For high-fidelity wall-bounded turbulent flow simulations, it is essential to have an efficient inflow condition that introduces physically accurate turbulent fluctuations into the domain. Common methods for this purpose include numerical tripping, superimposing random disturbances on a given mean flow, utilizing precursor simulation data, the recycling-rescaling method, and using reduced-order models. Although several studies have been devoted to devising methods that reduce the development length to reach the expected turbulent behavior, the suitability of such methods for aeroacoustic simulations remains underexplored. Such an investigation requires a test case that is sensitive to both the turbulent characteristics of the incoming flow as well as the spurious inflow noise. In a flow over a forward-facing step, the dominant noise source regions are located behind and after the step where two recirculation bubbles are formed. The upstream recirculation bubble interacts with both the oncoming boundary layer and the downstream separation region. Therefore, the fluctuating dynamics of the recirculation bubbles and the resulting acoustic sources are sensitive to the characteristics of the turbulent boundary layer approaching the step. On the other hand, at low to moderate Reynolds numbers the spurious inflow noise can mask the noise generated by the flow around the step. Thus, an accurate aeroacoustic simulation for this case, requires an inflow condition that is sufficiently silent in addition to being accurate in producing the expected turbulent boundary layer. For this, we perform a Direct Numerical Simulation (DNS) of the turbulent flow over a forward-facing step at Re = 16000 based on the step height, using the spectral element method. We use various inflow conditions including a reduced-order-model-based inflow using vector autoregression, precursor-based method, the recycling-rescaling approach, and synthetic eddy method to investigate their effects on the turbulent flow field and the acoustic sources. Subsequently, we conduct acoustic simulations based on the DNS with various inflow conditions and compare the results to the experiment.