I am using MetalKit and have a complex rendering pipeline. The results are rendered into a MTKView.
Now I would like to feed the contents of the MTKView to a SCNScene and using a SCNCamera to perform post-process effects like HDR.
How is this possible?
I do not want general directions. I want specific calls if possible.
You should ideally perform your post-process as part of your Metal rendering pipeline. The procedure you are suggesting requires unnecessary resources since you will be rendering a 2D plane in 3D in SceneKit just to apply some HDR effects.
Nevertheless, you can achieve what you want by rendering your Metal pipeline output to a texture and then simply applying it to a plane in SceneKit.
First assign your texture:
plane.materials.first?.diffuse.contents = offscreenTexture
Then override your SceneKit rendering to your Metal rendering loop:
func renderer(_ renderer: SCNSceneRenderer, willRenderScene scene: SCNScene, atTime time: TimeInterval) {
doRender()
}
Then perform your Metal rendering with the texture as target, and once that is done you render your SceneKit scene:
func doRender() {
//rendering to a MTLTexture, so the viewport is the size of this texture
let viewport = CGRect(x: 0, y: 0, width: CGFloat(textureSizeX), height: CGFloat(textureSizeY))
//write to offscreenTexture, clear the texture before rendering using green, store the result
let renderPassDescriptor = MTLRenderPassDescriptor()
renderPassDescriptor.colorAttachments[0].texture = offscreenTexture
renderPassDescriptor.colorAttachments[0].loadAction = .clear
renderPassDescriptor.colorAttachments[0].clearColor = MTLClearColorMake(0, 1, 0, 1.0); //green
renderPassDescriptor.colorAttachments[0].storeAction = .store
let commandBuffer = commandQueue.makeCommandBuffer()
// reuse scene1 and the current point of view
renderer.scene = scene1
renderer.pointOfView = scnView1.pointOfView
renderer.render(atTime: 0, viewport: viewport, commandBuffer: commandBuffer, passDescriptor: renderPassDescriptor)
commandBuffer.commit()
}`
Full example project: