You should have your render texture setup as render target in a dedicated camera which looks at the Label. Then the label should be rendered to the target RT
- Your label’s layer should match the capturing camera’s visibility, ideally one dedicated layer for each capturing camera, this will avoid mismatch and over draw.
- Your target RT should be set to a material as main texture to be rendered, normally you will be using quad geometry with an unlit material.
Your issue is probably due to the first notice, in previous version the engine handles canvas camera visibility and its nodes’ layer automatically, during the render process, we assign a layer for each canvas as the camera’s visibility and its contents’ layer. In v3.0, as you noticed, camera is separated from canvas and we leave the control to user. There are two reasons
- We need camera to be free controlled for 2d tiled games and parallel scroll games
- We will support 3D UI soon, 2D elements like sprite, label should be able to be rendered outside canvas, so node’s layer must be considered.
Anyway, now the Camera visibility and Node layer is the only relation determines which camera renders a given node.