The measurement in that review may have some flaws. First of all, the scale is a bit odd, making it seem worse then the way these graphs are typically scaled. It only uses a 30 dB scale in the amplitude range. That will makes peaks and dips look worse than they do in normal graphs which use about a 50 dB scale. Second, it is a bit strange to take the graph all the way out to 40 kHz. Just because you can doesn't mean you should. Third, I am betting this response came from measuring this speaker too close. Klipsch designs their larger towers for a 3 meter use. If you measure at 1 meter for a speaker of this design type and size, you will definitely see a dip at the crossover. A speaker like this needs space for the sound of the drivers to integrate. I think that if measured at 2 to 3 meters, that dip would shore up. I don't know how they measured those speakers, but my hunch is that they probably measure better if done so with a technique that is more befitting of that design. Furthermore, an on-axis response is not enough to characterize the sound of a speaker. I wouldn't be so quick to write the RF-7 IIIs off by looking at that measurement. Don't read too much into that measurement, I don't think it is telling us very much about that speaker.
