Describe the bug Not a bug. I am just wondering what prevents building filament using the native Microsoft compiler in Visual Studio? Is it just a matter of C++ code not building? I saw a reference about assembly code in another issue. Is this still a problem or is there equivalent c code?

This is an initial draft of a C wrapper for #139.

This will likely require major rework before it is merged, but I wanted to get the discussion going nevertheless.

Current the idea of API.h / cfilament.h is that the same symbol name is used in the C header, but is directly aliased to the C++ class if the header is included from the cpp file.

I am also not sure whether duplication of documentation makes sense here. The interface is most likely going to be used by writers of FFI wrappers for other languages that do not use the header directly. Actual users of C will likely have access to an actual C++ compiler and can link against the static library directly without use of this wrapper.

Describe the bug I use filament via the SceneForm SDK. But I suspect it is a filament bug. When I put this material on an object then on some devices (I have not yet been able to determine which ones are 100% functional and which are not.) the objects behind it are displayed in black.

This effect only affects GLTF and OBJ objects. FBX models are correctly occluded.

The question now is whether it is a filament bug or rather an ArCore bug.

This is my occluder material:

material {
    name : "Occlusion material",
    shadingModel : unlit,
    colorWrite : false,
    depthWrite : true
}

fragment {
    void material(inout MaterialInputs material) {
        prepareMaterial(material);
        material.baseColor = vec4(0.0);
    }
}

To Reproduce Simply place a model with this material in front of a GLTF object. That is the result.

Expected behavior It would be correct if the object behind it were not visible. Like this:

Screenshots See "ToReproduce" and "Expected behavior" sections

Smartphone (please complete the following information):

• Device: [Not Working: Pixel 3a / OnePlus 5T] [Working: for example Samsung S7]
• OS: [All Android Versions]

Additional context The render order of the objects was set as follows: Camera -> Occlusion -> Others

If this is not a filament error I apologize. But I have the feeling that the people in the SceneForm Repo are being let down.

Maybe someone from this team can help. What surprises me is that it only does not work with GLTF and OBJ objects.

I am using readPixels with a callback on iOS/Metal. The PixelBuffer uses PixelDataFormat::RGBA and PixelDataType::UBYTE. I am reserving an uint8_t array of size width * height * 4 for the storage, and call renderer->readPixels(0, 0, width, height, std::move(pb)).

The swapchain has SwapChain::CONFIG_TRANSPARENT | SwapChain::CONFIG_READABLE as parameters, and I have metalLayer.opaque = NO to match that.

Unfortunately, when the callback fires, the buffer is all zeros. Any ideas?

(will try on other backends tomorrow).

I want to render the gltf model with a transparent background in android-

the code is

  ` surfaceView = SurfaceView(this);
    choreographer = Choreographer.getInstance()
    modelViewer = ModelViewer(surfaceView)
    surfaceView.setOnTouchListener(modelViewer)
    loadGltf("BusterDrone")
    loadEnvironment("venetian_crossroads_2k")

}

private fun loadEnvironment(ibl: String) {
    // Create the indirect light source and add it to the scene.
    var buffer = readAsset("envs/$ibl/${ibl}_ibl.ktx")
    KtxLoader.createIndirectLight(modelViewer.engine, buffer).apply {
        intensity = 100_000f
        modelViewer.scene.indirectLight = this

    }

    // Create the sky box and add it to the scene.
    buffer = readAsset("envs/$ibl/${ibl}_skybox.ktx")
    KtxLoader.createSkybox(modelViewer.engine, buffer).apply {
        //modelViewer.view.blendMode=View.BlendMode.TRANSLUCENT
        modelViewer.scene.skybox= Skybox.Builder().intensity(0f).color(0.0f, 0.0f, 0.0f, 0.0f).build(modelViewer.engine)
        //modelViewer.scene.skybox=null
    }
}

private fun loadGltf(name: String) {
    val buffer = readAsset("models/${name}.gltf")
    modelViewer.loadModelGltf(buffer) { uri -> readAsset("models/$uri") }
    modelViewer.transformToUnitCube()
}

private fun readAsset(assetName: String): ByteBuffer {
    val input = assets.open(assetName)
    val bytes = ByteArray(input.available())
    input.read(bytes)
    return ByteBuffer.wrap(bytes)
}`

