The very best subwoofer blending (for a mono sub or for stereo subs) comes from using a combination of crossover slopes that were developed from theory (not by me) and first proven in the early eighties in pro-sound installs. The THX company has taken these useful slopes and named it "the THX crossover", but did not invent/develop the concept:
You would apply a 2nd-order (12dB/octave) Butterworth filter to the ProAc speakers and a 4th-order (24dB/octave) Butterworth filter to the subwoofer, at about 60Hz.
Why?
Because at ~60Hz, the ProAc woofer is mechanically rolling off its output at 12dB/per octave. This would be measured best up close to the cone's surface (I realize this is a ported speaker).
Add 12dB/octave electrical rolloff to that acoustic (mechanical) rolloff and you get 24dB/octave rolloff.
A good subwoofer, measured close to its cone, has a nearly dead-flat amplitude response over the proposed 40-80Hz crossover range, before any electronic x-over filtering. Thus, when you give it an electrical 4th-order rolloff at that same 60Hz, you would measure the subwoofer's upper-range rolling off at 24dB/octave, to at least past 150Hz.
So you effectively have a 24db/octave highpass filter on the ProAc's, and a 24dB/octave lowpass filter on the sub.
Which is the only filter combination that results in those two drivers being 360 degrees out of phase -one wavelength- which means the peaks and dips of any shared continuous sine waves line up. It would mean that the sub(s) and the ProAc's are phase coherent.
But they would not be TIME coherent- those two filters make one driver start moving a half of that 60Hz's period (1/120th second) sooner, and the other start a half period later. Which is thus "one period (360 degrees)" of difference. They are "In Phase", yes, but the beginning and ending of a tone burst at 60Hz are not clean. This info is in any filter-theory book.
What also mucks this up from being a perfect blending is the sealed-box sub itself, w/o any EQ, is rolling off at 12dB/oct starting down at, say 25-30Hz. This adds more time delay (phase shift) all the way up past 100 Hz, a delay that also changes with frequency (decreasing with increasing frequency). The max time delay, BTW, is at the 25-30Hz tuning frequency, equal to 1/4 of 1/25th-1/30th second (1/100th to 1/120th second). If this is a ported sub, that 1/4-period delay occurs probably up at 60 Hz, for even more unwanted time-delay through the crossover region. Any EQ in the sub adds more time delay.
This means the sub probably sounds best placed a little closer to the listener (via tape measure) than the ProAc's woofer, because it already is 12-15 inches farther away, acoustically/mechanically, at 60Hz than the ProAc's woofer.
If one could obtain a sub that has its sealed-box resonance down at 10-12Hz, then that "12-15inch closer" spec would not be needed, as the sub's own inherent LF mechanical time delay would be nil up at 40 Hz (though delayed 1/40th to 1/48th of a second down at 10-12Hz). Do consider that, if you have two subs, an "equal" tape measure distance could very well be reached by swinging the tape over to the sidewalls, for the subs' placement.
But that 10-12Hz sealed sub does not exist, so for "even greater perfection" from a conventional sealed sub, one would have to modify the electronic crossover, by adding an "all-pass" lattice, a passive-component network (even more complicated for a ported sub) to it. An all-pass network adds time delays to a particular range of frequencies, without changing the amplitude response, except for the loss at every frequency from going though all those passive parts. It is part of any analog telephone system- the low and high tones of the voice are delayed by the capacitance and inductance of the long copper lines, and the lattice network selectively delays the middle tones, so that all tones arrive at the same time, albeit later.
So try 24dB/octave on the sub at 60Hz or so (somewhere between 50-70Hz) and 12dB/octave on the ProAc's at probably that same frequency, and place the sub a little closer to you than the ProAc's woofers. Crossovers having those slopes can be built for you by Marchand Company and by dbSystems. A "THX" crossover has those slopes, seen in a "THX" home-theater receiver, but is often fixed at 80Hz, for a sealed 5.25" two way, or for a small, ported 6.5" design.
How do you know it is "right", by ear? Run the range with string bass, listening for mumble and stumble. Then apply a steady kick drum (Fleetwood Mac/Dire Straits), which is ~50Hz, and listen to the attack when moving the sub closer/farther than "tape-measure equal". Check with Stereophile Test CD #1, by quickly jumping from track to track, like a series of "burps", so that you hear more of the direct sound, and less of the room's build-up (which fools a an SPL meter, compared to what you hear on the transients of music).
Based on a great deal of experience, this 12/24 x-over at 50-70Hz would be the quickest way to improve the blending. These slopes are different if your upper-range speakers were panel dipoles like Magnaplanars, BTW.
