Binaural models’ output for HRTFs that were preprocessed with different methods and interpolated for spatial orders 1 to 44. Top left: left-ear perceptual spectral difference (PSD [56]) with the reference, averaged across the approximate 110-point Lebedev grid (lower is better). Top middle and top right: lateral and polar localisation precision, as estimated by the model of Reijniers et al. [26] for the same 110 directions (lower is better). Bottom left: externalisation, as estimated by the model of Baumgartner and Majdak [27] for 180 median plane directions (higher is better). Bottom right: spatial release from masking (SRM) in dB, as estimated by the model of Jelfs et al. [28], averaged for 180 masker positions in the horizontal plane. Reference data (black dotted line) was obtained from the original HRTF. Detailed data in tabular form is provided in [64].