I would recommend using Static pressure to track pressure oscillations. Depending on your application and what you are looking for, you can do this at the outlet (using a area-arveraged value) or by defining point receivers in your domain. I suspect your point-graph does not work because you specified area-average for a point (which has no dimensions).
Regarding your time step, it depends on your frequencies of interest for your Fourier analysis. All frequencies of interest must be spatially and temporally resolved from the source to the receiver location. The source zone will require sufficient mesh and temporal resolution to capture scale-resolving turbulent flow, while the region of propagation (transmission zone) needs to be resolved with at least 15-20 grid points per wavelength for the highest frequency of interest. Under-resolved waves will numerically dissipate through the transmission zone. Then, depending on your smallest grid size dx and the speed of sound c0 of your fluid, you should select a time step dt to achieve CFL of approximately 1 (CFL=c0*dt/dx).