Lighting

[Jorzi]

Yes, the landscape has always had gouraud shading (one of its key features when it was released) and therefore I was wondering why the models didn't...
Of course it would be good to be able to disable it, since most current models wouldn't look very good with it.

About smoothing groups, they don't need to be separately supported since they simply work by splitting the mesh and using duplicate vertices.

[MaNGusT]

Of course it would be good to be able to disable it, since most current models wouldn't look very good with it.

Agreed. it will be good if we could switch them on/off in debug menu for a test purpose only.

[Jorzi]

Agreed. it will be good if we could switch them on/off in debug menu for a test purpose only.

Hmm I don't know why they would be needed in debug mode, I just meant that old models should probably be rendered as they are, while new models would be smoothed with sharp edges pre-defined by splitting the mesh (smoothing groups).

[MaNGusT]

Hmm I don't know why they would be needed in debug mode, I just meant that old models should probably be rendered as they are, while new models would be smoothed with sharp edges pre-defined by splitting the mesh (smoothing groups).

I meant that till the new models aren't completed, there is no need to write a double code for the old and new models.
We can test a new models by pressing "on" button in debug menu and when they all will be completed then smoothing could be set statically "on".

[Jorzi]

Ok, but that would make it impossible to use unsmoothed models, which could be useful for some buildings etc, which is why I'd prefer to simply have the gouraud flag back...

[MaNGusT]

Ok, but that would make it impossible to use unsmoothed models, which could be useful for some buildings etc, which is why I'd prefer to simply have the gouraud flag back...

yep.. it would be awesome..

[Safety0ff]

The normals are calculated in a very naive manner. This should be improved one day. I think the right place to do this is in the C++ code.

I disagree, vertex normals should be incorporated into the model format, this way we can have the model editor do all the heavy lifting, and the same normals that are used in the editor are used in game (this matters for tangent space normal mapping).

Btw, gouraud shading is turned on, but like per said, the normals are specified in such a way that it is the same as if it used flat shading.

[Jorzi]

I see....

Having the normals integrated into the .pie format sounds sensible, keeping the game code a bit simpler.
On the other hand, it breaks backward compatibility and makes the .pie format a bit cluttered, especially since we probably want it to store the tangents (and binormals) as well...

This is the reason I prefer object space normal maps, despite taking up lots of texture space

[Per]

I disagree, vertex normals should be incorporated into the model format, this way we can have the model editor do all the heavy lifting, and the same normals that are used in the editor are used in game (this matters for tangent space normal mapping).

Oh, if we can get the model editor to do that job for us, I'm all for that. </lazy>

[Safety0ff]

On the other hand, it breaks backward compatibility and makes the .pie format a bit cluttered

I'm not too worried about this, I'm thinking whole new format, scrap backwards compatibility.

[Jorzi]

Ah yes.. after all, you know what you're doing while I'm still learning syntax and stuff...

[milo christiansen]

Kind of like resurrecting the WZM format?

[Safety0ff]

Kind of like resurrecting the WZM format?

If by resurrection you mean something completely different with the same extension, then yes.

[Jorzi]

I was about to request an additional specmap texture for use in the fragment shader, but since we apparently need a new file format anyway, I guess the implementation can wait...
However, a question about lighting, why is the main light currently a point light and not a directional light?
Instead of having a constant direction vector for the light, it seems to have a specified position and the light vector is calculated for each vertex by subtracting the vertex coordinates from the light coordinates...

[MaNGusT]

I was about to request an additional specmap texture for use in the fragment shader, but since we apparently need a new file format anyway, I guess the implementation can wait...

did you say about the specular color texture?