I'm getting the black background when I'm using this code in skybox modelViewer.scene.skybox= Skybox.Builder().intensity(0f).color(0.0f, 0.0f, 0.0f, 0.0f).build(modelViewer.engine)

and twhen I'm using this code in skybox. modelViewer.view.blendMode=View.BlendMode.TRANSLUCENT modelViewer.scene.skybox=null

I'm getting like this

Please tell me how can I show the model without any background?

Describe the bug in windows, command "gltf_viewer my.gltf", that point lighting effect is visible; however, command "gltf_viewer -s my.gltf",that point lighting effect is gone. However, we do not want to scale the scene.

Expected behavior how could be modify the code to control this?

Describe the bug some devices are works, but some not.

crash info bugreport files

crash detail info refer to file tombstone_00

device info https://www.tme.net/device/vivo-x20a/

Android 8.1.0

Describe the bug I updated from 1.15.2 to 1.16.1 and the lighting of my skin meshes completely broke. It looks as if the normals aren't computed properly, I'm not even sure the light is coming from the correct direction. No problem on static geometries.

• I already rebuilt the materials.
• I don't think I have any kind of scales in my hierarchy

To Reproduce Steps to reproduce the behavior: I don't have repro steps and can't provide screenshots yet, this will take me a while to provide (or is there a way to provide screenshots privately?)

I'm opening the ticket as a heads-up and welcoming suggestions if this problem could be on my side. In the meanwhile I've temporarily reverted to 1.15.2.

Desktop (please complete the following information):

• OS: Windows
• GPU: NVIDIA GTX 1080
• Backend: OpenGL

I create a View and set a RenderTarget which contains a color and depth attachment.

The texture for the color attachment is created like:

fila_texture = filament::Texture::Builder()
      .width(settings.width)
      .height(settings.height)
      .levels(1)
      .usage(filament::Texture::Usage::COLOR_ATTACHMENT | filament::Texture::Usage::SAMPLEABLE)
      .format(filament::Texture::InternalFormat::RGBA16F)
      .build(settings.ctx->getEngineRef());

I've set the clear color using ClearOptions of the Renderer to an arbitary color with an alpha of 0.

Is this the correct way of rendering into a transparent buffer?

I'm asking because the result I get seems to have an alpha of 1.0.

Bug description We suspect that the Filament 1.4.5 version, present in last Sceneform version (1.16.0), does not deal (properly) with UV Scaling in GLTF/GLB files.

To Reproduce This kind of transform is carried-out by means of the GLTF 2.0 extension KHR_texture_transform. Said this, when this kind of UV Scaling is used in a GLTF file under ARCore, in order to obtain a repeatable large texture, these are the wrong results obtained: Please note the wrong repetition problem in the platform and the rail.

Expected behavior Of course, the texture is shown in a correct way when we visualize the GLTF by means of a WebGL viewer (e.g. VSCodium GLTF Extension with BJS rendering):

Smartphone (please complete the following information):

• Device: Motorola MotoG6 and Samsung Tab S4
• OS: Android 9.0

I am trying to build the filament libs into my own project, but my project already includes opengl for other libs and I am getting the following errors: https://pastebin.com/aNCA9DEd

Do you guys have some sort of workaround for this just like the unwindows.h file?

Move the initialization of VulkanContext fields to the class itself. This will clean up how the various extension support flags are deteremined.

Also disable debugUtils extension for Mesa driver (#6192).

Describe the bug When we try to load a GLB model frequently, After 3-4 iterations, it renders a skeleton or a skin-less model for some microseconds. We are unable to eliminate it. I've attached a video and a photo of the same

To Reproduce Steps to reproduce the behavior:

1. Load a GLB model
2. Again load the GLB model
3. It'll cause flickering ( skin-less,texture-less model to render)

Expected behavior It was expected that the Model will load with texture properly every time.

