Box plots of ratings, averaged over the three presentations and the three loudspeaker positions, as a function of the BRIR set, for all attributes. A rating of 1 corresponds to “bad”, 9 to “excellent”. Statistically significant differences (post hoc pairwise tests for equal 20% trimmed means based on the bootstrap-t-method, p < 0.05) are indicated by lines in the bottom part of the facets.