Home |
Search |
Today's Posts |
#241
Posted to rec.audio.high-end
|
|||
|
|||
Vinyl's Comeback - featured NYTimes article
On Fri, 26 Feb 2010 17:52:11 -0800, Scott wrote
(in article ): On Feb 25, 12:02=A0pm, Audio Empire wrote: On Thu, 25 Feb 2010 06:56:41 -0800, C. Leeds wrote (in article ): Steven Sullivan wrote: Steve Hoffman has only the faintest idea what a blind test is, which doesn't stop him from advising his minions that blinds tests are of little use or meaning to an 'audiophile'. In practice, blind testing is of little use to the typical audiophile. = A proper blind test is not a trivial matter to conduct correctly; done improperly, the test is useless. The time spent conducting a test is time taken away from the raison d'entre for being an audiophile, which is to enjoy music. And, because it's a hobby and not a professional pursuit, it isn't necessary for an audiophile to "prove" anything to anyone, other than himself. Most of us recognize sound that we like, an= d sound that we don't like, and that suffices. Problem is, people write reviews and make pronouncements based upon these nebulous feelings about the sound of things. Why is that a problem? That is the nature of subjective review be it audio, food, movies or anything. *It is subjective.* The readers ought to know that. If they don't the problem isn't the reviews.... Worse, some lass than scrupulous or even well meaning but misguided individuals make and sell products of dubious worth to audiophiles who have no way of ascertaining whether or n= ot they've been duped. "Worth" like so many other things is actually subjective. If an audiophile thinks something was worth his or her money it's hard to argue. The point is to have fun no? If someone is having fun hard to argue value. In these cases, only double-blind tests can tell whether the "differences" heard between components are, in reality, the result of expectational/sighted bias (if you just spent $4000 for a pair of 1-meter interconnect cables, believe me, they are going to sound MUCH better than= the cables that they replace - even if they are, in actuality, identical in sound), or, if those differences are real. =A0 Of course, some people relish conducting equipment tests, and enjoy the rigor of managing a blind test. That's fine, of course. But, to suggest that typical audiophiles must also practice blind testing is just silly= . No one is suggesting any such thing. But, others have conducted such test= s, and the results have been published and are, in many cases, available on = the internet. One should avail themselves of these test results where possibl= e. Many tests have been "published" the problem is most of them are anecdotal and some who advocate blind testing as needed for "proof" have been caught cherry picking from the anecdotes. One can see clear as day hopw easily one particular anecdote was attacked due to the undesirable results. If you notice, I was speaking SPECIFICALLY about DBTs and was answering a poster who thinks that expecting the average audiophile to conduct DBTs is "silly". I've participated in a few blind audio tests. I found them interesting, but tedious - even when I was just a participant - and not especially useful to me. But, I don't design or market audio equipment, or I'd likely feel differently. Understand that double-blind tests are useful mostly for showing that the= re is a difference between the sound of two similar items, but not so useful= for determining which of the two is "better". No. DBTs are useful for removing bias effects. That can be applied to any test where bias effects are in play. There is nothing unique in audio about telling differences between two aleged similar items. Fact is bias is in play and has an affect on preferences even when gross differences in sound are present. That some audiophiles would limit their use of bias controls to try to prove a point they already beleive about differences seems futile. If you think about it. removing bias from the audition process is far more important when there are audible differences than when there are not audible differences. Think about it. I have. Double blind tests show that difference either exist or do not WHEN biases are removed. Seems to me we are saying the same thing. Remove the sighted or expectational biases, and even "true believers" can see the truth of these so-called "differences" (although many still won't admit it). For some, only a blind test answers their questions about the sound of audio equipment. For others, simple extended listening suffices. Yes, I know that the two methods are not mutually exclusive. But when listenin= g is sufficient, rigorous testing is unnecessary for the typical listener= . Like I said, DB or ABX tests are really for finding differences, not for determining which is better. That is simply not true. DBTs are really for removing bias effects. Bias effects are in play always when we are talking about subjective evaluations of perceptions. I dunno, when sighted tests find differences that DBTs show not to exist, then I would say that it's good at revealing whether or not the differences are real or imagined. In other words, we're saying the same thing, you just like the way you word it better 8^) I don't conduct blind testing of other consumer products that I purchase, either. For most products, all one needs to know about a product can be gleaned f= rom a spec sheet or a simple demonstration. That will not eliminate bias effects at all. Who said it did? |
#242
Posted to rec.audio.high-end
|
|||
|
|||
Vinyl's Comeback - featured NYTimes article
On 26 Feb 2010 02:31:44 GMT, "Harry Lavo" wrote:
Raises an interesting point. Wonder how many people could tell the difference in a Chevy Malibu vs a BMW 330i, if they were blindfolded, had earplugs, and were riding as a passenger? So do we start berating anybody who drives a BMW as being "autophools"? That is indeed an interesting point. However pretty much the whole point of owning, driving and being seen in a prestige brand car is to display one's wealth and superior genetic stock, peacock-style. Which is not the same as convincing oneself that an expensive piece of audio equipment *sounds better* than a cheaper piece of equipment because it looks and costs more... |
#243
Posted to rec.audio.high-end
|
|||
|
|||
Vinyl's Comeback - featured NYTimes article
I wrote:
