I appreciate the work done in setting up this review, guys, and the results are interesting.
Now, to the forum monkeys:
Gentlemen, there is little meaning in agonizing over their choice of methodology. Take their qualitative observations into account and realize that the frequency response graphs are pointless to analyze. All the speakers in question are rated to +/-3dB or better across their playback range on-axis in an anechoic environment. It therefore follows that room effects and measuring equipment created the variation.
What I'd like to see is a completely subjective "relevant to me" type of review sometime in the future. In my case that would mean:
-a 2.1 channel setup, with a reference quality sub.
-a proper highend software or hardware equalizer. A software eq would probably work the best, since you could control the audio stream in the digital world & avoid analog signal manipulation
-get a proper, calibrated microphone with a known spectrum response, hooked up to a balanced soundcard with known response & distortion values.
-Establish 3 volume levels to test at: "sitting alone sipping scotch in the afternoon", "watching a movie", and "party time".
-Create fine-tuned EQ profiles for each pair of speaker at each volume. (ie, account for dips & bulges & level them out as much as humanly possible!)
THEN after the speakers have been fully tuned to precision:
-get a handful of volunteers who don't have audiophile bias
-get their hearing tested. That's right, get their hearing tested. The human ear's frequency response can be remarkably bad.
-run 3 sets of listening sessions for each reviewer:
-without equalization
-with speaker/room correction equalization
-with speaker/room/human correction
Do NOT use music, or atleast any music the reviewers have ever heard. That way, there will be no bias in their brains as to what that music is supposed to sound like (coloration memory).
Subjectives to test:
-microdynamic detail at bass, low-mid, mid-high-mid, high frequencies
-listener preference
Objectives to test:
-use a microphone (at room/speaker equalization settings) and see how comparable the output is to the input wave form. Use an autocorrelating function and see how close you can get. In theory the closer signal wins.
...If you want to do it scientifically