Background Rendering

[Wumpf]

We need something spacy in the background! The following lists try to gather requirements and possible solutions. Discuss and edit

Must have:

• Stars (big, small, white, colored)
• Large Nebulas - probably just "the galaxy"
• Small Nebulas, "objects"

(very) optional:

• Slow animation \ Probably just rotating the whole thing super slowly is nice enough
• Different depending on position - travel through space!

Solution ideas:

• Precomputed skybox
  • from disk
  • Just take some open available skyboxes, there should be plenty of them
    • Looks nice and realistic out of the box
  • Most likely unmatchable in quality
  • License? Probably hard to find good ones?
  • Draw some by yourself
  • procedural
  • Simple results are easy, good results are its one scientific field
  • insert link here; some of the old sources I died apparently
  • General Advantages
  • Usually very fast to implement and in runtime performance
  • General Disadvantages
  • Need gigantic resolutions for accurate star rendering
  • Forget about animation!
• Rendering billboards on the fly
  • slower!
  • difficult to place stuff
  • much overdraw for rich environement
  • high precision, almost no resolution bounds
  • can mix photographs together
  • animation very easy
• A single post process pass on the fly
  • Need some demo-coder like ultra-formula
  • very high precision, no resolution bounds
  • animation easy
  • most complex objects are difficult
  • probably rather fast (only rendered where nothing else can be seen)

Links to other games

• Kerbal space program and their modders doing great
• TODO

I'am not sure which solution is the best. We definitely need to investigate where people before us succeded/failed. Also we need photos/screenshot that express what we are aiming for. We also should take into account that we do something non-photorealistic, consisting out of cubes (MANY cubes).

Since this is not maximum priority at the moment, but one of the most important graphical features, we should try to create a first non-perfect solution.