• Advertisement


  • Content count

  • Joined

  • Last visited

Community Reputation

127 Neutral

About Zellski

  • Rank
  1. Seamless Noise

    bull_dog: it looks to me like these parameters map out the squares within the 4D space within which the circles are drawn. In other words, they define where in the 4D space the samples should be taken (X1) and over how large an extent (X2-X1). Since this type of noise is deterministically pseudo-random, i.e. it always comes out exactly the same given the same input parameters, you have to alter (X1, Y1) to get different chunks of noise. You'd pick (X2, Y2) to be (X1, Y1) plus however much of the 4D space you want to consume. If you grab a lot of it, it's going to be more chaotic. If you grab a small portion, it's going to be smoother. It seems to me that really these parameters could be (X1, Z1, Radius1) and (Y1, W1, Radius2), but I admit I can't think of an example where this would be all that useful. Thanks JTippets for this post!
  2. Quote:Original post by Yvanhoe I think the difference between spore and most GA simulations is that GA simulations are made with robotics in mind : create a program that could pilot a robot with real constraints, real physics, etc... whereas the task spore must achieve is simpler : make the most realisticaly-looking walking pattern for a given creature. I don't know the first thing about Spore, really, but yeah, this is a pretty important distinction. Almost all previous research on generating motion for virtual creatures starts with the assumption that it must be physically valid (and high level constraints such as precise foot placement). The fact is that for entertainment purposes you can relax things a fair bit and things still look just fine, and take a fraction of the time to compute. Fully physically compliant evolved controllers don't really look that hot. The top results in the research and commercial world, as far as I can tell, can all learn to do things like walk and pick up guns or whatnot, but it is an undeniable fact that they look like they've suffered minor brain damage at some point. In terms of real grace, you can create vastly better motion with spacetime techniques, where optimization algorithms construct a path through the state space that minimizes something along the lines of 'metabolic energy consumed in the creature's muscles'. There is much literature on this subject, search for e.g. Witkin/Kass' paper on Luxo (that's the original one from 1988) or Michael F. Cohen's follow-ups through the early 90's. More recently, in the vein of relaxing the demands for physical accuracy (the full equations of motions are extremely non-linear and very difficult to solve quickly), Zoran Popović et al have done a lot of interesting work in this area in the past few years, especially working with momentum conservation and the like. Take a peek at http://grail.cs.washington.edu/projects/charanim/ as an example. The problem with these approaches in general is that they are basically 'offline' in nature, and of limited use for simulation. Once again, I don't know what Spore does -- this was just a general ramble . Feel free to PM/Email me if any of this stuff interests you...
  • Advertisement