A vibration dampening speaker mounting assembly includes a flush-mountable frame for mounting to a wall or ceiling, at least one vibration dampening member, and a baffle for supporting a single or multi-driver transducer array. The at least one vibration dampening member is disposed between the baffle and the frame to reduce acoustic distortions coupled to the wall/ceiling from the baffle. Fasteners are provided that permit installation of the baffle to the frame without tools while avoiding significant acoustic coupling between the baffle and the frame through the fasteners.