CBT speakers are interesting, and I like how they sound. I've heard them a few times. An issue that they have, is that it's hard to get any kind of 'sparkle' from them.
Here's an unprocessed response measurement from a CBT owner on Reddit.
Here's the processed response. Note that the high frequencies droop by about 10dB. That's why it's hard to get that 'sparkle' with a CBT. Like all arrays, you get tremendous efficiency when the elements are close enough to sum constructively, but at high frequency they interfere with each other.
Here's their speakers.
They wrote:
"I've never subscribed to the idea of full range speakers for anything but Jazz and I'm afraid that CBTs turned out to be no exception. I really was hoping they would be though. The CBT24 effectively sacrifices the treble performance for the CBT midrange. On one hand, the use of the Dayton 2.5" drivers made them more affordable than any other CBT but at the cost of performance. I'm convinced that CBT is a good technology for loudspeakers, but I would reserve it for 2-way designs where cost is not a design decision.
For $1500 it's not hard to find a traditional dynamic loudspeaker that can deliver a flat on-axis response with decent off-axis. For music, KEF LS50 and Ascend Sierra-2 are simply a far more resolving speaker than the CBT24."
Here's an unprocessed response measurement from a CBT owner on Reddit.
Here's the processed response. Note that the high frequencies droop by about 10dB. That's why it's hard to get that 'sparkle' with a CBT. Like all arrays, you get tremendous efficiency when the elements are close enough to sum constructively, but at high frequency they interfere with each other.
Here's their speakers.
They wrote:
"I've never subscribed to the idea of full range speakers for anything but Jazz and I'm afraid that CBTs turned out to be no exception. I really was hoping they would be though. The CBT24 effectively sacrifices the treble performance for the CBT midrange. On one hand, the use of the Dayton 2.5" drivers made them more affordable than any other CBT but at the cost of performance. I'm convinced that CBT is a good technology for loudspeakers, but I would reserve it for 2-way designs where cost is not a design decision.
For $1500 it's not hard to find a traditional dynamic loudspeaker that can deliver a flat on-axis response with decent off-axis. For music, KEF LS50 and Ascend Sierra-2 are simply a far more resolving speaker than the CBT24."
Here's what CBT shading looks like.
I was tinkering around in HornResp, and I think it might be possible to improve upon the CBT concept by doing two things:
1) I don't like the idea that the loudest driver in a CBT is on the floor. I know the floor is supposed to act like an acoustic reflector in a CBT. But I personally prefer that the acoustic center is in, well, the center.
2) Once we accept idea number one, I think it may be possible to put ONE tweeter into a CBT, in the center.
Here was my first stab at it. This isn't finished, it's just to give you an idea of what I'm getting at.
There are seven elements. The center element would be a tweeter in a waveguide. The tweeter is flanked by four midranges. The entire thing is shaded and curved, like a CBT.
I added a couple of large drivers at the end, the idea that I had was that the system would behave similar to a 2.1 system, where you have a midrange unit covering 200hz and up, and then a sub doing 200hz down. I don't know if I'm going to go that route, but I thought it's worth explaining what I was doing here, and why the woofers on the edge are larger.
The system covers an arc of about 90 degrees. The actual vertical beamwidth would depend on the shading, and it would be frequency dependent too. (Ideally one would optimized the frequency dependent beamwidth so that the beamwidth gets gradually wider as you go lower in frequency. You could accomplish this by manipulating the angle of the baffles and the crossover between tweeter and woofers.)
Here's a cutaway of the design, to scale.
Sounds like what you want is frequency dependent shading. At low frequencies the shading is typical CBT. Towards high frequencies the shading shifts so the driver at ear level driver has the highest output, and fading out towards the top and bottom of the array.
Therefore, drivers near the floor have a tilting-down frequency response, drivers in the middle have a tilting-up frequency response and drivers at the top are flat frequency response.
This could also account for the fact that soft floor coverings cease being a reflector at high frequencies.