What do You think, how would a listener disciminate different frequency responses of a speaker using an unknown program?
I think, that the often cited "Blauert'schen Bander" are mostly an artifact of the original investigation of Blauert, and shall not relate to stereophonics without a bullet proof explanation.
Consider a musical instrument e/g castanetes (snappers) of one brand versus another. One would emphasize a Blauert band the other would not. Does it mean, that a human is supposed to localize one brand, say coming more from 'above' than the other? I doubt that.
I argue that Blauert's results just reflect the head related transfer function when evaluated with the very same program material. The prerequisite of same program is not given with stereo, when used at home as inteded.