Fused camera sensor performance?

[gedalia]

Hello I've got a system that needs to capture the output of many camera sensors and I was hoping to get rgb/depth/segmentation all at the same time, in my experience with unreal that equals 3 capture objects and 3 additional rendering passes, making it pretty hard to hit frame rate targets. It seems like carla is using separate sensor captures for those kinds of tasks too, I imagine that each one is requiring a separate camera render pass. I'm wondering if anyone has looked at modifying unreal's capture code to support multiple output textures since different sensor results could just be different frame buffer attachments.