And what were the results? Did you publish them? I presume there were audible differences but was there a statistically significant preference for one cabinet over the other?
The results, when comparing a normal size 2 way with no bracing compared to the same driver array/crossover, in a vesrion of the cabinet with much lower amplitude response, was a detectable 'coloration' in the lower mid-range band. I can't give statistical 'preference' data, as it was ABX tests to detect only 'if' the difference was audible. It is entirely possible that simply adding 4 -5 simple cross braces into the 2 way cabinet could have reduced the cabinet audibility to zero - but I was only doing radical difference comparisons.
Another even more audible problem I have found is insufficient acoustical damping material used in speakers, especially lower priced ones. Simply installing a highly effective material in place of a poor one has made huge differences to mid-range definition/clarity.
I use a headphone with extreme low resonance behavior to compare for audible resonances in remote locations. I also have become extremely sensitive to resonances, far more than an untrained individual, due to my extensive auditory comparisons. So it's also possible that resonances that bother me, an average person may not notice unless they are trained to do so.
We did a pretty comprehensive study on the detection thresholds of resonances (and reflections) studying the effect of frequency, Q, time delay, level, program signal, and room reverberation on their audibility. Higher Q resonances are less audible than lower Q ones -- so dampening the cabinet resonances might actually make them more audible. Reducing the room reverberation, as you point out, can reduce the audibility of low Q resonances. U
I am quite familiar with that study and reference it often.
sing Finite Element Analysis you can model and simulate the resonances of the cabinet and then use our resonance threshold data to determine whether the resonances are below the masked threshold of audibility. With auralization tools you can even listen to different simulated cabinets/enclosures over headphones.
Unfortunately, I don't have access to this software, which I'm sure is not exactly cheap. I don't have a practical way to predict just 'how much' cabinet build is required to stay under the audible threshold, so I do build in a way to insure extreme low cabinet wall vibration amplitude.
I do have an idea/concept of how to actually acquire the raw cabinet amplitude response in a room vs. the driver response, then overlay them and refer to your resonance article to cross reference audibility. But I have not yet been able to get to this, as well as other important things I plan to do(I am going to build a large stereo automatic turntable setup for DBT of physical speakers behind acoustically transparent screens).
Harman certainly does not use 'lightly' built cabinets with any of their high quality lines, like the Infinity Prelude or Revel Studio, etc. Do they do this just for marketing?
I know the 802N pretty well, and have tested it against dozens of loudspeakers using over +300 listeners in controlled double-blind tests over the past 10 years. I have a large database of subjective ratings on its sound quality that can be largely explained by the comprehensive anechoic measurements I've attached above.
I submit that the EQ filters I specified earlier will correct it's SQ score by a substantial margin. The treble peaks will be very annoying, and I do not think it has sufficient baffle step correction for most placement situations that are far from a back wall. No matter how good a speaker, I believe it would sound 'annoying' with such a response. You should I.D. the $500 mystery speaker, though, as this may help a lot people.
The Behringer B2030P has power response that is superior to most speakers regardless of price. It's
measurements are very impressive, overall. A $150/pair monitor speaker, 2 way. But it will never be a contender against the 802 in it's stock form. It has no bracing, and the internal acoustic absorption material is marginal.
Excellent. Then you, better than anyone else, should not be forgiving of mediocre sound, especially when you are paying a premium for it.
One must be practical. Most commercial speakers are sub-par and it's an entirely different thing to design/build my own speakers compared to recommending pre-made ones. I can easily (and do) ensure text book perfect off axis response, and that drivers have no break up modes anywhere near their passband with my own designs as well as other important factors and it only costs me materials.
Why do no relatively low cost speakers exist that anyone can afford, that meet all of the ideal characteristics for optimum stereo playback without any significant compromises? I think it could be done, but no one that I know of has made such a device. Take the Infinity Primus 360/362; it has superb off axis response through most of it's passband and uses excellent drivers that have break up modes far past their passbands and the drivers it uses have very substantial dynamic range capability. It has a good crossover design. But it has a huge bass peak, it has very poor internal acoustic absorption material and it has a very questionable cabinet. With what seems would not be much more of a manufacturing cost, it seems they could have made this into a world - class speaker in the average Joe's cost range. Is that not done because it would screw up some kind of market balance for performance? Afraid it would cut into the 'high end' Harman products offering?
-Chris
