This is a good discussion, but I'd like to get back to where it started, with me saying: "Evidence to date is that in such a situation of superimposed spaces, the larger one wins. In other words, we can make a small room sound much larger, but we cannot make a large room sound small. The fact that "room sound" is largely incoherent, non-minimum phase, reflections makes it easier."
Some of you may know about reverberation enhancement schemes. One that I know well and have experienced in several venues is LARES (Lexicon Acoustic Reverberance Enhancement System), that was created by Dr. David Griesinger, who also designed the Lexicon reverberation units that have been widely used in recordings for decades.
http://lcweb2.loc.gov/master/mbrs/recording_preservation/manuals/Lares Lexicon Acoustic Reverberance Enhancement System.pdf
There are other such systems, and all of them are able quite convincingly to fundamentally alter the reverberation and to some extent therefore, the perceived size of a venue. I have been in a medium sized auditorium, relatively dead for speech intelligibility, in which pressing different buttons on a console changed it to a credible concert hall or even a semblance of a cathedral.
Reverberation is measured as a time domain phenomenon, but in reality, to a listener in the space, sounds arrive from countless directions and times. Naturally, augmenting these requires many mics, processors, amplifiers and loudspeakers. It was first done in Festival Hall in London decades ago (1960s I think), which was widely criticized for being too dead. Clever minds conceived of a solution, and being well aware of human nature, they installed and activated the system over a period of time, working in secrecy. Over time, attendees and music critics noticed the difference, commenting favorably on the halls improved sound, and attributing it to numerous false causes, including a form of "breaking in"
. When all was revealed, it was a shock that electroacoustic enhancement was at work.
There are now several multipurpose halls using these schemes.
Logically, Dr. Griesinger downscaled LARES to LARES LITE for home listening rooms. At Harman we installed it in our very dead home theater. It required only four loudspeakers and mics, a processor and amps. My favorite way of demonstrating its was to allow the audience to acclimatize to the natural acoustics, conversing and listening to a few stereo demonstrations. Then, during a musical selection with significant spatial effect/reverberation I would slowly fade in the LARES sound. It would go unnoticed until I hit the "Pause" control and we heard a decaying reverberation as if we were in the performance space. From then on, any sounds generated within the home theater - shouts, claps, singing, etc. - sounded as though we were in a concert hall-like space. Jaws dropped. Then I would play some close mic'd recordings, which then were heard with the spatial sounds added, or not. It was impressive. One could "design" a listening space and then play whatever reproduced or live sounds one wanted within it. There was at least one other such scaled down system in the marketplace, but they all were expensive and I truly don't know if they still exist. The Lexicon system never became a mass-produced product. Too bad.
So, what does this mean for stereo reproduction? First it means that we must acknowledge that two channels are not sufficient to create a persuasive sense of envelopment. The listening room will inevitably be perceptible, especially in recordings employing close mics and pan potting - i.e. most pop and jazz. More uncorrelated sounds will help, but the only real solution is multichannel, in which long-delayed (say, 30 to 100 ms) sounds can be delivered from directions other than the front. Then, what is heard will be entirely the responsibility of the recording and mastering engineers, their tastes, their skills, and their equipment to process the sounds. But done well it is impressive how convincingly we can forget that we are really in a small room. For me the most impressive demonstration of this ability was an Auro3D immersive music demonstration in which I was "inside" a concert hall and a cathedral, able to walk around much of the listening room without losing the illusion.
Summary: small rooms can indeed be made to sound much larger, but stereo alone cannot do it. Recording engineers are in control and they differ in the illusions they try to create. There is the "artists in the listening room" approach, the "hole in the wall" through which one hears an orchestra in a large space, and the "you are there" approach. The latter one is the least successful one with only two channels.
The most obvious audible "signature" of small rooms is the collection of low frequency resonances. If these are not rendered less audible by the employment of multiple subs and/or equalization nothing else will be completely successful in creating the illusion of very large spaces. Done successfully, this is a substantial aid in "eliminating" the listening room. It works! See Chapter 8 in my book.