Spiral Galaxy Simulation SG-3

A real time simulation of 10'000 stars building forming a spiral galaxy

Get the code here : http://www.ozone3d.net/demos_projects/spiral_galaxy.php?lang=2

How it works (in french) here : http://www.ozone3d.net/blogs/demoniak3d/?p=10

Category:

Film & Animation

Tags:

• space
• simulation
• 3d
• galaxy
• real-time
• rendering

[ZebNemeth]

no, you could'nt.

wit only 500 stars, a thousand frames take about 3 days to calculate.

this was done with thousands, and calculation time expands in powers.

so, if you are ready to wait 50 years, you could maybe do it at home.

[DrGoulu]

this calculation was done in REAL TIME on a standard PC :-)

the trick is that it doesn't simulate gravity. It is a purely geometric simulation. Follow the links for explanations and downloads.

[Capeau]

so, its just some particles flying around? thats not impressive man.

as a matter of fact, any serious animation package can do lots more then 10000 particles in realtime like that

[DrGoulu]

ok so do it with 10'000 and show me ;-)

The problem is to specify elliptic trajectories for the particles so that the spiral arms appear.

[aqsa852jaffar]

IDK why but for some reason that looks like a barred spiral galaxy!!!

[smitty121981]

awesome video! unlike Capeau below, I actually appreciate how hard it is to do this in real time! I do 3d for a living and most animation packages are so bloated with so many features that they could never do this real time. Good job!

[mathcore85]

excellent!!,

i would like to read about how it works in english please,

is this simulation based on the density wave theory?

[Capeau]

do a search on youtube. Theres already quite a lot of them.

The trajectory isnt any problem you can use any spline as a path.

[TheReasonWhyGuy]

Also, a way to simulate physical effects of gravity that is much faster but isn't as accurate is a tree method.

Look up,

N-body

Tree method

Particle mesh method

Using some of these other methods, you can simulate a much larger number of masses in a shorter time.

[CACBCCCU]

"the trick is that it doesn't simulate gravity"

Definitely. You can't get a tight double spiral like that by using Newtonian gravity.

Interestingly enough, if you want to give Newtonian gravity a low-energy-quanta characteristic, then multiplying Newtonian gravity by a long cosine wave is a possibility, and that should produce nice galaxy shapes. With a compact massive central body the negative cosine phase forms a weak spherical gravity hump for smaller peripheral masses.

[ZebNemeth]

Ah, that explains it! :)

Just 2 weeks ago, on astronomycamp, we made our own star cluster simulation, but with gravity.

That was what I was referring to...