I am using SceneKit to show a 3D model on iOS. I would like to use PBR. At the moment I am using the scene editor not coding to edit my materials. I can set any field to PNGs at the models material (metalness, normals, etc..) except the roughness. When I set it, the whole model turns pink. If I use float value, there is no problem. Do you have any idea, what the problem could be?
Related
I've replaced the tree.glb model in the ThreeJS placeground example (https://github.com/8thwall/web/tree/master/examples/threejs/placeground), but it's not showing. It works fine when using tree.glb.
To debug, I've also tried replacing it with the jellyfish-model.glb available in the examples, but it also doesn't show when tapping on the floor plane.
Is there something wrong with my code, or with the .glb models I'm replacing tree.gbl with?
const modelFile = 'tree.glb' // 3D model to spawn at tap
to
const modelFile = 'jellyfish-model.glb' // 3D model to spawn at tap
File structure on github: 8thwall-3js-test-github
Ideally, I'd like to replicate what I've done using Unity+Vuforia in this example (which basically places a .png onto a floor plane): https://www.youtube.com/watch?v=poWvXVB4044
I'd start by looking at the scale of the 3d model. The tree model in the link you provided is quite large, so it's being scaled down in size. See https://github.com/8thwall/web/blob/master/examples/threejs/placeground/index.js#L7-L8
Prove to yourself that the model is being loaded by adding a console.log('model added!') type statement into animateIn() (as that is the model loaded handler)
My guess is that your jellyfish-model.glb is there, just very small. Trying adjusting startScale and endScale to larger values and see if that helps.
I have developed iOS application to have AR functionality using ARKit. I have used this project.
The application works with Collada (.dae) file dynamically which means the the client uploads the Collada file and all of the textures somewhere and the model gets shown accordingly with the help of this solution.
The application works mostly fine but for this one we experience the black part of the model flashing.
As you could see in the pictures most of the floor should be completely black(apart from one small part which is black and white). When we zoom in the model (make the model bigger) the model gets shown with less white spots on the floor(in the first picture almost no white spot but sometimes we can see some). When we zoom out and make the model smaller, we see many white spots on the floor and other places which should be black.[zoomed in the model with less white spot]
Is there any solution for this problem?
The issue you are seeing is known as Z-fighting.
In the given model, the black (bottom/floor) plane and the black/white patterned plane are very close so as the user moves further away the renderer doesn't accurately differentiate which plane should be shown first.
One solution is to set the readsFromDepthBuffer property of the plane's materials to false.
Once you have a reference to the node/plane, you can set the property to its material.
if let geom = node.geometry {
for material in geom.materials {
material.readsFromDepthBuffer = false
}
}
Having problem with SceneKit object. Trying to get it to look like the first image below, but in the application it looks like plastic bronze.
Object in Xcode / SceneKit
Object in AR app
Can't really figure out what am doing wrong but have narrow it down to my material settings. Current settings:
Materials settings in Xcode
You're assigning specular map for diffuse, specular and metalness which is wrong.
PBR lights have four important components which are
Diffuse, Roughness, Metal and Normal
Assign these properties with the right map and you will have the expected result.
For More details on PBR check this link
Is there a way to change the appearance (size, color, etc) of the feature points in ARKit easily? (After setting debugOptions in the sceneView to ARSCNDebugOptions.showFeaturePoints I'm thinking I might have to iterate over the rawFeaturePoints and manually add custom objects into the scene at those points.
As its name suggests, ARSCNDebugOptions.showFeaturePoints is a tool to aid in debugging your app. Because the size and color of feature point indicators aren't essential to knowing where feature points are (for the sake of making sure your app is behavior correctly), Apple doesn't offer API to change their appearance. (Any more than they offer APIs for changing the colors of bounding boxes, physics shapes, and other indicators available in SceneKit debug options.)
If you want to create your own visualization for feature points, you'll need to do exactly as you suggest: read the rawFeaturePoints from the current ARFrame and use those to position content in the SceneKit scene. You might do this by creating a bunch of nodes with geometry and setting their positions. You might also look into whether it's easy to pass the entire buffer of points to create an SCNGeometry that renders in point-cloud mode.
I am attempting to experiment with Apple’s Fox game SceneKit example (link below) by adding a model with a simple animation like the ‘panda.scn’ and ‘walk.scn’ assets.
I can create a static model with no joints or animation that works: e.g. In Maya (2017) I add a simple sphere, export selection to FBX_DAE (COLLADA) file, drag it into the project in XCode and convert it to a SCN file. I can then drag that model into the ‘level.scn’, position and scale it as I’d expect.
However as soon as I add any animation or joints to my model I lose the ability to position and scale the model in XCode.
In Maya I add two joints to my sphere select the sphere and joints and export as above. When I examine the model SCN in XCode (either in isolation or as a reference within another scene) I find that I cannot apply any translation or scaling. XCode lets me move the xyz locators in the GUI and update scale but the model does not change. I can see the mesh and joints in the outline view and I have tried moving the joints instead of the mesh, but they do nothing. I have even tried ignoring the GUI and positioning the model in code just as the panda character is set up, but applying positions or transforms to the node does nothing - it always appears at the origin and default scale.
Ignoring that for the moment, my understanding from looking at the ‘walk.scn’ file is that an animation is just an export of the joints with keyframes. I have tried to reproduce that by exporting only the joints to a separate DAE and importing it, then applying that in code as they do with the ‘walkAnimation’. This seems to do nothing as well.
I have experimented with various settings of the FBX_DAE export dialog including baking animations (By the way - what is the difference between baking an animation on export vs baking it in Maya before export? The former seems to do something and the other does nothing in my tests.)
I would dearly love a workflow for creating some simple character animations in Maya and getting them into SceneKit. Any help is greatly appreciated.
For reference:
Apple’s source: https://developer.apple.com/library/content/samplecode/Fox/Introduction/Intro.html
Maya scene for my trivial ball model with two joints:
http://pat.net/misc/ball1.mb
and the DAE output from Maya for that:
http://pat.net/misc/ball1.dae
UPDATE:
Mnuages's answer appears to be correct in that when I added a parent control over the geometry and joints SceneKit then allowed me to move them. But even after brushing up on my understanding of how these nodes and their transforms relate in Maya I do not feel that I have a real understanding of how SceneKit is interpreting them. (Would love to read some docs on that would illuminate this more if they exist).
UPDATE #2:
I was finally able to create an animation by doing the following: 1) Export either the full scene or just the joint with the animation being sure to select the "bake animation" option in the DAE export dialog or bake the entire animation using Key->Bake Animation. 2) It only works if I load the DAE file in scenekit instead of converting it to an SCN. Converting to SCN format seems to lose the animation.
what happens is that the node named joint1 is animated by the joint1-anim animation. So even if you move joint1 in the editor, what you see on screen is the result after the animation is evaluated.
If you create an intermediate node, say joint1-parent, and make joint1 as child node of joint1-parent, then you'll be able to translate and rotate joint1-parent freely and see the effects on joint1.
As for why moving pSphere does not change anything, it's the same idea. Just like the animation overrides the position, the skeleton will reposition the mesh.