I'm intending to create a view/camera where the entire scene is situated in front of you, and you can rotate the scene on the x,y and z stationary axises. For example imagine you have a floating ball in front you, which with your hand you yaw it on the y axis(up) by spinning it left or right, or you pitch it on the x axis(right), by spinning it away or towards you. The viewer's axises never change and the ball always rotates via these axises. I hope I've made that clear, my question is on the method of implementing this. Would it be best to rotate the world matrix via the stationary axises and have the camera stationary, and then multiply all the scene matrices to the word matrix before transforming them? or should i spin the camera's position, up, right around the scene? Which seems a little more difficult but better because it lessens the amount of multiplying matrices stated above. Also with this should I stick to matrices or would it increase the smoothness if I use a quaternion? thx