glTF: Designing an Open-Standard Runtime Asset Format - GPU Pro: Advanced Rendering Techniques - page 391

COLLADA glTF glTF+Open3DGC glTF+Open3DGC

Model

ascii binary

XML gzip raw gzip raw gzip raw raw bin

gzip JSON

bin:102k bin:81k ascii:29k ascii:19k bin:18k bin:18k

JSON:11k JSON:2kb JSON:11k JSON:2k JSON:11k JSON:2k

336k 106k 113k 83k 40k 21k 29k 20k

Figure 2.10. Compression test with Open3DGC.

Similar to the GL APIs, glTF will support profiles and extensions to allow
use of functionality not available everywhere. For example, a profile for modern
OpenGL or OpenGL ES would allow using binary shaders to improve load times.
The current content pipeline in COLLADA2GLTF is a good start, but more
stages are planned to fully support COLLADA and better optimize the out-
put glTF. These stages include tessellating splines, flattening meshes and nodes,
optimizing and minifying GLSL, and minifying JSON. There is also a lot of par-
allelism within and across stages that can be exploited to improve the latency of
asset conversion. Finally, improving the architecture so the optimization stages
can be done independent of converting to glTF will make the content pipeline
useful beyond glTF. A high-quality open-source content pipeline for glTF is key
for its adoption and success.

2.11 Acknowledgements

Khronos, WebGL, COLLADA, and glTF are trademarks of the Khronos Group
Inc. OpenGL is a registered trademark and the OpenGL ES logo is a trademark
of Silicon Graphics International used under license by Khronos.

Bibliography

[AMD 14] AMD. rest3d GitHub Repo. http://github.com/amd/rest3d, 2014.

[An alytical Graphics 14] Analytical Graphics, Inc. Cesium GitHub Repo. https:
//github.com/AnalyticalGraphicsInc/cesium, 2014.

[Fa brobinet 14] Fabrobinet. MontageJS Viewer GitHub Repo. https://github.
com/fabrobinet/glTF-webgl-viewer, 2014.

[Fielding 00] Roy Thomas Fielding. “Architectural Styles and the Design of
Network-based Software Architectures.” PhD thesis, University of California,
Irvine, CA, 2000.

[Ke nnytm 11] Kennytm. Rapidjson GitHub Repo. https://github.com/kennytm/
rapidjson, 2011.