Interesting, but I am not sure what will change in @yaustar’s demo:
In camera-to-texture.js, I’m not even sure how these lines are working and how they need to change. The variable layer is not used elsewhere, but still, changing the renderTarget of that object change the behavior.
var layer = this.app.scene.layers.getLayerByName(this.layerName);
layer.renderTarget = this.renderTarget;
The renderTarget is basically the place where the calculated pixels should be drawn. The original approach used a layer to mark all included objects, which is a bit more cumbersome. With the camera renderTarget everything seen by this camera is drawn to the renderTarget (screenshot texture).