How can I use this tool to simulate near-field speech? For example, in an AI glasses array simulation scenario, I want to simulate the sound source located at the wearer's mouth position using this tool. However, after creating a simulation script based on the actual array configuration, I found that the simulated data does not appear realistic (I have already trained a model on a large amount of real-world recorded data, which I can use for validation). Could it be that near-field simulation itself inherently lacks realism?