Bprice2 is absolutely correct about the bipole Omegas, in my opinion.
I have built bipole speakers going back to the late 80's, though not commercially.
You can think of the front baffle of a speaker as a 180 degree pattern horn, as it in effect focuses the radiation into a 180 degree angle (front hemishere) down to the frequency where the baffle edges are 1/4 wavelength away from the center of the driver (or, down to the frequency where the baffle is 1/2 wavelength across). At lower frequencies (longer wavelengths), the baffle is physically too small to control the radiation pattern, and the energy starts to wrap around the cabinet. This is commonly called the "baffle step", with the "baffle step frequency" being the frequency where the enclosure is 1/2 wavelength across, and the result of the baffle step is a theoretical 6 dB loss in on-axis sound pressure level in the deep bass relative to the midrange (above the baffle-step frequency). It's not an abrupt drop - it's a gradual shelving down, and in practice it's more like 3 or 4 dB.
Now note that the power response (summed omnidirectional response) of the speaker is unaffected by the baffle step. So whether or not to compensate for it is a juggling of tradeoffs - but that would be another post for another day.
Back to the bipole. The bipole very elegantly compensates for the baffle step in this way: At the frequency where the sound waves start to wrap around the front baffle, guess what - they also start to wrap around the back baffle! So the wrap-around from the rear-facing driver compensates for the baffle-step rolloff!
Now unfortunately it's not perfect - there is a path-length-induced time delay, which will cause a notch in the on-axis response at the frequency where the rear-facing driver is 1/2 wavelength farther away from the listener's ears than the front-facing driver. There are geometrical parameters that can be juggled to minimize this notch, and I can tell by eyeballing Louis's bipoles that he is using an appropriate geometry. In addition, this notch is a phenomenon that looks much worse on paper than it actually sounds. The ear really doesn't notice it. However, if Louis ever submitted a bipole speaker to SoundStage or Stereophile, its on-axis anechoic frequency response would not look very good in the lower midrange region compared to other speakers. When the day comes that there's a measurement protocol that correlates very closely with subjective preference, I think Louis's bipoles will score very high. The on-axis anechoic frequency response is not an accurate predictor of subjective preference.
I can think of at least six different ways to configure a bipolar speaker. Mirage used two of them, Definitive Technology patented a third, Omega uses the most simple and elegant, and then there are two more that I haven't seen anyone use in a commercial design yet.
In the deep bass region, where the path length difference between the front and rear facing drivers is an inconsequentially small fraction of a wavelength, an Omega or Mirage style bipole gives about a 3 dB boost. The designer can take advantage of this and tune the box lower than normal, resulting in fairly smooth and extended bass. Now note that the rear-facing driver is eating up half of the wattage you're putting into the speaker, so this extended bass comes at the expense of efficiency to a certain extent. Tradeoffs, always tradeoffs.
In my opinion, the bipole has some very desirable characteristics. Dollar-for-dollar, it has to compete with considerably more expensive forward-facing-only drivers. So its commercial niche probably lies with particularly high-value drivers, which Louis is very good at finding and getting the most out of.
I used to be a dealer for Louis, and think very very highly of his designs. He is getting performance out of fullrange drivers that I couldn't begin to duplicate. If I wasn't off playing speaker manufacturer in my own corner of the playground, I'd still be an active dealer for him.
Duke