The last row of the bone transformation matrix always is 0,0,0,1 so
there's no point in uploading it. Also reducing the max bone count to 80
which means the uniform array will fit into the available space on 6000
and 7000 series Geforce GPUs.
Instead of transforming all vertices on the CPU every time an animation
progresses, we now only recalculate the skeleton, leaving the heavy
lifting for the GPU. This also means we no longer have to push all
vertices onto the bus every frame, because the mesh isn't changing and
can therefore be stored in a GL_STATIC_DRAW VBO when it's first loaded.
The downside of this approach is that there's only a limited number of
uniforms and vertex attributes we can pass to the shader. At the moment
these limits are a maximum of 128 bones per skeleton, and no vertex can
be influenced by more than 8 bones at once. So far this is no problem,
as the most complex skeleton in the base game uses less than 64 bones
and no more than 6 bone weights per vertex.
Instead of having the default vertex shader hard-coded into the engine,
allow to load it from Graphics.ocg. There's still a fall-back version
wired into the engine because we can't return an error from
GetVertexShaderCodeForPass.
- Increase light fade-in/fade-out quite a bit (4 times stronger)
- Fixed normal jump when multiple lights intersect (will make things a
little less nice with only one light, but this takes priority)
- Don't draw objects without light in light-debug mode
This allows to see the whole landscape without any areas covered by FoW
in the global viewport. Basically the ambient lighting is set to 1.0
independent of the ambient light map. In the course of this, a second
shader for the landscape has been introduced.
This allows to ignore slice declarations using `#define slice(x)`, which
mill be useful for custom mesh material shaders, allowing to write them
such that they can be used standalone in a mesh viewer but also as slices
for OpenClonk, in which case lighting and color modulation will be applied
automatically.
This is the "lazy" fix for the problem that lightness never returned to
full zero. On the other hand, it probably is also the right one - after
all, with smoothing the last 8 or so light intensity values are lost to
us anyway.
Turns out we actually can't optimise out the second texture normal
lookup - it makes a noticeable difference in-game. Also there was
possible slice collision, it might be worth warning about this kind
of stuff.
Also removed unused variable, and made error messages more consistent
in terms of whitespace.
The idea here is that we compose shaders out of "slices", which can
come from the engine ("built-in"), from files or possibly even from
models. This should allow us to more easily share the code between
different rendering shaders (e.g. for lights / normals).
TODO: Workarounds not yet implemented, so this might degrade less
gracefully.
This introduces a new texture, an ambient light map, that is generated
automatically at the beginning of the round by the sky portion of the
landscape. This basically makes everything that is close to sky visible
by default.
The shaders have been adapted so that they deploy direction-independent
lighting for the ambient component, and the current (diffuse) behaviour
for the diffuse component. This makes the shaders use an additional
texture unit that represents the ambient light. We can think about merging
this information into the light texture, but the coordinate systems are
different at the moment, so this could be performed at the stage of light
texture generation.
For meshes, the ambient material is not actually used, but instead a
diffuse light from the front is used. This makes many meshes look more
interesting, maybe also because the ambient material setting of most
meshes are not set correctly at the moment.
Drop the paper in the player selection dialog entirely, use transparent
black instead. Adjust the widget sizes in the scenario selection dialog
slightly.
Also set default to 8. This might be a pretty controversial change, but
the amounts of screenshots we have at zoom levels beyond 4 just calls for
a more high-res approach.
Pretty awkward compromise, but the array version seems to confuse a lot of
drivers, without causing an actual error. So this is probably better in
practice until better drivers are more common.
Also set default to 8. This might be a pretty controversial change, but
the amounts of screenshots we have at zoom levels beyond 4 just calls for
a more high-res approach.
Pretty awkward compromise, but the array version seems to confuse a lot of
drivers, without causing an actual error. So this is probably better in
practice until better drivers are more common.
- The light direction change isn't really consistent, as I can only give
GL the vectors at the triangle edges. So for deep rays, the light
distribution often ends up being significantly different than for short
rays.
- I *am* distorting the normal map vectors, quite badly actually. I'll leave
it for the moment, because it definetely looks best for the light vectors,
but I will have to correct that a bit once I have the time.
Some compilers are less strict about this than others, unfortunately.
The Gallium one complains with
preprocessor error: Macro names starting with "GL_" are reserved.
beliar reports that his driver doesn't support the number of fragment shader
uniforms that we're using in the landscape scaler. By making the reported limit
available to the GLSL code, it can use a workaround using a 1D texture to
transfer data.
Use by having texmap entries of the form "Mat-Tex1-Tex2-...". Right
now the speed is hard-coded to one phase per second.
The general idea is that the 3D texture contains all texture
transitions somewhere in the form of two textures with neighbouring
3D coordinates. There's some room for optimization here, of which
the code exploits some. Being smart can be arbitrarily hard actually.
+ CustomMessage now uses a normal picture of a definition/object
+ BigIcon is now the small (and only) "Portrait" the player chan choose
+ In future, one could of course make the max size of the BigIcon bigger
Also means we will have to think about mip-mapping on the scaler
soon. We might want to disable the scaler for zoom levels < 1
anyway - but right now the shader isn't flexible enough for that.
Using the walkaround mechanism we can now safely try whether the
driver has a problem with texture2DLod getting used in the
fragment shader. It might be against the standard, but I see no
reason why I shouldn't try - it did improve things for my old
driver.
Now using the best algorithms I could think of, meaning that quite
a lot of information is stashed into the scaler texture in order to:
1. Reliably find the "other" pixel to interpolate towards using the
scaler. The approximation used before would use invalid texture
coordinates when interpolating.
This now allows us to calculate the shading for this "other"
pixel as well, which actually caused even more visible artifacts.
2. Decide how to calculate the shader for the point rendered. In
case we are in the middle of a material, we want GL to interpolate
the normal so we get a nice smooth sub-pixel gradient. On the
edges, however, we would get interpolation between normals of
different materials, which looks ugly.
Now especially has a few debugging options that make it easier to
spot when texture coordinates are not calculated accurately. Use
DEBUG_SEPERATOR_TEXTURES to add nasty stripe textures that should
(in theory) never be visible. Set DEBUG_SOLID_COLOR_TEXTURES on top
of that to replace all materials with dark solid colors to make the
stripes more clearly visible.
Nicolas Hake