The latest issue of EDGE (183) features an excellent article on Automatic City Building for games. Focused on Gamr7, it details their middleware aimed at enabling artists to quickly design and create detailed urban environments.The cities are grown based on the meaning of buildings according to activity. Cities are of course complex and this procedural approach is interesting, a movie detailing the system to date is embedded below:
Note the above movie is only indicative of earlier work, a new demo is coming soon but for now due to NDA's this is all the info we have.
We look forward to seeing the latest version...
Gamr7 aim to have a working release by mid 2008.
I was actually unimpressed with this.
ReplyDeleteI think procedural generation in general is great with things like terrain, trees, (texturing) and well objects which are more organic based, but for objects like a city, which you can't seem to make look good automatically I think you just need to to build it by hand if you want to achieve any sort of realisim and it does not have to take as long as you think as it's mostly just the front.
I'm sure a small team of Artists could knock out dozens of these fascades in a month which would be more than enough for most games.
Secondly the placement takes very little time so even if you can load in your own library of buildings and use this tool as palcement, I don't think the task it replaces is THAT time consuming due to the scope/ease of use of copy/paste/rotate/scale. You also get more control over the design doing it by hand too.
I guess this is ok for really unimportant stuff but I'm not convinced myself.
Jamerio
Ex game designer perfect dark, now doing Architectural visualisation.
I sort of agree with you, the video is unimpressive but its quite old. (Smithee, could you add it on the blog post ?)
ReplyDeleteOur current code is mush more advanced. I cant' post anything right now (NDAs) but we are working on a demo, so stay tuned.
I worked on big games and believe me, the placement is very time consuming.
I have many war stories of making a believable city in a game. building one is very costly/painful.
The library approach you propose doesn't scale very well for big or high-quality cities where density, diversity and editability (at the city level) are needed.
Anyway, I am quite sure our next demo will help convince you.
Best regards,
Have you seen this article about Introversion Software's procedural city generation?
ReplyDeleteYes, we have seen it. It's based on fractals and l-systems.
ReplyDeleteIt's the right tool for they want to do but it don't think fit in a more traditionnal/bigger project when you have more people on the team.
Their tool, while excellent, is not versatile enough, but it's ok, because it has been designed for one unique game. Again, Our next video should explain that much better.
Introversion is quite unique in the industry. Very small team, unique games and (mostly) procedural content.
I have deep respect for what they do. I really like their work and am waiting eagerly for their next game (Multiwinia).