First of all, full disclosure up front - I'm a Harman (Revel / JBL) and Paradigm dealer, so feel free to be skeptical about what I post. However, there is solid science behind what I have to say - research you can easily verify.
Harman has the only double blind speaker testing facility in the world, where each speaker can be matched not only in volume but in physical location as well. Each speaker is put on the pneumatic speaker shuffler, matched precisely in volume to all models utilized in the test, and then hidden behind an acoustically transparent screen so neither the listener or test taker knows what speaker is playing at any given time. Each speaker is moved into the identical positions in under 3 seconds using the shuffler, and then the listener compares and rates each speaker using an iPad to enter the results. Revel's policy is that they won't release a speaker until it can reliably beat the competition in these double blind listening tests. B&W is a speaker very often used in the listening trials, since it often competes directly with Revel in terms of price and prestige.
The results with B&W have been all over the map. Some of their speakers do well, others very poorly. What astounds the engineers at Harman is the inconsistency. B&W has the resources to make a good speaker, yet their designs constantly lose to the Revels during the tests. Of course, Harman has the engineering brain trust and facilities to constantly tweak a design until it beats all competitors. The issue I am speaking to is how poorly certain B&W designs do when compared not only to Revel speakers, but other competitors as well. And the testing reveals exactly the same problems that
@DS-21 mentioned in his post:
B&W entered their "sound power problem" phase. Instead of flat+smooth on axis response and smooth off axis response, they use "minimalist" (an audiophool euphemism for "misengineered") crossovers and too high midrange-tweeter crossover points. The results in both colored response on axis, and off axis response problems.
Here you are referring to terms like "soft" and "muddy" in relation to perception of speaker sound. No, there is no "industry standard" that identifies those specific characteristics, but the correlation between measured speaker response and listener preference has actually been documented very well. Harman has published, peer reviewed research that shows that they can predict which speaker will win double blind listening tests with 86% accuracy just by looking at a set of graphs called the "Spinorama." (
https://www.soundandvision.com/content/15-minutes-harmans-audio-guru-sean-olive) To me, that clearly demonstrates that speaker preference is not simply a matter of personal preference. In almost all cases, the speaker that wins the blind listening tests is also the most accurate. This makes sense when you think about it - just like we want our video displays to accurately portray colors, we want our speakers to accurately portray the sound of instruments and the human voice. And, as far as an existing measurable industry standard goes, we have exactly that with CEA2034, which proposes that the Spinorama become the standard by which consumers can choose speakers based on a standardized set of graphs:
https://standards.cta.tech/apps/group_public/project/details.php?project_id=165
Right now speaker manufacturers are all over the map when it comes to offering specs for their products; adopting CEA2034 will mean that consumers could actually make meaningful comparisons.
It's no surprise that speakers that measure like this (the B&W CM1, graph from
www.soundstage.com):
Don't do nearly as well during comparative listening tests as a speaker that measures like this (the Revel M106, graph also from
www.soundstage.com):
(I tried to compare comparably priced bookshelf speakers from either manufacturer, and use a reliable third party source.)
The speakers that win the double blind listening tests not only have smooth on axis response (the sound coming directly from a speaker), but smooth off axis response as well (the sound coming from the sides of a speaker, out to about 60 degrees of axis).