Originally Posted by R Harkness
I find that the excuses offered by many audiophiles against the idea of blind testing to be frankly pathetic. I've seen otherwise science-oriented, rational friends become, in their audiophile mode, completely inconsistent and resorting to essentially "I'm confident I heard what I heard. If blind tests don't agree, then it's the tests that are wrong, not me."
I agree that competent and applicable DBT should be given its due weight.
I would be hesitant to criticize advocates of any competent DBT unless of course it was inappropriately applied to the wrong use case.
One such IMO misapplied and also inherently flawed study is the often cited Harman room EQ study that restricted itself to a single channel through a single speaker in a single room with three programs of pop music and (to all appearances anyway) no loudness compensation let alone matching of such compensation via calibrated reference level in the programming.
Despite these limitations that circumscribe the study and restrict its applicability I find it difficult to fault Harman. Who am I to tell Harman how much of their own money to spend on promoting their own room EQ? Their study is fine as far as it goes.
What I object to is the misapplication (specifically, overgeneralization) of that study as if it holds some sort of universal truth concocted and promoted by audio gods. Someone advocating Harman's curve once replied to my question as to where the default curve came from with an allegation that it took about 2 hours to come up with it.
This is how one develops a general-purpose default EQ curve that allegedly trumps all others?
Turning off the 'bright' Audyssey room EQ of my system is supposed to improve the sound of the bass in my room according to Harman expert listeners who preferred no EQ over Audyssey. What happens instead is that the unequalized response becomes thick mud (especially through the 'acoustically transparent' screen hanging in front of the disproportionately long dead wall) with boomy room modes on the bass and dull treble coloring the sound of every speaker.
IMO the inherent technical limitations of reproduction demand that case-by-case tradeoffs will compromise something
along the way in the practical world we actually inhabit. Engineers are intimately familiar with such tradeoffs and sometimes the hardest part is the balancing act.
Medicine stands out as the most outrageous scientific counter-example I can identify, with the supposedly advanced US spending approximately double on health care compared to competitive nations but ranking nearly tied in lifespan with Cuba that has been under embargo and dismal dictatorship for half a century. Rumor has it the Cubans might have cured cancer too with their recent announcement of a vaccine. By all objective measures Cuba should rank far below the US yet the practical reality is that all those extra expenditures in the US are wasted on redundant 'privatized' administrative costs and profit margin funneled to the very top earners.
This example demonstrates that DBT in medicine only applies in the lab not the marketplace yet here on this forum so many people seem to marry the DBT and the 'best practices' promoted by the experts as if the holy grail of audio nirvana cannot be achieved in any way other than blindly following the expert recommendations no matter what financial mayhem or suboptimal result that leads to.
I suspect that the neglect of real-life practicalities on the part of consumers and the experts they hired has derailed far more system builds than any lack of alleged DBT rigor or alleged bias in sighted listening tests or alleged self-promotion by manufacturers ever has, or ever will. IMO DBT has its place but it does not belong in your living room.
I had an outboard DAC, a Meridian CD player and a decent Sony CD player that I was sure were quite sonically distinguishable. I did some blind and double-blind testing between them and was able to identify each with almost perfect scores.
Unless you applied strict engineering and statistical methods and used dedicated, special-purpose hardware to eliminate all the 'tells', there is substantial reason to be skeptical that your testing was actually DBT.
To my knowledge, there have been no rigorous DBT of actual hardware DAC, at least none that were published in any credible journal. Also, DAC performance targets zero distortion in all aspects and things like load or other external factors are eliminated by design from the entire environment they function in to the greatest extent possible.
To my knowledge, amps are a very different animal, with some designs such as tube amps and tube-sounding amps employing deliberate design tradeoffs to introduce plainly audible even-order distortion that is sometimes pleasing (at least to heavy metal fans
) plus the intended use of an amp plays a huge role in its actual performance (particularly regarding output power and load impedance).
My limiting practicality is that competently designed amps without any deliberate signature and operating within their capacity all sound the same to my untrained and gradually failing ears. I suspect that many aging audiophiles face the same limiting practicality, especially those who have accumulated enough wealth over a presumably longish life to afford audiophile grade pro equipment with superior performance that they can no longer appreciate fully if at all due to hearing loss.
If so then it begs the question why there are so many vociferous advocates of DBT on this forum especially considering that the competently designed and published ones can be counted on the fingers up one nose.
Hard to base any purchasing decisions on a theoretical concept with no practical applications of it available to consumers.
My pal had the Quad ESL 63 electrostatics and (as I remember) a pair of Spendor BC1 speakers...I was startled at how hesitant I became in telling them apart! It was remarkable how alike they sounded to my ear. Ultimately I could identify each, but it was much closer than I'd ever have guessed from sighted listening to each speaker.
What used to be obvious distinctions between speakers has vanished and now everything sounds pretty much equally dull to me, even analog cassette recordings vs digital.
I suspect that the same happens to most aging audiophiles with aging ears so do not feel so bad about being unable to distinguish the sound of dipole from monopole. Just let it be a sign that you no longer have to spend so much on your equipment to maximize your enjoyment.
