How do you high pass your main speakers?

@tony1954 wrote:

I have been very happy with the distributed bass array I added to my system, but from what I hear, the optimum method of integration is to high pass the main speakers.

Two questions:

1. What are my options for accomplishing this? Does this need to be a feature built into my amplifier or is there another component that needs to be inserted in the chain.

Preferably you’d need a DSP unit (digital crossover)/electronic crossover inserted between your source/preamp and power amp to your main speakers, instead of high-passing your mains through the (potentially available) high-pass filter function of your subs.

2. What crossover point would be ideal? What frequency and amount of rolloff would be best if my speakers are ATC SCM19’s which have a frequency response of (-6dB) 54Hz-22kHz.

Your ATC’s are passively filtered already, and running a DSP prior to your amp in addition - to some degree dependent on the specific DSP unit chosen - is really the least of a filter function interference compared to a passive filter.

What you gain high-passing your mains, on principle - certainly with a XO-point no lower than ~80Hz, and relative to the chosen filter steepness - is two-fold: the ATC’s are freed of a large chunk of LF signal, meaning less excursion, more headroom and a cleaner sound in this (great) driver’s band. And, similarly the amp driving your ATC’s will be rid of LF-signals, so it’ll operate more effortlessly and with lower distortion as well.

Will it be worth it? Who knows, only you can decide given your specific setup context and different factors that are at play here. I’d be inclined to say though: leave it as it is. You’re happy with your DBA-implementation - good, that's something to be appreciated.

Make no mistake: I’m an ardent proponent of high-passing main speakers in the 80-100Hz range crossing over to subs (to make the most of the added headroom this allows for), but I do this fully actively sans any passive XO’s, and so my speaker setup as an outset, incl. separately housed subs, is really approached as a whole, interconnected system per channel - like you would a single box; high-passing the mains in my case is not an extra "layer" of filtration, but rather designing each channel, in its different bands, as a complete setup from ground up.

@tony1954 wrote:

Actually my MiniDSP crossover is set to 80Hz for the subs so there would be a redundancy of frequencies from that 80Hz threshold and where the ATC's roll off around 54Hz. 

Perhaps it might be best to just tweak the crossover point a little bit lower to mitigate any spike between 54Hz and 80Hz.

Some prefer a slight overlap between the subs and mains, but I would use the supplied lower knee specs by ATC for your SCM19's and work from there setting an appropriate low-pass point juggling through the range here. The ATC's being a sealed design with a shallow lower end roll-off I suspect you would be better off lowering the LP from 80Hz and experiment with slope steepness. Using a DBA arrangement and setting delays, depending on the specific implementation, must be close to a nightmare to get right - although some may argue it's less important, but I digress. Your ears must decide, so whatever works, works (in my own actively configured speaker setup I use the same LP and HP frequency between the subs and mains with 36dB/octave Linkwitz-Riley slopes, so very little "redundancy" at the crossover. Proper delay settings here is of vital importance). 

Doesn't your MiniDSP have extra output channels? You already have all you need then to experiment high-passing your mains. 

Re: the OP and his sealed ATC SCM19 main speaker design with its huge magnet, large 3" and underhung voice coil SL-woofer, the distortion numbers here are already quite low up through the midrange, and my previous "leave it as it is"-answer (i.e.: don’t mind high-passing them) was partially in relation to this context. Granted, as a sealed design there’ll be excursion maxima at the tune, and thus there will be some influence into the mids at more elevated SPL’s with low frequency material exciting the cones into prodigious movement, not least if/when the voice coil exceeds the gap and its linear motion, but at more typical listening levels it’s a lesser issue.

Question remains: when will the efforts of high-passing them be worth it, and naturally this depends on the specific implementation of a HP-filter among other things. Given a fully active DSP-configuration steeper HP-slopes can be easily chosen, and it’ll make the HP-function and its sonic implications all the more effective. A mere 1. order slope here will still have some "bleeding" into LF-territory and is really only a half-baked solution - even with an 80Hz HP. Myself I’m implementing a HP over my mains at ~85Hz, 36dB/octave L-R slopes, and that’s with dual 15" woofers per channel w/100dB sensitivity and 800W continuous power handling - that is, tested with material into the bass region, so add a bunch of dB’s of effective headroom when high-passed as described. What’s interesting with such a powerful main speaker system is the difference a HP still provides for here; natively a ported design tuned at a relatively high 40-ish Hz, high-passing them removes the character of the ports, cleans up the mids even further with sharper leading edges, and seamlessly "delivers" them to a pair of very efficient 1/8 space loaded tapped horns that are quite "characterless" as well - not least when compared to the sonic imprinting of a variety of dual 18"-loaded direct radiating subs. All of this makes for a more seamless integration with the horn section from ~600Hz on up (111db sensitive). Indeed, here a HP over the mains makes a world of difference.