Noise is getting increasing attention as environmental factor. Communication of potential impact to affected citizens is part of a participatory approach to many projects like e.g., in windenergy projects. Sound, however, is hard to grasp by non-specialists looking at commonly used metrics. Noise maps from sound simulations are often used to communicate expected noise levels from certain sources but only provide abstract insights that still can be hard to transfer for nonexperts. Extended reality (XR) in combination with real-time spatial audio allows to explore the impact of future projects in the built environment as a direct experience. While the generation of a visual environment poses no greater challenge by using geospatial data, realistic modelling of a spatial soundscape from multiple point-sound-sources is not trivial. This paper reflects on the challenges to present an audio-visual windfarm in XR, outlines established techniques that have proven useful as well as challenges and opportunities in creating, delivering, and evaluating virtual soundscapes.