I have to admit very often I'd be hearing discontinuities, suck-outs, colorations, generally artificial sound, many of which would have me thinking I was listening to mid-fi, or a sub-too-small-sat system, with missing frequencies...I'd also habitually check the sound coming from the speakers to the "real" sounds of objects making noise and people talking about me, and this routinely exposed just how far away most sound systems (and microphone recordings etc of course) were from portraying the organic quality of real humans.
Ad-hoc demonstrations at shows are subject to severe acoustic limitations. You cannot expect them to perform at their top potential unless installed professionally in an optimized environment. Your example just points out that the in-room installation is a hugely important factor in the sound. IMO expo environment is for advertising not critical listening.
Regarding 'realism', that is another factor that needs to be tuned for. Do you want a movie scene filmed within a 'room' to sound like it is playing inside a room? Easy, just install a quality receiver/speakers and TV in your living room. Do you want an outdoor scene to sound like it is outdoors? Not so easy, now you need something akin to anechoic chamber and a really detailed mix such as full-fledged Atmos. Do you want a concert hall experience? A partly live dedicated home theater with a room simulation algorithm such as Audyssey DSX might do the trick but be mindful that to accurately reproduce the dynamic range of live performance you might need hefty speakers and amps too.
This 'realism' point is hugely interesting. What aspect of reality is being targetted? It makes all the diffence. A player piano sounds realistic, except it cannot improvise.
As for my use of tube amplification...Anyway, just letting that out of my system....
I have no use for tubes. Big, heavy, hot, power hungry, noisy, consumable, unnatural sound... everything about them is wrong. They were great in the 1950's. To each his/her own...
Originally Posted by R Harkness
Oh, one more thing about the HK blind-speaker tests, and speakers designed using the resources of the NRC... And they generally have left me unmoved, cold. The main reason that I think I identify is that although many of them sound "pleasant" insofar as sounding fairly full, easy to listen to, etc, they rarely sound "real."
Despite the Primus line having all the modern Infinity advantages including CMMD or whatever diaphragms, small cabinets and a great price, I find the prominently intelligible and forward sounding midrange of 3-way Primus towers inferior to the laid-back midrange of my 10+ year old aluminum woofer Sapphire/TSC 3-way towers designed by a co-founder of Infinity (Cary Christie). It is a matter of 'does it sound like the helicopter is taking off in my living room' versus 'does it sound like I am outside with the helicopter taking off in front of me'?
Whether or not either of these budget designs benefitted from an anechoic chamber to tweak them is an open question. Does anyone actually advertise such information on a speaker-by-speaker basis? I suspect that all the final tuning is done in an actual listening room rather than an anechoic chamber anyway. Can anyone shed light on this? I find it hard to believe that all the pro speaker designers out there have anechoic chambers of their own or even rent time in one.
The most typical (and depressing) difference I find in stereo systems is the lack of timbral realism...
IME Acoustic treatments and room EQ go a long way toward eliminating the 'boxy' sound of speakers in a room, but until your speaker system is capable of morphing itself into a cello and your room is capable of morphing itself into a concert hall you are bound to be disappointed.
My personal reaction to the "NRC" school speakers have left me wondering "why?" And I wonder if this has to do with what is being tested in the first place. For instance, the HK tests seem set up to discern which speaker characteristics are judged "most pleasing." But that isn't necessarily the same as "most realistic."
Maybe that is what I am hearing in the Primus vs. Sapphire. Undoubtedly they were designed by different people with different goals and different technology. I suspect that plays a more important role than where they were measured.
Originally Posted by FMW
If audible differences do not exist then amount of training will make them appear in a bias controlled test.
Self-explanatory and obvious.
No any sound will do, actually. The bias controlled testing we discuss here isn't qualitative. We simply seek to know whether an audible difference exists or not. We don't care about the quality of the sound, only the uniformity of its reproduction.
over my head. To me, fidelity is fidelity. Anyone with functioning ears can tell the difference between good and bad sound. When trying to tell the sonic difference with subtle variations in tech using trained listeners, how does the goal suddenly become shifted away from fidelity to 'uniformity'? The ultimate DBT is reality vs. Memorex. As difficult as that DBT is to construct, it is still the holy grail AFAIK
What am I missing? According to your statement it almost seems that as tech improves the minor differences in implementation start to be treated as variations on reality as opposed to deviations from reality. Are we dealing with a multiverse DBT?
I can't translate what you wrote into anything I understand but preferences don't play a part in bias controlled testing. We simply don't allow them.
Then according to your statement the Harman room EQ study I mentioned does not qualify as a bias controlled test because it stated its outcome in terms of preference, not accuracy.
You are concerned about what causes the bias. We ignore that. We don't allow any kind of bias in our testing.
Sorry, this sounds to me like you are claiming that your DBT is conducted by perfect machines instead of humans. Bias is an inherent quality of humanity. Only fictitious supernatural beings have no bias whatsoever. Your testing may control for bias but it cannot remove bias by definition or any other means. That is impossible.
Not trying for a 'gotcha' moment, just trying to understand. Contrast "We don't need no stinking experts" with "I ain't no stinking expert" and you get my drift.