If you were building a xover from scratch and wanted to experiment, and have your results listenable quickly, then doing it with DSP hardware and separate amps would be the answer, but building for manufacturing wouldn't work that way. It's great for experiments, but for mass-production, if you can whip out a crossover with a hand-full of passive parts inside the box, that's cheaper and more controlled than the active way. The DSP device you need to experiment this way would be fairly expensive, though, or would be a stack of cheaper units. And, some of the control software for cheaper DSP devices isn't conducive to optimizing a crossover, more of a sound system EQ tool. Don't forget, if you're starting from scratch with this, you need good analysis tools too.
