I would concede that CEA has merit in one aspect
of a sub's performance, that of output within certain targeted distortion limits within a certain band. However, as a measurement it does not tell the whole story of output, and it does not tell the whole story of distortion, and this is because it only measures a confined band of frequencies, 20 to 63 Hz. There is more to a subwoofer's performance than that. According to Tom, no one should listen to me since I am not published in the JAES, but if you want proof, just play some tones that sweep above and below 20 to 63 Hz. Hear that stuff? That's what CEA doesn't measure. Tom would love you to pay attention to CEA's region alone because outside of that band, his bread-and-butter driver just does not do very well
. CEA does not cover real deep bass, it calls 20 to 31.5 Hz bass 'ultra deep', which is laughable to many of the enthusiasts around here. It does not cover bass above 63 Hz which is actually a very crucial range and typically has more soundtrack content in it than 20 to 63.
Another aspect where CEA2010 falls short is the equal loudness contour
vs the audibility of harmonic distortion at deep frequencies. The harmonics of deep frequencies are way more audible than the fundamentals themselves. CEA does not account for this because their THD thresholds are fixed for every frequency. A superior metric would account for that. PSA does not want to acknowledge that because of the relatively high levels of THD their subs start hitting at those frequencies. But don't listen to me, I am just a crazy person who hasn't been published in the JAES.
Furthermore, what you won't see Tom acknowledge is that CEA has too much variation between for the kind of comparison he wants to make. CEA is good for a rough idea of how these subs might compare, but it is nowhere near precise enough for his sham 'value factor', where every decibel counts. Example discrepancies include audioholics and data-bass' results for the same exact FV15HP unit, the many different results for the VTF15h, and S&V's results for the SVS subs. Brent Butterworth even wrote a detailed four page article describing the limits and variations of CEA
in this respect, but that didn't stop PSA from using Brent's own measurements of Deftech's sub from being excluded from their bogus comparison chart!
Tom has dodged the critique of his subs poor CEA frequency response, by seemingly claiming that everything within the CEA averaged regions sound the same because CEA can be stated in averages, but I doubt Don Keel would agree. The averaging is what you do when you run out of print space:
CEA-2010 mandates a specific way to present the measurements, which looks like this:
Ultra-low bass (20-31.5 Hz) average: 110.4 dB
20 Hz 101.1 dB
25 Hz 107.3 dB
31.5 Hz 116.3 dB
Low bass (40-63 Hz) average: 118.6 dB
40 Hz 117.2 dB
50 Hz 118.8 dB
63 Hz 119.5 dB
Of course, you can arrange these numbers as you wish; the important thing is that they're all presented. Actually, you aren't required to break out the numbers for all frequencies; you can just report the averages if you choose. But unless you're pressed for space (i.e., in a print magazine), you should report all the data.
So either frequency response accuracy matters, or PSA ran out of webpage space and couldn't fit in all those extra numbers.
Regarding the XV15's inflated CEA data:
Originally Posted by Tom Vodhanel
The minor scaling that the reviewer in question has confirmed is correct on one model---the XV15. (adding about 1dB to his XV15 measurements because the driver was actually about 7foot from the mic versus 6 foot for other brands). Is is all simple math, and math we have performed accurately.
What Tom isn't saying is that, would the sub have measured a little bit more output, sure, but it would have measured a lot more distortion. By turning a subwoofer away from the listener (or mic), higher frequency output will be masked. If anyone needs this proved, just listen to your speaker facing away from you. So by having the XV15 tested as he did, Ricci was masking much of its harmonic distortion. If you want to know how that would have affected any output test which has a distortion threshold, ask any audio scientist or audio engineer. Ask Josh Ricci himself.