CEA-2010: A Good Baseline for Bass?

The result

Now I had my measurement procedure as good as I could possibly get it — and at least as close to perfect as anyone else’s. To my disappointment, the numbers I got fell in between the results from Hsu and Audioholics. My Ultra-Low Bass average for the VTF-15H in maximum extension mode was 117.2 dB — 1 dB below Hsu and 2.3 dB above Audioholics. Full results will be posted on the VTF-15H review.

And after measuring a few more subs, I found that my results were generally within a couple dB of the ones I was getting originally with the spectrum analyzer.

What I’ve learned from this exercise is that it’s probably impossible for any two technicians measuring different subwoofer samples in different environments with different gear to get the same CEA-2010 measurements. Any number of factors could throw off one or both technicians, and errors created by these factors can combine to create larger errors. These factors include:

1. Limited accuracy of calibrators; most are specified ±0.3 to ±0.5 dB, meaning you could have as much as 1 dB variance between two calibrators.
2. Sample-to-sample variation among subwoofers under test.
3. Differences in testing environment, such as ambient temperature.
4. Performance and level setting differences among the USB preamps used for testing. No particular model is specified by CEA-2010; I’m using an M-Audio Mobile Pre, a unit commonly used for speaker measurements.
5. Minor subwoofer positioning differences. Subwoofer positioning is specified in CEA-2010, but the specifications are somewhat vague and don’t cover some driver/port configurations.
6. Common measurement-to-measurement variation. As anyone who’s ever measured speakers knows, you tend to get slightly different results every time you run a measurement. Averaging can help (I set the CEA-2010 software to average results for 8 tone bursts) but it doesn’t eliminate the problem.

One change I’ve made to our lab reports as a result of this exercise is that we’re now including the CEA-2010 results at all frequencies, instead of just the Ultra-Low Bass and Low Bass averages. This change shows more clearly where a subwoofer’s output starts to dive.

For example, the VTF-15H averages 121.8 dB from 40 to 63 Hz, while the Cadence CSX-12 averages 121.5 dB. Practically the same, right?

But if you look at the full numbers, you see that the VTF-15H outputs 122.7 dB at 63 Hz, 121.2 dB at 50 Hz and 121.6 dB at 40 Hz. Compare this to the CSX-12, which puts out 125.6 dB at 63 Hz, 121.9 dB at 50 Hz and 117.2 dB at 40 Hz. Obviously, the CSX-12 (which has a smaller 12-inch driver and a smaller cabinet, and is less expensive than the VTF-15H) is tuned for maximum upper-bass punch at the expense of low-bass output. Thus, the full CEA-2010 results show that these are two very different subs, even though their CEA-2010 Low Bass averages are almost identical.

All this effort leaves me with one piece of advice for S+V readers: Evaluate CEA-2010 measurements carefully, especially if they’re coming from two different sources

It’s likely that CEA-2010 numbers from one entity will be reasonably consistent — i.e., if S+V measurements show that subwoofer A puts out 5 dB more at 20 Hz than subwoofer B, then subwoofer A will definitely give you more rumble when the depth charges go off in U-571. But if S+V says subwoofer A puts out 110.1 dB at 20 Hz and Audioholics says subwoofer B puts out 108.3 dB at 20 Hz, don’t assume subwoofer A is better than subwoofer B.

Furthermore, don’t assume a subwoofer that gets an extra 2 dB compared to another sub is significantly better, no matter who measured them. Day-to-day variations in the environment and tiny variations in measurement procedure and setup could account for a difference that large. And even if that extra 2 dB is real, whether or not you’ll actually hear it in your living room is questionable.

So as I’ve said for 20 years, and as other audio writers have been saying for at least 50: Read the whole review, don’t just skip to the end.