No announcement yet.

Destructible Environments

  • Filter
  • Time
  • Show
Clear All
new posts

    Destructible Environments

    Hi Guys,
    I am after some information please, currently I am conducting research for my dissertation at university regarding destructible technology in games. Im hoping that any level designers out there might be able to give me an insight into what considerations would arise as a result of including a destructible environment. Im interested in knowing what approach you might take if the environment was fully destructible (as far as it can be) or if the amount of destruction was to be limited (for example indoors where layers of plaster/bricks could be destroyed).
    Also I am after the opinions of environment artists regarding the considerations when creating assets for a destructible environment and what approach you as an artist would take regarding workflows, potential problems etc.
    Thanks in advance

    You'll hit into pathfinding issues if everything is destructable. The game may be annoying to play if you got to jump and crawl through rubble all the time. The current fracture mesh system will probably not suffice if you want to do this on a large scale, it doesn't offer enough control. A major challenge will be in getting large things to fall down and collapse after removal of the majority of supports. For example have a building collapse as soon as you remove the carrying walls on one side of the building. That is hard to do.

    If you keep the scale down you should be pretty alright.


      You can achieve this if you have bigger structures that aren't destructible but should fall when their supports are removed.

      You can add physics to it and give fracture to the supports. When they are destroyed the floor/ceiling should fall. (You could also add an animated sequence for it)


        Performance is a pretty big consideration too; lots of destructible objects will probably lead to an in-game significantly larger number of objects with some form of physics simulation (i.e rigid bodies), as well as them all being dynamically lit.


          Thanks for the speedy replies guys! Are there any other issues you could think of while designing a destructible environment? Are there any issues that an artist may encounter?
          Thanks again for the replies.


            Isn't a dissertation supposed to be original work?

            To avoid reiterating what everyone else said, I will leave out performance issues.

            Personally I've had trouble trying to get destructible objects to do what I want them to (both with the UDK and other engines). As an example, a building that's textured with both wood and stone, the two materials are totally different but making them react exactly how I want them to is often a pain in the ***.

            With the UDK, there are a lot of issues to be had, as previously stated by others, because of the lack of control of it. Obviously not all materials are going to break apart into shards and chunks like they do (or do for me, although having not really played with it that much my opinion on it is mostly invalid). I'm using it for a project right now and for prototyping (but really I do nothing anything above absolute minimum for prototyping because I don't like doing it that much), and I'm trying to figure out how I want it to work exactly by shooting stuff out of the city a ways and seeing what happens to it (for science). I COULD probably get what I want by making everything in pieces, but I am not exactly keen on having fifty thousand pieces to place (that may be a hyperbole, I have no idea how many pieces I would need to use but it's a lot). Which obviously is stupid for a multitude of reasons.

            What would be preferable would be a far more intuitive system, and likely a lot more powerful computer than I have to run, that would do some calculations beforehand, but could work mostly at runtime (i.e. a handgun is going to do a lot less damage than something like a large caliber rifle, regardless of what it's hitting, so logically they should not blow out the same area). Again though, this is on my limited experience with the one in the UDK, so someone might reprimand me later.

            And, irrelevant to the topic, think I could read it when you finish with it? I'm doing a similar paper on AI (I guess not really similar, but hey, they're both video games), although it's for a class assignment.