Screenshots

Smartphone (please complete the following information):

• Device: [Realme 9 pro, Poco x3 pro, many devices]
• OS: [ Android 12.0, 13]

Additional Details Glb Size: 8-10MB

@romainguy and team

As the picture showing, Creating Skybox or IndirectLight by KTXLoader.createXXX will create cubemap Texture. Normally, we may use Engine.destroy to release the Skybox or IndirectLight, but the cubemap Texture will not be destroyed unless the engine is destroyed or destroying the cubemap Texture explicitly(it is very oddly because nobody can realize it).

version: 1.28.1 platform: Android crash info:

Build fingerprint: OPPO/PFVM10/OP522D:12/SP1A.210812.016/R.202211071000:user/release-keys
Revision: 0
ABI: arm64
time: 2022-12-26 11:54:47
pid: 6127, tid: 6127, name: main  >>> com.re <<<
signal: 11 (SIGSEGV), code: 2 (SEGV_ACCERR) fault addr: 0x7637e0200c
si_errno:0, si_errnoMsg:Success, sending pid:0, sending uid:0
    r0: 0xb40000774abfe000  r1: 0x00000000000000ba  r2: 0xb4000077b21fa000
    r3: 0x0000000000000073  r4: 0x0000000000000000  r5: 0xb4000077b2586740
    r6: 0xb4000077b2585d0c  r7: 0x0000000000000000  r8: 0x0000000000002cc0
    r9: 0x0000007637e01040  r10: 0x0000007637e0201c  r11: 0xb4000077b21fafe0
    r12: 0x0000000000008000  r13: 0x0000000000000002  r14: 0x0000000000000000
    r15: 0x0000000000003c00  r16: 0x0000000000008001  r17: 0x0000007637ed2300
    r18: 0x6adacb2a790c2a3a  r19: 0xb4000077b2520e80  r20: 0x00000000000000b3
    r21: 0xb4000077b2558000  r22: 0x00000000000000b3  r23: 0xb40000785c8115b0
    r24: 0xb4000077b251c860  r25: 0xb4000077b654f108  r26: 0x0000007798853f98
    r27: 0x0000007fc22f95f0  r28: 0x0000007fc22f9570  r29: 0x0000007fc22f95b0
    r30: 0x00000077989b989c  sp: 0x0000007fc22f9440  pc: 0x0000007637f07744
    pstate: 0x0000000020001000

#00 pc 0000000000104744 filament::FSkinningBuffer::setBones(filament::FEngine&, filament::backend::Handle<filament::backend::HwBufferObject>, filament::math::details::TMat44<float> const*, unsigned long, unsigned long) + 180 [arm64-v8a]
#01 pc 00000000001a6898 updateBoneMatrices (/Users/re/work/other/filament/libs/gltfio/src/Animator.cpp:562) [arm64-v8a]

java:
com.google.android.filament.gltfio.Animator.void updateBoneMatrices()(Animator.java:66)
com.re.gltfplay.impl.GltfView$FrameCallback.void doFrame(long)(GltfView.java:718)
android.view.Choreographer$CallbackRecord.run(Choreographer.java:1216)
android.view.Choreographer.doCallbacks(Choreographer.java:972)
android.view.ChoreographerExtImpl.checkScrollOptSceneEnable(ChoreographerExtImpl.java:319)
android.view.Choreographer.doFrame(Choreographer.java:859)

I am trying to rotate the final rendered image by pi/2 or -pi/2 by adjusting the projection matrix. That works fine, except that light produced by spotlights changes. This is what I do:

math::mat4 rot;
double angle = 3.1415/2;
rot[0][0]= std::cos(angle);
rot[0][1]=-std::sin(angle);
rot[1][0]= std::sin(angle);
rot[1][1]= std::cos(angle);
rot[2][2]=1;
rot[3][3]=1;

auto projmat = camera->getProjectionMatrix();
camera->setCustomProjection(rot * projmat, camera_near, camera_far);

Directional lights behave fine (i.e. their effect looks the same with or without the rotation), it's just the spots that change.

Am I doing this wrong?