I never mentioned Real Time or even Operating System.
As soon as you put a micro/cpu some form of software has to be there, USB requires several stages of operations and scheduling of the USB traffic, even with one device.
Devices still have to be enumerated, and endpoints handled.
Cameras normally have two types of flow, control and data. Where the data is isochronous mode bulk transfer due to its nature. These are often handled as different endpoints or in some cases a composite device with effectively two devices.