I have a glb model which I am loading into my Vue project via Three js. I have managed to import several other models for practice, but the one I actually want on my page will not load. I have tried playing around with different scaling, positions, background colors (since the object is mostly black), and camera angles but I am not able to get it in frame no matter what I do. I am able to see the model perfectly in any regular gltf view, but I cannot see it in my project, what am I doing wrong here?
Edit: As a side note, I also tried changing the scale in blender, but that did not change the result.
const loader = new GLTFLoader();
let me = this; // must refer to the instance in vue in order to be added to the scene, have tested this with other models
loader.load(
'pantalla_ball_2.glb',
function(gltf) {
gltf.scene.traverse(function( node ) {
if ( node.isMesh ) { node.castShadow = true; }
});
gltf.scene.scale.set(1,1,1) // Have tried several scales from 0.01 to 200
me.scene.add(gltf.scene);
console.log("added") // added is successfully called every time
},
function(xhr) {
console.log(xhr);
},
function(err) {
console.log(err); // no errors appear in console
}
);
The object loading perfectly in 3d viewer below
Thanks to JP4 for his helpful comments, the issue was the ambient lighting was not strong enough to illuminate the object sufficiently enough to be seen. The axis helpers showed me that the object was there thanks to the gaps in the axes.
Related
I am trying to load quite complex .3ds model into three.js using TDSLoader exactly the same way as in three.js/examples: https://github.com/mrdoob/three.js/blob/master/examples/webgl_loader_3ds.html
but what I get looks like this Result when it is supposed to look like this Loaded using https://3dviewer.net/. I can see that this online viewer uses three.js so it is possible. I also know that TDSLoader loads textures but does not map them for some reason (if I move texture out of the folder I get an load resource error). Code looks like this:
var loader = new THREE.TextureLoader();
var normal = loader.load('textures/normal.png');
var loader = new THREE.TDSLoader();
loader.setPath('textures/');
loader.load('textures/CAT_336D.3ds', function (object) {
object.traverse(function (child) {
if (child instanceof THREE.Mesh) {
child.material.normalMap = normal;
}
});
scene.add(object);
});
My model has 23 types of materials but the loader loads only 2 - black and yellow, you can see that glass, metal and others are missing. How do I map all textures and materials? Maybe it does not work because the model is too big? (over 160 meshes, around 30MB)
I am trying to display a textured plane with Three.js. I'm working with Forge RCDB.
At first, I managed to display the plane, but instead of being textured, it was completely black... I made some changes and now nothing is displayed anymore...
Here is my code :
render () {
var viewer=NOP_VIEWER;
var scene=viewer.impl.scene;
var camera = viewer.autocamCamera;
var renderer = viewer.impl.renderer();
renderer.render( scene, camera );
}
and in the function supposed to display the textured plane :
new THREE.TextureLoader(texture).load(texture, this.render);
tex.wrapS = THREE.RepeatWrapping //ClampToEdgeWrapping //MirroredRepeatWrapping
tex.wrapT = THREE.RepeatWrapping //ClampToEdgeWrapping //MirroredRepeatWrapping
tex.mapping = THREE.UVMapping
At the beginning I used loadTexture(). I managed to display my plane, but it was all black, and no texture was applied on it.
Then, I use THREE.TextureLoader().load(), in this case, I believe it is trying to find the image on localhost. The image is downloaded, I can see it on the console.
But now I get these errors :
Uncaught TypeError: scope.manager.itemStart is not a function
and :
Uncaught TypeError: renderer.render is not a function
Now the object is not displayed, even in black.
So I think this may be linked to render, but I don't understand how...
I found this, and it answers my question partially.
Finally, I decided to keep THREE.ImageUtils.loadTexture(), and I replaced MeshLambertMaterial by MeshBasicMaterial.
No need for render.
I know this is a well known question but I was not able to find a good answer.
My usage
I use three.js to display 3D models created through drone pictures (here an example).
The problem
I can't render heavy models (1M vertices, 2M faces) : Chrome or WebGl crashed.
What I tried
I used Threejs.org examples for all my tests to be sure that it was not my code that didn't work. I made my test on a x64 Chrome with --max_old_space_size=6144 flag.
Import model in .dae with ColladaLoader --> 2.5Go of RAM is used and Chrome crash (displaying "not enough memory")
Import model in .obj with OBJLoader + MTLLoader --> 2.8Go of RAM is used and WebGl crash
I red many posts about Three.js and memory allocation but many of them speak about remove an object from the scene
Possible solutions
I saw that .stl (binary) file are more compact, but as far as I know
there is no texture with these files, so I can't use it
Use BinaryLoader (which has GeometryBuffer output) but I need to convert .dae or .obj to binary files and I don't know how to do that
Load my model in multiple parts to not load in one shot ? But I didn't saw any example or posts with that kind of treatment
How to reproduce
For the code, I use basics examples on Threejs.org. For models :
If you want to try with .dae you can find on this folder a working example (WorkingModel.dae / .jpg) and the model who don't work (BigModel.dae / .jpg)
If you want to try with .obj you can find on this folder a working example (WorkingModel.dae / .jpg / .mtl) and the model who don't work (BigModel.dae / .jpg / .mtl)
Any ideas to load the big one ?
