As most Google Tech Talks, this is way above my virtual tea cup - which prefers to just travel and be amazed - but even grasping not the complete 50% this is interesting. For virtual worlds to have a future, we will definitely need well-build ‘cities’ - or urban areas - most probably even ones that resemble their First Life bothers and sisters very close. The problems pointed out in this video, the ‘Content Challenge’ is definitely true for Second Life:
Architectural content like cities, buildings and interiors is extremely important - but also very complex.
No tools available for efficient creation of detailed architectural 3D models.
The solution to this presented here is ‘shape grammar’ which leads to automatic content creation. They call this procedural modeling. Very, very, very much simplified, this means, that once you have your ’shape grammar’ in place, you give the computer a map and it figures out which buildings to build, where to put roads, what the buildings should look like, … which gives a very uniform results. Of course, did I mention this is very much simplified? ;)
You then can fine-tune the resulting city, buy making small are larger changes to it’s ’shape grammar’. For instance, you’ve build a Roman city, but used the wrong time area and inserted Doric columns in stead Ionic ones? Easily corrected by changing the program’s grammar. Again, I must add ‘very much simplified’ to this.
The procedural modeling workflow shortly summarized:
Architectural design idea
Analyse design and it’s parameters.
Create and define:
Create or get elements/textures.
Define city layout /initial shape(s).
Encode design: rule set(s)
Add stochastic behaviour to rules.
Apply rule set(s) and export models.
All this has lead to a commercial program called The CityEngine that does procedural modeling of CG architecture and is set for release May 2008. The use they see for this? Gaming, entertainment, urban planning, archeology, … . I’m sure you can think of some use for this of your own. Impressive, nèh?
Google Talk explenation on ‘Urban Reconstruction and Modeling for Building Virtual Worlds’: Creating digital content for virtual worlds remains a significant challenge, especially for urban environments, which are among the largest and most complex. As display capabilities improve and audience expectations grow, procedural modeling techniques are becoming an increasingly important supplement to traditional modelling software. In this talk, we present grammar-based, image-based and interactive methods for the efficient creation of urban environments. Thus massive architectural models of high visual quality and geometric detail can be produced at low cost. Selected examples demonstrate solutions to previously unsolved modeling problems, especially to consistent mass modeling with volumetric shapes of arbitrary orientation. Furthermore, we show massive urban models with unprecedented level of detail, with the virtual rebuilding of the archaeological site of Pompeii as a case in point.