It would be interesting to try such experiments with a coaxial speaker such as those from KEF and Genelec. I suspect, once ones preconceptions have been removed, the "height" effect will be, too. With a two-way system, due to there needing to be physically offset drivers with a cross-over, some component of the audio spectrum that we normally use for sound localization will be carried by the midrange/woofer as well, and this can induce ILDs that we could possibly interpret as cues with respect to height. Same goes with room reflections as they can emerge from differing trajectories on their way to the listening position. Or even comb-filtering, as others have pointed out, could mimic the spectral coloration induced by the ears that we use to determine height in the first place. The effect of the speaker topology and room have a powerful impact of how we perceive things. From a review of the
FR30, the reviewer noted that it had a very deep soundstage, which is likely due to the inclusion of the super tweeter on the back of the speaker. This will add a sense of depth due to there being sound coming from both the front baffle as well as reflecting off of the back of the room. Its important to note that this will be artificial, even though its likely to be subjectively pleasing from a listening standpoint.
I'm not saying height is impossible in a two-channel recording. But with conventional microphones there is no mechanism in place to convey it. It would genuinely be interesting to repeat his experiment on height with proper controls, but I suspect the results would be to dispel the existence of height in conventional stereo recordings once biases have been removed. I think we know enough about psychoacoustics to have a reasonable degree of confidence at this point to draw that conclusion. But, I am also genuinely open to being proven wrong since that is where all your interesting discoveries come from
On a side note, since I have the equipment to do so, I repeated one of his experiments, namely
Temporal resolution by bandwidth restriction and my results were not good. They showed that Kunchur had neglected, at a minimum, two things. The first being having a control to keep the fundamental energy in all the waveforms the subjects were subjected to constant, and the second to actually do high resolution FFT of the actual output from the apparatus to ensure there are no IMD components clustered around the harmonics of the square wave since these alone can give unintended cues to the subjects. I know my generator does the same thing, and I suspected I could hear a difference just from those alone. But the upshot was once all those aspects were removed and I repeated the experiment, its clear he was getting a false positive. Once the harmonics have been pushed out of the audio spectrum leaving only the fundamental,
any waveform chosen simply sounds the same as a sine wave at the same frequency. Its simple physics at that point. Its unfortunate that he falls prey to such things, but I think we can say that any paper he has written on audio is, at best, junk science.