Is Speaker design evolution stagnant

Jack,

There are many ways to get into it, but in my experience it has been very rewarding to go active and take the learning curve. I echo what Erik suggested about diyaudio (multi-way forum).

I started my journey 4-5 years ago. I had a great preamp (Lamm LL2 deluxe) and loved it, but thought a system optimized to do one thing very well would be better, for a given amount spent, than a system attempting to do several sources as good as possible. I bought a multichannel DAC, took the Lamm out, built an "audiophile" computer to be my only front end, and implemented active crossovers between my subs and mains. The tube amp was happy not attempting to reproduce below 80Hz, and so were the B&W 804 speakers. Sound improved.
Then I purchased "class-D" amp kits from Hypex, assembled, and now have those 400W driving each B&W bass section. So I removed the passive crossover and have the Hypex directly connected to the woofers, running from 70 to 350Hz. For now I still have my McIntosh tube amp driving the midrange and tweeter section, where the passive xo between them is still in place. So it's a 3-way active.

In the process I learnt a lot in many aspects. I am now working on a pair of DIY 4-ways, fully active, to replace the B&W. I really love tubes, and DIY active allows to design to maximize the benefit. For example I want to use SET for the mid and treble sections and SETs are constant voltage sources so high and constant impedance speaker is a real benefit for them...so that narrows it down for me pretty good.

I belive conditions have matured to go into this direction. Digital signal processing is enabling stuff unheard of some years ago. Low power Class A amps aren't very expensive these days (SETs or Nelson Pass' First Watt, for example), class D amps of high power and good enough performance for bass are cheap these days, professional speaker drivers are pretty good and a lot cheaper than exotic audiophile stuff (which I reckon I also like). A Rythimk subwoofer kit is cheap. You can build a 3 or 4-way system that will sound super. There is also a lot to be gained with good box (or lack of) design practices - things that tend to be too labor intensive for manufacturers to adopt).

It's been a rewarding experience for me. The big trade off is the time I've spent, and will continue to spend. But that's part of the fun for me.

Jack,

 A significant point Bache made and I neglected to mention is the need for measurement. This means a mic and learning how to do it. USB mics are easier to implent but not as good as the internally do analog to digital conversion with a clock that is different from your DAC. But an easy way to get going, which in my mind is most important.

Incremental improvements rather than radical changes. The learning curve for radical changes is even steeper than the upfront cost and likely to frustrate.

Seanheis,
In my view that is a low probability scenario and wouldn't detract me. Don't know why you would ship a speaker. 

Kosst,

What you describe is true, sometimes. As usual it depends on implementation. Active digital xo aren't that complicated to implement and can solve the issues you describe if done right. At a fraction if the cost of passive. But the system has to be designed to work that way - not just the xo.

Why this isn't used more frequently? The learning curve is steep. It's a lot more involved. You need to be able to measure. It's a lot simpler to connect an amp with a pair of wires to the speaker,  play music and call it a day.