Solar Model 2.1 - 3D real time modeling of Solar System

[hybrid]

Ok, you forgot to recalculate the mesh bounding box which made the tests against it quite wrong

[greenya]

Ok, you forgot to recalculate the mesh bounding box which made the tests against it quite wrong

am i ?
i think that you talking about the line:

Code: Select all

mb->recalculateBoundingBox();

but it was present and the object still disapers some times.

p.s.: the source code can be checked at few posts sooner.

[hybrid]

No, a mesh->recalculateBoundingBox() is missing, which sums up all bboxes from the meshbuffers.

[nikitale]

Cool!!! I really enjoyed pressing 4 and chasing Earth

[Jesus.Net]

Go, it is incredible, it has left me of a stone. I am completely impressed.

[greenya]

Hi again

Solar Model 1.21 Released

a little bit more changes:
+ Added Asteroid Belt.
+ Add Kuiper Belt.
+ Fixed Saturn polar diameter.
+ Improved source compatibility (now the same source compiles under Windows and Linux).
+ Updated to latest IrrLicht library (version 1.3, revision 636).

Really hope that now it is easy compiles under Linux (thanks for help to hybrid)

Hope looks not very bad for you:


Asteroid Belt you can see on the image,
Kuiper Belt is visible after Neptune orbit, so you should fly long distance, the objects moves very slow at that distance from the Sun (press "4" for estimation).

More screenshots, binaries and sources can be downloaded from http://www.ffsoftworks.com/solar_model/downloads.php

P.S.: For compatibility reasons Solar Model runs in 800x600 windowed mode, BUT if you want, you can change this -- edit config.dat. In 1280x1024 fullscreen mode it looks very nice.

[sio2]

Very nice.

The shading on some of the planets (and other spheres) doesn't look great though. This is due to the fixed-function shading.

Rather than just complain, I've done something about it. I've added the ability to specify a shader in the material definition and added a GLSL shader to compute per-pixel lighting.

Example:

Code: Select all

AnimatedMesh
media/sphere.x
{
	Object {
		Name Mercury
		Rotation 90.0 0.0 2.0
		Scale 0.05 0.05 0.05
	}
	Material {
		LayerTexture media/Mercury/diffuse.jpg
		ShaderGLSL media/shaders/glsl/PerPixelLighting
	}
<other stuff>
}

Before - note the zig-zag shading (interpolated per-face lighting):


After:


I need to correct the ambient lighting and then I'll upload the new demo and source.

[Ico]

Maybe I just haven't checked close enough but I miss some functions (that should be quite easy to implement):

Display the orbits of planets and moons.
Reset the timer.
Set the timer to the current real time (I'm sure the data for this is somewhere out there).
Would be nice to be able to (i.e.) see eclispes etc.

[arras]

sio2 >> doesnt Irrlich support per pixel lighting?

[greenya]

I need to correct the ambient lighting and then I'll upload the new demo and source.

Cool. I'm waiting
P.S.: i was wondering why the shading not so good, but i though that it is my video card ... now i see that was wrong

2 Ico
I plaining to implement displaying the orbits of planets and moons. About real current positions of planets -- if i will found the info about it -- i can implement it ( but i'm not sure about moons ).

[hybrid]

If you want to improve the shading for fixed pipeline you have to use more vertices in the sphere. In case you're using the Irrlicht CSphereNode you have to raise the initial parameters. this will reduce the artifacts.

[sio2]

If you want to improve the shading for fixed pipeline you have to use more vertices in the sphere.

So you're belittling my contribution? My method isn't good enough? I sometimes wonder why I bother...

Do it Hybrid's way. I'll keep my stuff to myself...

[hybrid]

I do wonder what you IrrSpintz guys think about yourself. Either try to calm down and reread my post or think about ignoring my posts completely (which seems fortunately unnecessary nowadays ).
But just in case someone else could really misinterpret my last post: I simply said what would help in case you want to keep the app compatible with old hardware. This is also perfectly compatible with sio's suggestions (just do a driver check for shader support and use higher detail models otherwise, or do both if FPS is no problem on high-end cards).
Or are you just fishing for compliments? Than this was a pretty lame one. Next time ask directly, maybe I find a possiblity to rock you to sleep (But seriously, I never critizised any of your work, was sometimes the first one to post admiring words to your examples. What's the point?)

[greenya]

If you want to improve the shading for fixed pipeline you have to use more vertices in the sphere. In case you're using the Irrlicht CSphereNode you have to raise the initial parameters. this will reduce the artifacts.

I do not know any CSphereNode class.
I know irr::scene::ISceneManager::addSphereSceneNode() which returns ISceneNode* .

Indeed I using single model for all planets and moons because I do not know how to specify texture coordinates for the node after creating it using addSphereSceneNode().

P.S.: the model file contains mesh and texture coordinates so I just specifying another texture to the model, scaling it, rotating and getting new planet or moon.

P.P.S.: I do not like idea to increase vertices count because for now i getting result like:
- on Radeon X200M ( it is like Radeon 9100 ) -- 9-15 fps ( also Doom III gives 10-12 fps on the same config )
- on GeForce 7600GS -- 60 fps ( because of VSync enabled ) .... but It is too expensive for such graphics as renders Solar Model, don't you think ?

[Ico]

At least you can now tell everyone you wrote a program with such beautifull graphics that it's running as slow as Doom 3.

Hybrid was talking about the addSphereSceneNode().

Haven't taken a closer look but how do you draw all the small asteroids etc.? Maybe you could code some simple check that makes them disappear once the fps drop below 15-25 fps.