Best regards,
Roy Johnson
Green Mountain Audio
You would apply a 2nd-order (12dB/octave) Butterworth filter to the ProAc speakers and a 4th-order (24dB/octave) Butterworth filter to the subwoofer, at about 60Hz.
Why?
Because at ~60Hz, the ProAc woofer is mechanically rolling off its output at 12dB/per octave. This would be measured best up close to the cone's surface (I realize this is a ported speaker).
Add 12dB/octave electrical rolloff to that acoustic (mechanical) rolloff and you get 24dB/octave rolloff.
A good subwoofer, measured close to its cone, has a nearly dead-flat amplitude response over the proposed 40-80Hz crossover range, before any electronic x-over filtering. Thus, when you give it an electrical 4th-order rolloff at that same 60Hz, you would measure the subwoofer's upper-range rolling off at 24dB/octave, to at least past 150Hz.
So you effectively have a 24db/octave highpass filter on the ProAc's, and a 24dB/octave lowpass filter on the sub.
Which is the only filter combination that results in those two drivers being 360 degrees out of phase -one wavelength- which means the peaks and dips of any shared continuous sine waves line up. It would mean that the sub(s) and the ProAc's are phase coherent.
But they would not be TIME coherent- those two filters make one driver start moving a half of that 60Hz's period (1/120th second) sooner, and the other start a half period later. Which is thus "one period (360 degrees)" of difference. They are "In Phase", yes, but the beginning and ending of a tone burst at 60Hz are not clean. This info is in any filter-theory book.
What also mucks this up from being a perfect blending is the sealed-box sub itself, w/o any EQ, is rolling off at 12dB/oct starting down at, say 25-30Hz. This adds more time delay (phase shift) all the way up past 100 Hz, a delay that also changes with frequency (decreasing with increasing frequency). The max time delay, BTW, is at the 25-30Hz tuning frequency, equal to 1/4 of 1/25th-1/30th second (1/100th to 1/120th second). If this is a ported sub, that 1/4-period delay occurs probably up at 60 Hz, for even more unwanted time-delay through the crossover region. Any EQ in the sub adds more time delay.
This means the sub probably sounds best placed a little closer to the listener (via tape measure) than the ProAc's woofer, because it already is 12-15 inches farther away, acoustically/mechanically, at 60Hz than the ProAc's woofer.
If one could obtain a sub that has its sealed-box resonance down at 10-12Hz, then that "12-15inch closer" spec would not be needed, as the sub's own inherent LF mechanical time delay would be nil up at 40 Hz (though delayed 1/40th to 1/48th of a second down at 10-12Hz). Do consider that, if you have two subs, an "equal" tape measure distance could very well be reached by swinging the tape over to the sidewalls, for the subs' placement.
But that 10-12Hz sealed sub does not exist, so for "even greater perfection" from a conventional sealed sub, one would have to modify the electronic crossover, by adding an "all-pass" lattice, a passive-component network (even more complicated for a ported sub) to it. An all-pass network adds time delays to a particular range of frequencies, without changing the amplitude response, except for the loss at every frequency from going though all those passive parts. It is part of any analog telephone system- the low and high tones of the voice are delayed by the capacitance and inductance of the long copper lines, and the lattice network selectively delays the middle tones, so that all tones arrive at the same time, albeit later.
So try 24dB/octave on the sub at 60Hz or so (somewhere between 50-70Hz) and 12dB/octave on the ProAc's at probably that same frequency, and place the sub a little closer to you than the ProAc's woofers. Crossovers having those slopes can be built for you by Marchand Company and by dbSystems. A "THX" crossover has those slopes, seen in a "THX" home-theater receiver, but is often fixed at 80Hz, for a sealed 5.25" two way, or for a small, ported 6.5" design.
How do you know it is "right", by ear? Run the range with string bass, listening for mumble and stumble. Then apply a steady kick drum (Fleetwood Mac/Dire Straits), which is ~50Hz, and listen to the attack when moving the sub closer/farther than "tape-measure equal". Check with Stereophile Test CD #1, by quickly jumping from track to track, like a series of "burps", so that you hear more of the direct sound, and less of the room's build-up (which fools a an SPL meter, compared to what you hear on the transients of music).
Based on a great deal of experience, this 12/24 x-over at 50-70Hz would be the quickest way to improve the blending. These slopes are different if your upper-range speakers were panel dipoles like Magnaplanars, BTW.
Best regards,
Roy Johnson
Green Mountain Audio