If you don’t like the power alley that results from uncoupled subwoofer arrays and you do have six or more subs and enough real estate, you can try a compromised approach by aiming the left and right energy away from the center, improving isolation and lowering variance across the audience.

I should make it clear that the result appears to reduce the power alley only in contrast to the rest of the audience. There is still interaction in the middle, it’s just lower in level compared to on-axis with the sub arrays.
A reduction of 4dB at 63Hz is found at the center of the audience.

Download the MAPP3D file and run your own tests.
Why six or more?
The 2-element end-fire array is a one note wonder. It cancels at a single frequency in the rear. A more efficient option would be the gradient array, although there are exceptions.

How do you design an end-fire array?
Space the elements in a line so that their operating frequency range fits nicely in between the preferred filters recommended in Subwoofer Array Designer.

What’s the least amount of cabinets that can be used effectively?

As the number of cabinets goes up, so does the range of cancellation and consistency of coverage.
- 2 cabinets: Cancellation at 1 frequency. Could be useful for a fighting a single resonant frequency on stage. Otherwise prefer gradient array.
- 3 cabinets: Cancellation at 2 frequencies. Better an 2. Option to convert to 3-element inverted gradient stack.
- 4 cabinets: Cancellation at 3 frequencies. Now we’re talkin’.
- 5 cabinets: Cancellation at 4 frequencies. Even better.
- 6 cabinets: Cancellation at 5 frequencies. Begin to approach the point of diminishing returns.
Four elements is the most common end-fire quantity because it is effective and reasonably practical. Economizing to three units sharply reduces the randomization in the rear, leading to frequency-dependent reduction. Never end-fire with just two elements. It’s a one-note-wonder on the back side. Use the gradient in-line instead (same physical, different settings). We don’t have to stop at four, bearing in mind that the horizontal pattern narrows with quantity. Get crazy! RF antennas will end-fire 10+ deep.
McCarthy, Bob. Sound Systems: Design and Optimization: Modern Techniques and Tools for Sound System Design and Alignment (p. 321). Taylor and Francis. Kindle Edition.
A more consistent polar pattern?

I had never considered this before Tamas asked about it and I was excited about the possibility. Unfortunately, my experiments do not reveal a significant improvement using 2nd order all-pass filters over pure delay.
Frequency | Opening Angle w/Delay | Opening Angle w/APF |
40Hz | 172º | 160º |
50Hz | 152º | 152º |
63Hz | 134º | 144º |
80Hz | 120º | 116º |
100Hz | 106º | 100º |
100 – 40Hz | 66º | 60º |


One interesting side effect was the development of a MATLAB script to calculate the ideal frequency and Q parameters for each APF. Let me know if you’re interested in hearing more about that and I can update the article or send you the script.

How do I use SubAligner with end-fire arrays?
Measure the distance to the main array as you normally would, but measure the distance to the furthest subwoofer. All of the other subs in the array are aligned to it as well.



Here’s a direct link to this alignment if you’d like to use it in SubAligner.


Don’t we need to add 4th order filters, as suggested in Subwoofer Array Designer?
Normally, yes, but in this case there is already a native low-pass slope of 24dB/oct.

Have you tried end-fire arrays on your shows? What were your results?
Leave a Reply