We study noise and recording properties of nonoriented barium-ferrite particulate media using micromagnetic modeling. The packing and orientation distribution of the barium ferrite particles are reproduced very well with our packing algorithm. The distribution of switching fields is determined from experimental hysteresis loops. Using these parameters, we perform recording simulations of periodic waveforms written at various frequencies and extract broad-band signal-to-noise ratio (BB-SNR). Comparison to experimental measurements shows very similar signal and noise characteristics, with however a 6 dB offset in BB-SNR values. Other (e.g., mechanical) noise sources would need to be included in the simulations to account for the difference. Nevertheless, the simulations prove very useful to understand and quantify how particle and media parameters contribute to the signal-to-noise ratio. For the present nonoriented particulate medium, we verify that the particle volume distribution affects the noise power according to the existing analytical expression for particulate noise. The particle anisotropy distribution is found to significantly affect the signal roll-off. The related variation in BB-SNR is found to be in good quantitative agreement with experiments. The effects of other parameters on BB-SNR, such as dipolar coupling, medium thickness, and average head-medium spacing, are also presented. An analytical slope model is proposed for nonoriented media that reproduces the simulations and experimental data very well.