Once I had a pair of horns playing above 140 Hz, I naturally wanted high effeciency speakers for bass, too. Horns would be nice, but the dimensions need to be a decent fraction of the lowest wavelength being reproduced. In other words, bass horns are HUGE. They can be folded, but the result is not just huge, but complicated to build too. The designer can make comprimises to make the enclosure smaller, but they might not sound as good.
In my search I began to frequent an online forum for fans of high efficiency speakers. Someone there kept mentioning an old design called a Karlson. It is about the size of a ported enclosure for an efficient woofer, but it seems to have some of the characteristics of a bass horn, and according to him, it sounds better than either at that size.
Unfortunately, no one seems quite certian how it works. That wouldn't bother me except there are no guidelines for selecting a woofer like there are for a conventional ported or sealed enclosure. It does seem fairly forgiving of speakers that aren't an exact match, though. So I finally picked an Eminence delta12LF which I thought would sound good, in a ported enclosure if the Karlson didn't work out.
I decided just to build one for now. Even though there are much bigger speakers to choose from the Karlson seems quite large in my living room. Maybe someday I could sneak another one in unnoticed. Until then, one sounds quite good by itself. Imaging is fairly good, and the overall sound is much better than when I had two 8" woofers providing stereo bass in conventional ported enclosures.
Amplification is from push pull tubes. The output uses two EL34s wired in triode mode. They operate in class A with no global negative feedback. Actually the amp used to be a single ended stereo amp. I replaced the output transformers with one push pull transformer. The input comes in to one channel only. The input for the other channel comes from a voltage divider between the grids of the output tubes. The cathodes of the input tubes are tied together to improve the AC balance of the signal.
I used a first order passive crossover to mix the two channels and reject signals above 250 Hz.