As someone who has built speaker boxes in the past, I can't help but notice that your designs are not baffled.
As any speaker DIY-er can attest, a "speaker" is really a system, consisting of a driver (the part people tend to call a "speaker"), a crossover (if applicable) and a baffle (i.e. the box in which the driver is mounted.) A driver alone is terribly inefficient, since the air around it is being filled with both positive and negative-going waveforms - the air is being pushed in one direction from the front and the other (equal and opposite) direction from the rear. These waves tend to spill out on their opposite sides, and reflect off of surfaces in the room, so phase cancellation is a big problem.
The very simplest speaker box design is an "open back" baffle, which is the common way of doing it with guitar amps. The sides of the box tend to keep the waves coming out of the rear from mixing with the ones coming out of the front. The drawback is a lack of bass response, so you don't see this used on bass amps. On electric guitars, the bass response isn't very low anyway so it doesn't really matter.
Early hi-fi speaker designs used a "horn baffle" design which sent the rear waves through a complex chamber of reflecting surfaces that acoustically flipped the phase of the waves by the time they exited the front of the speaker box. The drawback was that these boxes are typically huge, since the exit had to be large enough to accommodate the lowest waves, so the hole was usually as big as the driver, at least. Another way was to use a "passive radiator" which was in effect a driver with no power, that was pushed by the rear waves as they bounced off the rear of the baffle box, resulting in a (mostly) in-phase combination.
But both systems take up lots of space. Consumers wanted smaller boxes they could put on a shelf.
The simplest way to deal with the problem (and the main design of bookshelf speakers through the early 1960s) was the "infinite baffle", which basically was a heavy sealed box filled with some kind of dampening material like fiberglass fuzz or foam rubber to kill the rear waves. That way only the front-firing waves are heard. But this decreases efficiency of the speaker by at least half (only half of the power generated is going to make sound waves.)
Then two Aussies named A. Neville Thiele and Richard H. Small came up with means to calculate the resonant frequency of a speaker cone, and a formula match it to an "exit hole" in the front, with a tube attached that reaches toward the rear of the box. The "Thiele-Small parameters" of a speaker driver determine the size of the box, diameter of the hole, and the depth of the tube needed to result in phase-coherent waves and flat response in the front for a given size and mass of driver.
I've never seen a real RCA theremin speaker "in the flesh", but I'm guessing by the vintage that they were open-back designs, which nearly everything was in those days. I don't know how deep the sides were, but they were definitely there. They were not simply a driver mounted to a flat baffle board hanging from a stand. I can't help but think you're canceling certain frequencies and boosting others through the frequency range by not having sides on your speaker mounting boards. In practical terms, that means certain frequencies will be perceived as louder than others due to phase cancellation effects, especially toward the lower end of the scale. Do you notice any problem like this?