Thanks !
EDIT 1 :
I tried to put a breakpoint in the SuccessCallback to see if the overload of RAM is during the load or after. I was not able to hit the breakpoint, so the overload of RAM is before the SuccessCallback.
Then I went step by step in the ColladaLoader to find what is using so much RAM. Here is the "callstack" :
myCollada.load()
ColladaLoader.parse()
Geometry.parse()
Mesh.parse()
Source.parse (hit 3 times) = +400mo in RAM
Vertices.parse = +0mo in RAM
Triangles.parse = +1500mo in RAM
this.geometry3js.computeVertexNormals() = RAM go over 2600Mo and chrome crash
Could I do any other tests to find the reason of this problem ?
Thanks
Your textures are waaaay to big, and besides, you don't need them because your model has baked vertex colors, which includes baked lighting. Your model therefore does not require UVs eiher.
Use ColladaLoader2, and this pattern. It should work.
var loader = new THREE.ColladaLoader();
loader.load( 'BigModel.dae', function ( collada ) {
var dae = collada.scene;
dae.traverse( function( child ) {
if ( child instanceof THREE.Mesh ) {
child.geometry.removeAttribute( 'uv' ); // you don't need it
child.material = new THREE.MeshBasicMaterial( { // scene lights not required
vertexColors: THREE.VertexColors // you have them, use them
} );
scene.add( child );
}
} );
} );
three.js r.86
The following has worked well for me with svg files in R76 and R77 of Three.js, but in R78 I can only get it to work with pngs and jpgs
var floor = new THREE.TextureLoader();
floor.load('layout.svg', function ( texture ) {
var geometry = new THREE.PlaneBufferGeometry(4096, 4096);
var material = new THREE.MeshBasicMaterial( { map: texture } );
var mesh = new THREE.Mesh(geometry, material);
mesh.rotation.x = -Math.PI / 2;
scene.add(mesh);
} );
Since the problem arose, I added the progress and error functions to the load arguments...
function ( xhr ) {
console.log( (xhr.loaded / xhr.total * 100) + '% loaded' );
},
function ( xhr ) {
console.log( 'An error happened' );
}
...which tells me that in r78, the svg is 100% loaded, but it doesn't show up in the scene.
I should add that I also used this with phongmaterial for the reflection properties, and with transparency to show svg elements apparently floating in a skymap, great fun.
My question is, how can I get this working?
In R77, the svg would map to a planeBufferGeometry in FF or chrome.
In R78 it wont load, no errors shown on Firebug.
UPDATE
I investigated SVGLoader.js, SVGRenderer.js (which also requires Projector.js). Renders an svg file facing camera, responds to translations, but not rotations; i.e. no perspective view possible.
I added an xml header to the svg file, then a doctype, but threejs obviously isn't bothered by their absence. Using localhost makes no difference in FF, required in Chrome
So the problem seems to lie either with TextureLoader or WebGLRenderer, which have both had quite a reshuffle in R78.
I also briefly tried the dev version, which acts as r78 does.
Can anyone suggest where I go from here? I'm not sure whether to submit a bug, or a "feature restore".
Here's the desired effect.
Now my understanding is that prior to R78, svgs were internally converted to raster images before rendering.
To continue to use svg images in R78 onwards with Textureloader, convert them to pngs beforehand.
OR if the svg always faces the camera, use the THREE.SVGLoader.
OR If the separate svg elements are required in a three.js scene, convert them to paths and use this example to make a 2d shape or a 3d extruded shape.
My lesson here was not to rely on examples to just work. I should dig right in, break things, and stay up to date with issues on Github.
And use more than one browser; at the time, Chrome was going through a buggy phase not entirely unrelated to this.
In this project I am working on I have several Collada models being displayed and then removed when not needed anymore. There seems to be a memory leak somewhere in the project and I am looking for ways to get it to run as smooth as possible as time is not on my sideā¦
I feel like I don't remove the meshes the right way and that this might cause some of the memory leakage I have.
I load the objects with LoadObject(level_1_character, "Assets/Level_1_Character.dae"); for example, where level_1_character is a Object3D. This calls the following function:
function LoadObject(name, url) {
var mesh, geometry, material, animation;
var loader = new THREE.ColladaLoader();
loader.options.convertUpAxis = true;
loader.load(url, function(col) {
mesh = col;
geometry = col.scene;
name.name = url.toString();
name.add(geometry);
});
}
I add the objects to the scene depending on the level through scene.add(level_1_character); and then remove it by doing the following:
level_1_character.traverse(function(child){
if (child instanceof THREE.Mesh) {
child.material.dispose();
child.geometry.dispose();
}
});
I am not sure if this actually fully removes the object though. It seems as though the objects still are present in memory. Any idea what I'm doing wrong here?