However, if improving the directivity above 5 kHz will be more audible, then would not spend time or money for better bass directivity.
What do you mean by "improving" here? Wider? Narrower? Different shape?
IMO the biggest sonic distinction from R1 - possibly "improvement," I suspect that could be room dependent but maybe not- would be to extend pattern control lower. That's also the most complex and expensive direction to go.
There is a nice overview of these sort of speakers here
https://www.linkedin.com/pulse/directivity-control-part-1-survey-loudspeaker-systems-steve-mowry
and a follow up page with a design
https://www.linkedin.com/pulse/dire...se-article_more-articles_related-content-card
If rear rejection is not the most important aspect then a supercardioid has a higher DI in the forward direction and is what you tend to get with a rear cancelling driver like the W371.
That's a fascinating find - thanks! A similar design has been bouncing around in my head as a potential "phase 2" for our 2-channel system.
Isn't the author the brains behind a lot of the TC Sounds subwoofers in the past?
On the other hand both D&D 8C and Kii3 are extremely complicated speakers with multiple amplifiers. Kii3 has five power amps driven by an integral DSP in each speaker and has a remote control! Neither designs are within the realms of a DIY project.
I don't understand this. Why can't a DIYer handle a project (assuming stereo) 10 processing channels? Here's an easy route in terms of hardware:
miniDSP SHD or SHD Studio - gives you remote control, AirPlay streaming, other inputs (even vinyl, through e.g. a Parks Audio Puffin phono pre with digital output) room correction, 2 processing channels for your tweeters, and two channels to digitally pass to the next processor
miniDSP 4x10HD or 10x10HD - gives 8 processing channels with the ability to expand to even more over AES/EBU
So that's your remote, source, and processing.
Admittedly, miniDSP doesn't have a unified control software a la London/Audio Architect or QSC Qsys, so you'll need to use two different plugins to set up the crossovers. Still, it's not that difficult to do. Certainly easier and less time-consuming than simulating passive crossovers, building, and tweaking. Plus Dirac Live if you want.
Even the amp section is very simple and inexpensive. You could, for instance, use of the Lexicon/Crown/Savant "Drivecore" 1RU 8ch amps amps for most of the channels, and maybe something with better top end FR or lower noise for the tweeters. If you want higher quality there are single-unit multichannel Hypex-based amps from Buckeye, Apollon, etc. Or used ATI made "zone" amps sold under different brands. That's at most 2-4RU worth of amps. Building the amps into the speakers is kind of dumb anyway, when you don't have to on packaging efficiency grounds. (Kii, D&D, etc. would need to design a nice case and also new shipping carton, etc. for a proper external controller box - much cheaper to build the amps into the speakers.)
Four channel DSP plate amps are rather rare. Therefore, it would probably be desirable, if possible, to limit the project to 3 channels.
This means that only one DSP channel is available for the bass module.
...
Or am I missing a simple solution possibility?
If three processing channels per audio channel is the limit, one could do a hybrid crossover - passive MT unit (passive parts to roll off the drivers and correct for individual driver breakup, use the processing channel to set the response, a la JBL 7-series i-line) + bass. That gives you 2 channels for bass. However, I kind of hate plate amps generally. I much prefer keeping all the processing and amplification components in one rack and only running low-voltage cabling to the speakers.