Beckwith's (2003) environmental sound identification task was re-administered to check the reliability of previous results. The participants were presented with 30 different environmental sounds from five acoustic categories (see Appendices G and H for details). They were required to identify each sound after ten seconds. Overall, MR produced significantly more errors than controls (z = 3.48, p < 0.001), suggesting his auditory agnosia is not specific to musical sounds but extends to environmental sounds. As shown in Table 2.6, most of his errors were for technical sounds or naturally occurring non-living sounds. A common feature that may be indicative of the root cause of MR's environmental auditory agnosia is that these are all rapidly changing broad spectrum sounds.
MR's ability to better control his musical illusions during re-administration of this task reduced his error rate for the human non-verbal, and naturally occurring non-living sounds in comparison with his previous results (Beckwith, 2003). He did, however, report a faint violin sound throughout the duration of the items, pointing to the presence of some form of illusion, albeit unobtrusive.