Our latest paper evaluates the numerical accuracy of simulated HRTF magnitudes with increased human head/torso model complexity, as well as their accuracy in terms of sound source localization response predictions: