summaryrefslogtreecommitdiff
path: root/vis.c
Commit message (Collapse)AuthorAgeFilesLines
* use the new light types constants.rodri2024-08-171-3/+11
|
* let Model.tex have precedence over any material.rodri2024-08-161-9/+9
|
* vis: add options for the clr color, blending and z-buf and a-buf.rodri2024-08-131-13/+93
| | | | | also stop setting alpha always to 1. it's not correct yet but it allows for transparent objects to be blended properly.
* use the new readobjmodel function.rodri2024-08-071-8/+4
|
* vis: use the new viewport upscale interface.rodri2024-08-021-13/+22
|
* improve redraw responsiveness.rodri2024-07-201-8/+11
|
* vis: don't allow peeking outside the framebuffer.rodri2024-07-151-0/+2
|
* vis: add viewport upscaling support.rodri2024-07-141-11/+32
|
* fixes.rodri2024-07-141-6/+10
|
* use the new camera interface. solar got faster with backface culling.rodri2024-07-131-27/+29
|
* vis: add knobs to control culling.rodri2024-07-131-0/+21
|
* fix the toon shader.rodri2024-07-131-1/+5
|
* get rid of redundant draw lock and adapt to libgraphics changes.rodri2024-07-131-11/+9
|
* adapt to recent changes in libgraphics.rodri2024-07-091-14/+9
| | | | | | | also set the normals correctly in the phong shader after mapping from tangent space. that way we can see the actual normals being used in the normals framebuffer.
* adapt to the new texture interface and fix the identvshader.rodri2024-07-051-11/+18
| | | | | | | | the identvshader was used for the toon shader as well, and that forced it to do a light computation when its fragment shader won't ever use it. split it in two vshaders, each for their own purpose, so we don't get slow identity rendering.
* med,vis: fix normal map sampling.rodri2024-07-041-2/+3
| | | | | | recent changes in libgraphics assume every texture resides in an sRGB color space, which is not the case for material resources like normals.
* vis,med: make qball rotations position-independent.rodri2024-06-161-3/+3
| | | | | the camera no longer needs to be on the first quadrant for the qball to work.
* vis: add some instrumentation to debug the framebuffers.rodri2024-06-141-1/+37
|
* use the new modulapt3 functions. don't clamp in the shaders.rodri2024-06-131-49/+30
|
* add skyboxes based on cubemaps.rodri2024-06-061-14/+24
|
* vis: same as previous commit, and add normal mapping to the phong shader.rodri2024-06-041-29/+53
| | | | | implemented tangent space normal mapping and now diablo3 looks sick!
* get rid of rendering latency by decoupling it from i/o.rodri2024-05-231-12/+28
| | | | | | | | | | | i was using a single alternator to mux both i/o and drawing in a single proc, and avoid locking. the problem is that it causes a fight to break up every time there's any input (steady 100Hz) or a lot of drawing requests from the renderer (up to 60Hz.) i added a drawproc to handle exclusively drawing requests, and a drawing lock for UI widgets that kidnap the display.
* fix the gouraud shader. improve the cube generator in med.rodri2024-05-231-28/+32
| | | | | | | | gouraud wasn't being applied when the vertex had no material, now it just takes the defaults. also added a new primtest.obj to visualize all the primitives in a single model.
* add a triangular basis and clamp the fov to [1,180]°.rodri2024-05-161-2/+2
|
* fix the qball.rodri2024-05-151-5/+4
|
* use real lengths in solar. bring a qball for vis.rodri2024-05-141-24/+34
|
* vis: print pipeline timings to stderr when profiling.rodri2024-05-071-0/+6
|
* adapt to the recent changes in libgraphics. add new test files.rodri2024-05-031-27/+19
|
* mkone → mkmanyrodri2024-04-241-0/+853
crucial step to begin exploration of different 3d environments and their tools.