Crossovers are vital when it comes to configuring an audio system. Anthony Grimani outlines the steps to finding the right frequency for your project.

Setting the subwoofer crossover frequency. Something you’ve done a thousand times, right? Kind of seems old hat by now, yet it’s one of the most important decisions you’ll make. You may be thinking, “Come on, Anthony. That sounds like an overstatement. This is just a geeky obsession thing.” Let’s examine.

About the least-bad thing that can happen if you choose the wrong crossover option is that you’ll simply lose some extension in the bottom octave. More serious would be limited bass headroom, really bad standing wave problems, or gaps in the frequency response between the main speakers and subwoofers. Powerful, dynamic bass is probably number one on everybody’s wish list for home cinema, so your clients won’t be happy if their systems have bass problems.

The worst thing that can happen is drivers hitting excursion maximums, power compression, and ultimately failure of expensive woofers. You might be tempted to blame the client for “playing it too loud.” But it’s actually your fault for configuring the system incorrectly. Making that kind of mistake will instantly erode your client’s confidence in you and could open you up to thousands of dollars in repair and replacement costs.

Ok, we’ve established crossovers are important. How do you know you’re setting them at the right frequency? Whom do you trust? Here are some common practices you may be following.

• Don’t set a crossover at all (full range). This is not good for host of reasons, including loss of bass extension, low bass output, poor standing wave performance, audible woofer distress, and high risk of damaging something. But what if the main speakers are really big? Define big. If they don’t have one 30 cm or two 25 cm woofers in a large enclosure, then they probably won’t survive reference sound pressure level all the way down to 20 Hz. In short, there’s really never a good excuse to go “Full Range”.
• Just pick 80 Hz for everything, because that’s what THX said to do. THX actually said that for THX speakers, which they tested and certified to play down to 80 Hz with sufficient volume. Who is certifying your speakers to do that? Speaker companies don’t. The buck stops with you. Here’s something to think about: unless your speaker has dual 17 cm woofers in a reasonably large enclosure, chances are it won’t make it loud down to 80 Hz. That little bookshelf speaker with single 17 cm or 13 cm you love so much is going to need a higher subwoofer crossover.
• Set the crossover at the speaker’s specified low frequency cut-off (-3 dB point). That’s the low cut-off at the level that was used for testing the speaker’s frequency response, which is not as loud as you need it to play. When you push it harder, that -3 dB point may rise.
• Choose a crossover that is 10-20 Hz above the manufacturer’s rated cut-off. This is going in the right direction, but rather unscientific. The crossover may need to be even higher than that, such as in a very large room. Or, you may put it higher than necessary and get subwoofer localisation issues. 125 Hz is the mean at which people are able to detect localisation, so it’s risky to go above than that. You may need a beefier speaker.
• Go with the in-room measured low frequency cut-off. This is a popular choice with some automatic EQ systems. Unfortunately, it can be influenced by room gain and standing waves, which may cause a speaker to measure as having more bass output than it does. To use this method effectively, you need to test that the speaker can actually play as loud and low as you need without hitting excursion or power limits. Automatic EQ systems typically don’t do this, as the high SPL required for testing would scare the dingus out of consumers and pose some risk of damaging the speakers.

So how should you be setting the crossover? Honestly, the industry is still kind of figuring this out, but there are some steps that will set you on the right path.

You need to test every model speaker in a system. Start with setting the crossover in the Full-Range mode. Grab your analyser (free REW works great) and set up a spatial average about 6-8 feet away from the speaker. Put on broadband pink noise and adjust the volume to get 60 dB SPL. Capture the response curve.

Now repeat with the volume for 95 dB SPL. Overlay the two curves and compare. The 95 dB curve should not deviate in shape from the 60 dB curve. If bass rolls off higher/faster in the louder measurement, locate the highest frequency where the deviation first occurs. That’s the lowest you can set the crossover, but not necessarily the best choice.

If, in comparing the curves, you see that the louder curve drops below the softer by more than 3 dB at any frequency, you may be about to melt something. An AES standards group has concluded that deviations greater than 3 dB indicate imminent component damage. As such, you may want to step up the volume of your test more gradually (75 dB, 85 dB, etc.) rather than jumping straight from 60 to 95. Your call. You break it, you fix it.

Next, fire up your single tone generator and set the volume for 95dB SPL. Start at 500 Hz and sweep down. Listen for any obvious sounds of distress from the woofer: rattling, buzzing, motor-boating, distortion. These may occur at a higher frequency than the response deviation from the pink noise test. If so, set the crossover at or above the frequency where distress becomes audible. In other words, the speaker needs to pass the pink noise frequency response deviation test at 95 dB, and produce each individual frequency at 95 dB.

But, hey, you’re still not done! In the unlikely event that the speaker makes it down to 40 Hz or 30 Hz successfully, that doesn’t mean you should set the crossover that low! Producing bass from a single point in the room, like one speaker, is asking for trouble with standing waves. Bass needs to come from multiple points, with distributed subwoofers. The next phase of crossover setting is to avoid standing waves in the main speakers…but that’s a topic for another time!

Chase Walton ([email protected]) contributed to this article.