I have seen some discussion about the Databass measurements of FV15 in terms of distortion at max CEA output. I need to clarify a a few things.
First, the high distortion seen in those measurements (similar to the one done by Illkka) is caused the amplifier being push into clipping. Here is the typical distortion from an amplifier.
The distortion will increase when the output exceeds what the power supply voltage can offer (also known as clipping) and there is no graceful degradation between the linear region and amp clipping region. The old NAD 3020 has a soft clipping circuit with claim 20WRMS outputs. But when one disables the soft clipping circuit, the output actually goes up to 60WRMS. It does take that much buffer area to implement the soft clipping behavior. But even soft clipping has distortion in there except it is mostly 3rd order.
Second, depending on how Josh pick the test points, it can give very different impression to readers. This has to do with how he picks the db point in his measurement. A 3db difference (eg, 400W vs 200W from the above plot) can give a very different distortion picture. Even as small as 1db (eg 350W vs 300W) can give a very different distortion picture. So if Josh had picked just 1db or 2db too high for one sub vs another, the distortion plot can be very different.
Third, low distortion can be had using a limiter. What it does is it pick a point (such as 250W just for an example in reference to the above plot) and begin to limit the output above 250W. That itself is a distortion even though that method can game the harmonic distortion measurement system to give low distortion number. This is because sush systems can no longer provides a correct dynamic contrast of between, say, 35W to 350W dynamic range as that same dynamic range has been reduced to 35W to 250W.
So we need to get over the thinking that we saw a number of 15% distortion, we then infer it is better or worse than others because we don't have other information and that 15% number sticks. A subwoofer with 15% distortion at 110db output 20hz does not means it is 15% distortion at all other output levels or at any frequency point.