Methods, apparatuses, and computer program products are provided in order to provide 3D audio playback using audio head-mounted devices. The apparatuses may be configured to receive at least one of position and orientation of a first head-mounted device in relation to a first user device, wherein the at least one of the position and orientation received is used to train a model using machine learning. At least one signal quality parameter may be determined based on input data and a filter pair may be determined corresponding with a direction to which a spatial audio signal is rendered based at least in part on the at least one signal quality parameter and the model so as to control spatial audio signal reproduction to take effect a change in the at least one of the position and orientation of the first head-mounted device during rendering of the spatial audio signal.