An electronic apparatus and method for audio reproduction virtualization in an enclosed physical space. The electronic apparatus generates a 3D model of an enclosed physical space based on one or more images and applies a machine learning model on the 3D model, to determine a plurality of objects and one or more acoustic parameters of the plurality of objects. The electronic apparatus controls an audio capturing device to determine a first frequency response for a first sound in the enclosed physical space based on the determined one or more acoustic parameters. The electronic apparatus applies a plurality of combinations of operational modes, and 3D positions of the first audio reproduction device on the 3D model and the first frequency response to generate a second frequency response. The electronic apparatus selects first combination based on the second frequency response and displays output information associated with the selected first combination.