If you want to refresh VBOs (and thus have geometry changes reflected in renderer), you need to set dirty flags on geometry object.
There are separate flags for different buffers (as not always all buffers need to be updated and oh my is updating costly):
mesh.geometry.__dirtyVertices = true;
mesh.geometry.__dirtyNormals = true;
mesh.geometry.__dirtyUvs = true;
mesh.geometry.__dirtyTangents = true;
mesh.geometry.__dirtyElements = true;
That was quite tough feature, a lot of refactoring, yet performance is still quite bad :(
The biggest remaining bottleneck seems to be translation between Three.js internal data formats and buffers. I removed as much per-frame arrays creation as I could, but even just iterating through existing data and setting of values is still very costly. Also per-frame normals computation is expensive.
Unfortunately performance issues are still there and now already trickled down also to dev channel :(
I wonder what they did - fps numbers look more or less ok, problem is more with subtle yet very annoying choppiness. Like if somewhere suddenly bandwidth / memory is getting trashed.
Also first time I noticed issues with stats widget in latest Chrome canary 10.0.624.0 / Chromium continuous build 10.0.629.0:
- it doesn't show in demos with mixed canvas / WebGL content (OBJ converter demo, lights test)
- simply just adding it can significantly decrease performance (OBJ converter demo is choppy when invisible stats element exists)
- in some demos it has different transparency (Walt cubemap demo)
Looks like with some Chrome update in last few days we hit some new compositing or 2d canvas issue (both OpenGL and ANGLE versions share the same issues).
As expected, performance of one-to-one conversion from CanvasRenderer was horrible, so I had to change a bit how lines are handled.
Line object now has "type" property:
- default type is "THREE.LineContinuous" which should behave as before: geometry represents one continuous line (v0 ... vn)
- new option is "THREE.LinePieces" which tells renderer that geometry represents collection of individual line segments (v0,v1) ... (vn-1, vn)
(one Line object corresponds to one VBO)
(same limitations as with meshes - once VBO is baked, no changes are possible except object transforms - scale / rotation / position)
loader.loadAscii and loader.loadBinary now take just one parameter with following properties:
- model (required)
- callback (required)
- texture_path (optional: if not specified, textures will be assumed to be in the same folder as JS model file)
- bin_path (optional: if not specified, binary file will be assumed to be in the same folder as JS model file)
Example use:
loader.loadBinary( { model: "model.js", bin_path: "path/bin", texture_path: "path/img",
callback: function( geometry ) { createScene( geometry ) } } );
Code is still rough around edges (it can be made nicer), though it should already produce one-to-one results with ubershader.
Side-effect is that scene is no longer required as WebGLRenderer constructor parameter. Shader programs are now constructed upon first frame render and only then scene lights are examined.
Also changed MeshShaderMaterial demos to show how to use cloning. For these particular demos uniforms cloning is not really necessary as they use only one instance of shader material, but for example, if there were two normal mapped models in one scene, each model would need separate material with own uniforms.
For the moment, only one type of fog is baked into shader, depending on scene.fog initial value.
If use case arise for dynamic fog type switching, this could be changed, though than both fogs would need to be computed all the time :S.
Diffuse map and ambient occlusion maps are now optional in normal map shader.
By default, both are disabled, so they need to be enabled explicitly like this:
uniforms[ "enableAO" ].value = true;
uniforms[ "enableDiffuse" ].value = true;
This allows to work around ANGLE antialiasing issue appearing when compositing WebGL framebuffer with transparent background color with HTML canvas element background (while keeping antialiasing on).
WebGLRenderer constructor now takes JSON object with few optional parameters:
renderer = new THREE.WebGLRenderer { scene: scene,
antialias: true,
clearColor: 0x000000,
clearAlpha: 0 };
Here is how to change clear color in runtime:
renderer.setClearColor( 0xff0000, 1 );
Eventually I'll have to continue in the experimental branch. Want to give it a go to the ColorPass/TexturePass material API. The color + map + env_map setup feels too limiting...
Now that I have more knowledge of WebGL and shaders, this is how the ColorPass/TexturePass API looks like:
var material = [
new THREE.MeshLambertMaterial( [
new THREE.ColorPass( 0xff0000, 1, THREE.NormalBlending ),
new THREE.TexturePass( new Texture( <image>, new THREE.UVMapping() ), 0.5, THREE.NormalBlending ),
new THREE.TexturePass( new Texture( [ <image>, <image>, <image>, <image>, <image>, <image> ], new THREE.CubeRefractionMapping() ), 1, THREE.SubtractiveBlending ),
new THREE.ColorPass( 0x0000ff, 1, THREE.AdditiveBlending )
], { blending: THREE.NormalBlending } ),
new THREE.MeshBasicMaterial( [
new THREE.ColorPass( 0xff0000, 1 ),
], { blending: THREE.AdditiveBlending, wireframe: true } ),
new THREE.MeshPhongMaterial( [
new THREE.ColorPass( 0xff0000, 1 ),
], { ambient: 0x000000, specular: 0xbbaa99, shininess: 50, blending: THREE.NormalBlending } );
];
Makes things a bit longer but I think it's more intuitive this way (and pretty powerful). However, the simplest material would be like this:
var material = new THREE.MeshBasicMaterial( new THREE.ColorPass( 0xff0000 ) );
Mhh... Now I don't know whether they should called `*Pass`, or `*Layer`...
This is a workaround for Chrome ANGLE antialiasing issue manifested in geometry_terrain_fog demo:
http://twitpic.com/3iftyh
Problem happens in Chrome ANGLE when geometry is rendered with the same color as canvas background - then instead of expected nothing to be seen, there is a faint white outline at geometry borders.
This issue is not specific to fog, even just rendering geometry with MeshBasicMaterial having the same color as background produces the same outline.
TODO: create isolated test case and file ANGLE / Chromium bug.
For the moment, it works just on Basic / Lambert / Phong materials.
Fog must be added to the scene before initialization of WebGLRenderer and scene must be passed to WebGLRenderer constructor (like for lights).
I don't know yet how to solve properly MeshShaderMaterial & co :(
alteredq
Mr.doob
Alex Schworer