The problem with making ASSumptions on how a speaker sounds solely on measurements without actually hearing them can lead one to believe listening tests aren't even needed. Just buy with a graph. I spoke with Dan Roemer of Perlisten about this topic. It's fairly easy to make a speaker produce good CTA-2034 measurements and still sound like ass. The ML F100s speakers actually sound very neutral. Matt told me he put them up against his Perlisten S7c inwalls and they were very close in performance and neutrality. I've heard the F100s in several rooms and thought they sounded very good. The bass is their strong point since the woofers are low to the ground and the port is on the floor, which minimizes ground bounce. Also factor in the speaker fit and finish is excellent. It's the first time I'd recommend a Motion series product. Their bookshelf is a different story as James Larson found out. We rejected that speaker and asked MartinLogan to resubmit once they tweak the crossover to tame the highs. Last I checked they were making a running change. Not sure the status on that.
Yes they are a difficult load to drive but not for a good Anthem amplifier. I've seen most set ups running these with Anthem electronics, not a cheap AVR.
I agree with you partially. Certainly with current measurements you can have a speaker measure well, but not be a particularly good speaker. However, I do think measurements do uncover defects. The problem is that measurements are incomplete, especially in the area of power response versus frequency. However, my experience is that when measurements do uncover an issue, more likely that not it will be audible, and that is especially true of frequency response errors in the speech discrimination band.
This is where speaker testing is different from amp testing for instance. When you get an amp on the test bench you test it right to the point of destruction where the distortion curve gets just about vertical. You can not do that with a speaker, but only infer it from experience of the design concept.
The problem with many speaker designers is, they so frequently do not take on board the power resources required in the power spectrum of most program
You can only really get a handle on this by getting a deep understanding of where the fundamental frequencies lay in the audio spectrum of common instruments and the human voice. Most of the power is going to be required in the cluster of the fundamentals, but harmonic envelopes can not be ignored either.
This is a major area where speaker evaluation differs from amp testing. On the bench you can test it at full power throughout the audio range. Pretty much it will develop max power at any frequency you choose. Obviously a speaker will not to that and can not be tested that way without destroying it.
What this amounts to is that I have learned to take an musicalogical approach, for want of a better word, to speaker design. This leads me to devote power resources differently to a lot of others. This really is the nub of the limitations of speaker measurement, and sets it quite apart from measuring other parts of the reproducing chain.
Lastly the confounding instrument is the pipe organ which is essentially unique in being able to produce huge spls. across the whole musical spectrum and all at once! So that instrument is the speakers designer's biggest challenge. They are kings and lords of the castles, and hard to satisfy.
