I bi-amped my bedroom system about a year ago, with dbx 231 equalizers and Marchand crossovers. Now I’ve replaced those with with a pair of tiny Mini DSP 2x4 units. After a full five weeks of tweaking, finally got the system set up about right in my bedroom system. I say five weeks, but of course that's not full-time - I can only do an hour or two at a session - the spectre of madness requires this limitation. Can't work on it every day, either. So I'd say about 25 sessions. You can save your settings in a simple computer file, and then begin anew another day. BTW, once all the values are entered to the boards, the settings are non-volatile, and you need no computer when listening to music - just when you want to change any values.
This system is controlled by an HR-772 receiver, used as a tuner-preamp, with the power amp section only used to run my venerable Stax headphones. Those electrostatic "ear speakers" made a very good reference when voicing the system (a trick suggested by Bob Carver). The left and right preamp outputs split to run the headphone amps, and also connect to an ESC (not the "-U" type, but the earlier version, included as part of the first-run Silvers). The ESC has a Gundry Perspective switch as well as four other useful options. From the ESC the two channels feed a Trinaural Processor (analog), from which they emerge as three discrete (but not particularly discreet) channels, which then go to the DSPs that in turn put out my desired six channels (ribbons and woofers, L, C, and R) for the power amps - two 3-channel A-753x Carvers, over 400 wpc into four ohm loads. The 3 speakers are ALS Silvers, biamped directly - no power-sucking passive crossovers, so they need electronic crossovers and much equalization; it's a tough room, too. Also in the mix is a Sunfire Mk4 sub, so it's a three-way system. Proper balancing of all three frequency-arrays turns out to be relatively simple; voicing the ribbons and the upper bass is a female dog.
It's all about endless repetition of the same tracks, each cut having something to listen for. Toughest of all is midrange smoothness, from around A (440 Hz) to 3000Hz. All voiced by ear, mostly to various vocalists. Also dealt with: upper bass/lower midrange boominess (biggest peak around 235 Hz - it only happens in Silvers when the woofer array is in phase with the ribbons, and that is my chosen arrangement). Also: tizzy highs well above the 3300 Hz centered Gundry Dip (Bob says to put a filter in at around 6K - that helped, once I got the Q right). A smooth handoff from the Silver's woofers to the sub, from 80 to 50Hz, was relatively easy. [BTW, I just had Bill Flannery recap the Sunfire plate amp. Well worth it, and he did it fast. Sent it early, hoping he’d have it done and back in plenty of time to take to the Fest. Shit, he got it on a Monday and shipped it Tuesday!]
Adjusting the Mini's was largely a learning process. I’m using the parametric software, but even in this you have to deal with the fact that a tweak in one place will affect the frequency curve some undesired distance away. Going in, I pretty much knew what sort of a curve I needed, but making that happen is another thing. For extra fun, the gain is limited on this DSP model, so you have to learn how to avoid either overloading the first stage or lacking enough output voltage to drive the power amps to ear-bleeding levels. The shelving filters make it possible to get around this situation, but you have to learn their quirks, by trial and error as much as anything. Those high-shelf or low-shelf curves do some unexpected things, especially when you mess with their Q levels even a little.
Adding to the ordeal is the fact that I have three channels to do, and the equalizers are designed for just two, so a second board has to be plugged into the computer after disconnecting the first one, and all the latest settings for the center channel matched to the left and right before the next listening test, every time. So you can't tweak while you listen, except to one of three channels. If you cut some over-prominent frequency, you have to remember that ultimately you will be cutting it on two more channels, not just the one you are hearing. So it's slow going.
Luckily the trinaural processor has a subwoofer output (two actually, 180 degrees apart in phase), with an associated gain control; also a built-in optional sub-bass filter from 80 Hz, on the main outputs. My Living Room system (quite similar) has the 8x8 DSP, and that one big board can do all this by itself. But the 8x8 costs twice as much as these two little 2x4 models put together, and I thought I'd economize a bit for once. The company no longer produces a 4x8 model, unfortunately.
The 2x4 units are not too fussy in their DC voltage requirement, so my pair is powered by a junkbox 19 volt laptop power supply (the laptop being history, long since). It's rated 3.4 amps, and that's overkill for this job - it runs very cool.
The crossover point in the living room system was originally chosen as 200 Hz, higher than the factory's 150, to free the ribbons from having to deal with that half-octave that is really getting into their difficulty area - they're better for highs than lows, as everybody knows. And without the lows, the highs figure to be cleaner. I started this bedroom project using that same frequency (200), but at some point in the voicing I ran out of filters in the DSPs. There are six on each input channel, and six more after the crossover splits the signals, but it still wasn't quite enough. So finally I resorted to a little trick. Considering the original logic of raising the crossover point, and remembering that the sub relieves the woofer arrays of much of their requirement to cover the lowest notes, I simply raised the crossover frequency a bit more - to 235 Hz! Now, by raising the frequency setting a bit at the ribbon side and lowering the setting at the bass side, I can attenuate the response at that center frequency as much as desired, with no need for even one of those 12 filters, all of which are needed elsewhere. That freed up two – I was using some 235 cut on the input side, and more on the outputs. It’s too late for brevity, but I’ll at least spare you any explanation of that.
Having now biamped both Silvers and AL IIIs (both models have identical ribbons!) I can opine that the IIIs are easier to EQ, though I'd hate to have to choose which model comes out sounding better - any measurable difference could always be addressed by placement, tilt angle, room effects, and/or additional tweaks in the DSP. Having listened to both models, before and after, I can also say that there is no doubt that this bi-amping is a major upgrade, well worth the effort, and quite cost-effective.
If you go this route, beware: once the passive crossovers are disabled and the EQ and X-over work is done, you will never look back. I’d recommend starting with the 31-band equalizer software – it will get you to a fairly listenable version of the ideal curve faster, and teach you what Hz numbers are associated with various sorts of sound imperfections: growly, boomy, tubby, squawky, shrieky, tizzy, (and maybe make you Dopey, Sleepy, and Grumpy). Then, using what you learn there, go to the parametric software, to tweak and seek the final perfection; then you’ll be Happy.