Soundscape-to-panorama: spatialize auditory perception by linking acoustic environment to panorama

In order to gain a deeper understanding of the mechanisms that shape the human sense of place, more and more geographical studies are beginning to combine auditory and visual perception. However, majority of them combine acoustic environment with limited-view landscape images, ignoring the impact of space on the formation of acoustic environment and auditory perception. In fact, as scene sounds are multidirectional and propagate in 3D space, combining acoustic environment with visually restricted images will omit spatial and visual soundscape information. We developed a Soundscape-to-Panorama model that generates landscape panoramas from audio data to solve this problem. This model offers richer visual cues compared to the Soundscape-to-Image model. It enables a deeper exploration of spatial and place information within scene audio and demonstrates excellent performance in both task-specific and general evaluations. From the perspective of geographical research, our study is the first to integrate panorama generation with acoustic environment in the understanding of place perception. We have broken through the limitations of previous research, observed the importance of space in acoustic environment research, and opened up a new path for future acoustic environment research.