Soundstage Width and Depth

Duke,

The source "appearing" outside the speakers is "encoded" in the music with either mixing methods or microphone techniques, but and it is a big but, how that appears on playback is highly dependent on speaker, listening position, microphone or mixing technique and very important the listener themselves. The microphone technique or mixing (playing with timing) can have more impact on perceived playback that anything to do with the venue.


Keep in mind that how this works is not by recreation of a real source outside the speaker as would be the case with a first reflection, but by tricking the brain with a delayed signal from the other speaker hitting the opposite ear to generate timing information that the brain may perceive as equivalent to the timing information of a sound wrapping around the head to determine direction. Not everyone interprets it the same and the effect can be positive or negative and is influenced it appears by individual interpretation of timing and volume cues for direction, not to mention what works for one recording on one system could fall apart completely with a different recording or system. 


For those discussing "height", we have a hard enough time getting height in a real 3d environment. Nothing that comes out of your speakers (in a single plane) is height beyond that your room acoustics may generate and it would bear little reality to the actual recording environment (if it had any height at all in the first place).

Yes, this it the most current knowledge and there are no indication it is incorrect, but even with these cues, it can be difficult to accurately assess height. I spent a number of years doing R&D on hearing aids and similar audio "devices". Our group believed we were one of the first to look at how the design of the hearing aid could be improved with the goal of preserving positional cues most take for granted. Unfortunately that R&D was abandoned after I left as well as other programs to pump up the balance sheet before selling. It was a bit contentious at the time as well. It indicated issues with signal processing delay differences masking timing cues.

"Technically",  just as you have indicated, frequency filters that mimic the pinna, can provide a sense of height in head-phonic playback and encoded in only two channels. There has been a fair amount of research done with HATS (head and torso simulators) for recording, but, as you indicated, it requires tailoring to the individual to work properly. If you attempt that technique with speakers, you get not only the HATS transform, plus the listener ... and two pinnas are not better than one.  W.R.T. your particular situation, making a wild ass guess, the microphone above his head, if not omni and not pointed at him, created a filtering effect that simulated height with pinna filtering. Curious if the wavefront from the electrostats is less impacted by torso/head/pinna than would normally occur with dynamic speakers.  Interesting!  I may have to pick up a pair now and do some testing.


Speaking of interesting, to the last post about difficulty of creating a stable image outside the speakers, have you done much research on ambiphonics?

Regarding height cues out in the "real world", my understanding is that the way sound diffracts around the head and outer ear (the pinna) from above is what gives us height cues. I have read papers and articles about encoding these "head and pinna transforms" into a signal to convey height information, but to really do it right, the equalizations would have to be tailored to the individual's ears. (One possible application would be in the helmets of fighter pilots, so that an audible threat warning would also convey the direction. Head position tracking would of course have to be included.)  

I don't see how height information could be encoded in a normal two-channel recording... BUT something weird happened to me years ago.

Ambiphonics:  I would start with this:   https://cdn.website.thryv.com/7b2b654758d449b08935c9dfa207e8f9/files/uploaded/Ambiophonics_Book.pdf

Then read this article on methods that are more robust:  https://www.microsoft.com/en-us/research/wp-content/uploads/2013/10/Ahrens2013a.pdf

While there is criticality of listener position, it is much more robust than ops "fluke" that requires perfect everything to "maybe" work.

Duke,
w.r.t. your sensation of height with that speaker set, been thinking about it more. I still feel there was the potential for frequency filtering (and possibly reflection/sheltering) due to the microphone pattern and position, and this could have simulated a human torso / ear / pinna. Being line source, there would be limited ceiling and floor reflections, so what got to your ears would have limited room interaction from a height standpoint preserving what was in the recording. Again, interesting. Been checking the local classifieds for a tolerable line source speaker to do some experiments.