next »

[Marvin Fröhlich]

That's OK.  I guess I was a bit vague saying I wanted decals and actually wanting projection .  Yes, I think that the GLSL Shadowing will be fine.

Well, actually I I was a bit vague saying, too . What I meant was, that the GLSLShadowMapping algorithm could be partially reused for texture projection. You would spare the RenderToDepthTexture part and just project a pre-loaded Texture the same way, the ShadowMapping algorithm does, this would be faster, but of course less pretty. Is this, what you initially wanted?

Marvin

[kukanani]

Right, that's perfect.  Thanks  .

[karmaGfa]

1 feature that I wish Xith to have is to not force the user to load libraries that he won't use. In other words, I wish Xith to have less dependencies with the libraries it is using.

Actually, when someone wants to use Xith only for rendering (which is still its main purpose), without inputs handling, physics, particle engines or else (network?), he still need to have all the libraries.

Example of forced dependency: HAIL, which requires JInput, SWT, LWJGL, etc ...

Of course, we can choose to not have those libraries at runtime if we are sure (but that's hard to say in practice) that our program don't use them. But we still need to have them all when we want to compile Xith3d.

In my case, I just want to use the renderer without anything fancy, I am just using JOGL_SWING, and I am directly using JInput outside of Xith. The best I expect from an engine like Xith is to not force me to have included in my project each one of the libraries supported by it if I don't need them.

In other words, what I want is ... Freedom !!!!!!!!!!!!!!!!!! 

I know that it is easier to say that to have, especially when we still want to keep the implementation simple. But be sure that if I find a way to do it and keep the source simple, then I will let you know about it. 

[Mathias 'cylab' Henze]

I actually have a different opinion. If there is a library that offers some features, we shouldn't hesitate to use/depend on it. As long as you don't use the features you can freely omit third-party libraries in the runtime-installation, so I don't see any problem here.

[Marvin Fröhlich]

Well, the only library, that you need to run it, which you might not need, is HIAL. You definitely don't need to have LWJGL or SWT in the classpath the run from a xith jar.

What most of us do, is checkout the project and link your own project with the xith3d project. This way you don't need to have any libraries outside the xith3d project and your project will compile fine. You don'T have to worry about any dependency. I don't know, how this is done in NetBeans. But I'm sure, there's a way .

Does this help?

Marvin

[karmaGfa]

If there is a library that offers some features, we shouldn't hesitate to use/depend on it.

At least, I totally agree that we should not hesitate to use it.
But I don't see the need to depend on it, specially if we can find a simple way to avoid it.

As long as you don't use the features you can freely omit third-party libraries in the runtime-installation, so I don't see any problem here.

The problem that I wanted to point is that if xith depends on it, we can't be sure that it will still work without, even if we think that our program don't use it (as I said, in practice it is hard to make sure that we won't have a class load problem at runtime ).

To first depend on it and then to pretend at runtime that we don't depend on it is a kind of self lie, a hack. If we don't have to depend on it at runtime, then maybe it means that there is no reason to initially assume that we should really depend on it in the first place.

Vincent

[Marvin Fröhlich]

We really, really don't depent on LWJGL/JOGL/SWT at runtime, if it isn't used. Only the Xith3DTestLauncher (DisplayOptions) and the display-options HUD Menu will depend on it apart from the renderer itself, if it uses the appropriate layer.

About HIAL: If you have a good idea, how to avoid this dependency, just tell me. But currently it is highly integrated and is not easy to extract without a replacement.

Marvin

[Marvin Fröhlich]

I have added projective texture mapping support. Just have a look at the org.xith3d.test.texture.ProjectiveTexturesTest.

Enjoy .

Marvin

[kukanani]

Thanks.  It looks good, except for me...

http://xith.org/forum/index.php/topic,789.msg6944.html#msg6944

[Rasenger]

One very important feature that Xith lacks is per pixel or Phong lighting. I already thought of moving to jMonkeyEngine becouse it supports Phong lighting, but then I thought it would be too troublesome and I want to support this great project. No-one wants to look at vertex lighting while playing modern games, it just looks terrible. Even with very high polycount it's not as good as Phong with low polycount.

[Marvin Fröhlich]

One very important feature that Xith lacks is per pixel or Phong lighting.

This is not correct. It is supported. But you have to handle the shader programs on your own. Just have a look at org.xith3d.test.lighting.ViewSpotlightTest.

Marvin

['n ddrylliog]

C'mon, Marvin, that's not an answer.

We can also say, "OpenGL supports Phong lighting", but that wouldn't be fair. People use high-level APIs like Xith3D for a reason. Using phong lighting should be as easy as model.setLightingMode(LightingMode.PHONG);

Integration, that's what people want. There's no use in everyone reimplementing their phong version and having to climb the shader learning curve each time.

[Marvin Fröhlich]

C'mon, Marvin, that's not an answer.

We can also say, "OpenGL supports Phong lighting", but that wouldn't be fair. People use high-level APIs like Xith3D for a reason. Using phong lighting should be as easy as model.setLightingMode(LightingMode.PHONG);

You're definitely right. And Mathias and me are already working on a shader based rendering pipeline (on a long term). This way we'll have the opportunity to have a seletable ligting model.

Integration, that's what people want. There's no use in everyone reimplementing their phong version and having to climb the shader learning curve each time.

But it would be nice, if someone implements a phong shader in GLSL, would share the shader code back to the xith3d project . This is how it should be.

Marvin

[Mathias 'cylab' Henze]

I am currently working with a friend on a hobby project containing GLSL shaders for phong lighting, shadow mapping, bloom, tonemapping, normal mapping and depth of field. While the shaders might not be directly transferable, all these features will come to xith in mid term.

['n ddrylliog]

Well that's for some good news.

Marvin, I was afraid you'd react somewhat like in the old fortune, d'you know it? it's the one which goes like:


Q:      How many hardware engineers does it take to change a light bulb?
A:      None.  We'll fix it in software.

Q:      How many system programmers does it take to change a light bulb?
A:      None.  The application can work around it.

Q:      How many software engineers does it take to change a light bulb?
A:      None.  We'll document it in the manual.

Q:      How many tech writers does it take to change a light bulb?
A:      None.  The user can figure it out.