If I were to roll my sleeves up and design a speaker system from scratch, I'd not use passive crossover components connected directly to the drivers—as is done in 99% of all speakers. Being an electronics guy, I'd self amplify this mythical speaker, assigning the right amp for the right driver: class D mega amp for the woofer, tube input/MOSFET outputs for the midrange, simple all-tube design for the tweeter.
To separate the frequencies, I'd use simple passive networks of very high quality, small size, working into the high impedances of the three power amplifiers.
Of course, this is only a dream, one I wouldn't even think to begin.
Even under the best of circumstances, it'd probably wind up sounding like crap. Why? Because I am not a speaker designer. I would be like the proverbial wanna be culinary expert that knows just enough to be dangerous: best ingredients + great recipe + no skill = good, not great.
But the concept I laid out is still a better alternative—in the right hands—than what most speaker designers have to work with. The smaller components in the passive networks can be of exquisite construction, the amps tuned perfectly to the driver.
Yet rarely do you see such a setup because it is simply not commercially viable. Audiophiles like to pick their own amplifiers and, besides, even if that weren't true, designers would be taking a huge gamble that their tastes matched yours. Once you go down the rabbit hole of an inclusive design, you're limiting people's choices.
That can be a good thing, but also, a scary thing.
There are those brave designs like the Devores, Meridian, Genelux, Emerald Physics, and I am certain I have missed a few.
Tomorrow I'll tell you about a speaker project I built years ago and what made it unique - and yes, it was active.