphics everyday.

    • Neil Trevett

      glTF is ‘just’ a container for 3D assets such as meshes, textures and animation data. It does not prescribe in any way how a run-time should use that data – so an application is free to employ any rendering techniques it wishes. This is a major difference to standards such as X3D that do combine a definition of run-time behavior with a file format. Standards such as X3D can be valuable in vertical applications – but the industry needs a commonly agreed way to simply ‘transmit’ 3D assets

      • erikbjareholt

        Best description in this comment thread. Well done.