Home Technical Talk

Managing Units over Large Distances?

polycounter lvl 10
Offline / Send Message
paradoxical-pixel polycounter lvl 10
I am about to start a UDK project based-upon the International Space Station, and a very fundamental problem has arisen.

I want to create the ISS, Earth, Moon, and Sun in accordance with their relative proportions, positions, and orbits. However, this may not be possible due to the maximum coordinate limitations of most (if not all) 3D software packages.

The minimum distance I want to express is 1 cm (i.e. components inside the ISS), and the greatest is 150,000,000 km (i.e. distance between Sun and Earth). That disparity as a ratio is: 1:15,000,000,000,000. That is 150,000 times greater than 3DS Max's highest supported coordinate value.

Obviously, there must be a get-around since many other space games/animations already do a good job of representing these distances.

Must I reduce the distances between the ISS, Earth, Moon, and Sun, and then scale them down accordingly, or is there a more elegant solution? My concern with faking the proportions is that the parallax will be too great when moving the camera around the Earth's orbit.

Does anyone have a solution?

Many thanks,
Parapix.

Replies

  • Xazas
    Options
    Offline / Send Message
    Xazas polycounter lvl 7
    A while ago I read about the people making Kerbal Space Programm having the same problem. Although they already scaled their universe down by a factor of ten, I hope this helps you in any way:


    http://forum.kerbalspaceprogram.com/entries/54-Scaled-Space-Now-with-100-more-Floating-Origin!
  • paradoxical-pixel
    Options
    Offline / Send Message
    paradoxical-pixel polycounter lvl 10
    Interesting approach. Thanks. :) I hadn't even considered that jittering might arise from reduced floating-point accuracy!

    My thinking now is that the most realistic approach would be to reduce the distance between celestial objects (scaling them as required), and then parent the distant objects (i.e. Sun, stardome, etc) to the camera such that parallax is eliminated. This should be an acceptable method since everything will be seen from the vantage point of the ISS.

    Can anyone see any problems with this, or provide a better solution (preferably one that doesn't involve complex matrix transformations)? :icon_question:
Sign In or Register to comment.