I am in the process of developing a game that consists of three different scenes, each with unique functions. However, all these scenes require the same set of three.js objects. My main question revolves around rendering speed - would it be more efficient to utilize three cameras and reposition the objects when switching between scenes? Or should I create three separate scenes, each containing similar types of objects but as distinct entities which do not need to be relocated? To put it simply, imagine having a scene filled with three.js objects shaped like letters spelling out a paragraph. Now, if I wish to display a new paragraph using the same letters, what method would result in optimal rendering speed: moving the existing letters around or transitioning to another pre-loaded scene featuring the objects arranged in the shape of the new paragraph? I am open to exploring alternative approaches to this task, provided they are solely based on JavaScript. Thank you for your assistance!