Atomontage Engine

   Gallery
    + Destructible Vehicles, VR...
    + Mesh Voxelizer
    + Hi-Res Boulders
    + Destructible Geometry
    + more...

Download

   Dev Blog
    + #4 Destructible Vehicles,...
    + #3 Voxelized Geometry
    + #2 Evolutionary Design
    + more...

Contacts

   Donate
    + Make a Donation
    + List of Companies
    + List of Individuals

F.A.Q.

FOLLOW AE ON

Dev Blog

Dev Blog #1.3 | January 3, 2011 | Reading forum stats...
The Future Is Volumetric, Atom-Based 3/3

Jump to: Part 1/3 Part 2/3 Part 3/3

Part 3 of 3 - Destructible Geometry (1st half)

The third part focuses on the interaction with the virtual matter the content is made of. Atomontage Engine™ can handle large volumes of matter consisting of different physical materials exhibiting different physical properties.

Watch and comment this video on YouTube.
Visit our channel on YouTube.
Go to Gallery to view high quality screenshots.
Content Generation
The presented scene was populated with millions of unique pieces of geometry using single-pass and multi-pass generators. The sand as well as the stones were generated using simple single-pass generators. These generators analyze existing local geometry and plot a piece of new geometry based on a set of parameters.

The rock as well as the buildings were generated using multi-pass generators. In the first pass the large-scale structure was defined and stored volumetrically. In the second pass surface and volumetric details are added and voxels plot into the volumetric scene.

The generators set all the required attributes of the voxels automatically so that each voxel is in a way physics-enabled no matter whether it is a surface or a sub-surface one. Later the real-time part of the engine can identify the particular material of composite objects volumetrically on a per-voxel basis so that physics-based features can be performed in a plausible way. This feature can be very useful in various processes like erosion, explosions and similar. The engine automatically maintains the integrity of the visual as well as physical attributes of the content after a destruction event occurs..

First-person View
In the video it can be clearly seen that the spatial resolution of the voxelized content is too low to get decent picture in a first person view. Soon I will release screenshots or possibly a short demo video showing a higher LOD of the same content.

Material Composition of the Content
The scene in the video consists of seven different materials. For example there are about 18,174m³ of rock while only about 24m³ of mortar. Each of the materials responds differently to the simple erosion tool you can watch in action in the video. The tool removes particular voxels based on the strenght of the material.

It is the same with the wheeltracks. Currently the particle simulation is disabled. Therefore the wheeltracks are static and don't collapse the way it is expected. I will re-enable the particle simulation once the prototyping of the volumetric interaction functionality is finished.

Part 3 of 3 - Destructible Geometry (2nd half)

Watch and comment this video on YouTube.
Visit our channel on YouTube.
Go to Gallery to view high quality screenshots.
The possibility to work with a fine volumetric structure of composite objects will be a great feature of the coming generation of volumetric game engines. This functionality will make mechanical as well as chemical and thermal processes result in unprecedently complex, plausible behavior of things. We can expect the complexity of games and similar software to start to grow exponentially once physics-driven synergical effects start to emerge in the complex interactive atom-based world. One day synergy will become key in making playing games a truly realistic experience. And it's the atom-based representation that will make all that possible, soon.

Fast Processing of Voxel-based Content
As can be seen in the second half of the video the removal tool can remove large amounts of material quickly. The engine can access voxels one by one as well as in groups of few or few hundreds of voxels at once. The scene is clearly segmented and all the math necessary to access the data this way is trivial and easy to optimize. The removal tool can remove up to a few dozens of millions of voxels a second including the reconstruction of the new surface. The code is not optimized, yet.

See hi-res images in Gallery
Visit and leave a comment on YouTube: 1st half 2nd half Our channel
Make your name or message appear in the next videos and/or demo.