In practice, blind testing is of little use to the typical audiophile. audio empire answers in message No one is suggesting any such thing Please don't pretend to speak for others in this group. For example, in this thread, Arny responded to the very same remark (in message ): Simply not true as a general rule. Others in the past have made similar claims. I recognize that you agree that blind testing is often of little use to the typical audiophile. However, many cling to the notion that it's an essential part of the audio hobby. |
#244
Posted to rec.audio.high-end
|
|||
|
|||
Vinyl's Comeback - featured NYTimes article
On Feb 27, 7:25=A0am, Audio Empire wrote:
On Fri, 26 Feb 2010 17:52:11 -0800, Scott wrote (in article ): Many tests have been "published" the problem is most of them are anecdotal and some who advocate blind testing as needed for "proof" have been caught cherry picking from the anecdotes. One can see clear as day hopw easily one particular anecdote was attacked due to the undesirable results. If you notice, I was speaking SPECIFICALLY about DBTs I was speaking about blind tests too. and was answering a poster who thinks that expecting the average audiophile to conduct DBTs i= s "silly". He is right. It is silly to expect anything of other audiophiles. DBTs are useful for removing bias effects. That can be applied to any test where bias effects are in play. There is nothing unique in audio about telling differences between two aleged similar items. Fact is bias is in play and has an affect on preferences even when gross differences in sound are present. That some audiophiles would limit their use of bias controls to try to prove a point they already beleive about differences seems futile. If you think about it. removing bias from the audition process is far more important when there are audible differences than when there are not audible differences. Think about it. I have. Double blind tests show that difference either exist or do not WH= EN biases are removed. Seems to me we are saying the same thing. Remove the sighted or expectational biases, and even "true believers" can see the tr= uth of these so-called "differences" (although many still won't admit it). No we are not saying the same thing. what I am saying is remove the bias and one can get an unbiased opinion on sound quality even when the differences are gross. My point being that if two things sound the same and someone prefers one over the other because of biases that person does not get objectively inferior sound. He gets the same sound but enjoys it more. OTOH when differences are real but preferences are swayed by bias then one runs the very real risk of choosing objectively inferior sound because of bias effects. This audiophile really does loose out due to bias effects. So while some audiophiles are obsessing over who is right about what sounds the same they are missing the boat when bias effects can actually have a real detrimental effect on one's choice of components. I find that extremely ironic. Like I said, DB or ABX tests are really for finding differences, not f= or determining which is better. That is simply not true. DBTs are really for removing bias effects. Bias effects are in play always when we are talking about subjective evaluations of perceptions. I dunno, when sighted tests find differences that DBTs show not to exist, then I would say that it's good at revealing whether or not the differenc= es are real or imagined. In other words, we're saying the same thing, you ju= st like the way you word it better 8^) That wasn't my point. You asserted that DBTs were limited in use to finding differences. They have a much broader range of use which includes determining unbiased preferences. For most products, all one needs to know about a product can be gleane= d f=3D rom a spec sheet or a simple demonstration. That will not eliminate bias effects at all. Who said it did? You said all one needs for *most* products can be gleaned from a spec sheet or simple demo. Apparently in *most* cases you are not concerned with bias effects in one's process of choosing. Why the limited concern for bias effects when they are always in play if not controlled? |
#245
Posted to rec.audio.high-end
|
|||
|
|||
Vinyl's Comeback - featured NYTimes article
On Sat, 27 Feb 2010 13:19:56 -0800, allen wrote
(in article ): On 26 Feb 2010 02:31:44 GMT, "Harry Lavo" wrote: Raises an interesting point. Wonder how many people could tell the difference in a Chevy Malibu vs a BMW 330i, if they were blindfolded, had earplugs, and were riding as a passenger? So do we start berating anybody who drives a BMW as being "autophools"? That is indeed an interesting point. However pretty much the whole point of owning, driving and being seen in a prestige brand car is to display one's wealth and superior genetic stock, peacock-style. Which is not the same as convincing oneself that an expensive piece of audio equipment *sounds better* than a cheaper piece of equipment because it looks and costs more... Oh, no, not at all! 8^) Not to mention the difficulty one would have driving ANY car blindfolded. They don't give blind people driving licenses in your state do they? They would here in CA though, provided you paid the worthless swine in Sacramento enough.... But there are more differences between a Malibu and a BMW than there is between two audio amps such as, for instance a Crown XTi1000 at $500 and one of those gorgeous (to look at) Classe CT-2200 amps at TEN TIMES the Crown's price at $5000. Of course the $500 Crown has more power into 8 Ohms than does the Classe (275 W/channel vs 200), but they are otherwise very similar and I SUPECT that they would be indistinguishable one from the other in a DBT, as well. Sure, like the Crown vs the Classe, the BMW has more ostentatious "bling" than the Malibu and costs a LOT more (though not 10X). OTOH, the BMW actually performs measurably as well as observably much better than the Malibu. The Beemer, however, is much faster, has lap times around a race course that are probably at LEAST half that of the Malibu. The BMW accelerates quicker, has a higher top speed, and stops from any speed in a shorter distance and will do it all day. IOW, you pay, Maybe $30,000 for a Chevy Malibu. The BMW costs probably $80,000 for a 5 series (I'm guessing, of course, but the price delta between the two is maybe 3X or 4X not 10X). The point is that you actually and inarguably get more car with the Beemer than with the Malibu while you do not really get ANYTHING more with the Classe vs the Crown except the bling and the bragging rights: "This is my new power amp, I paid $5000 for it!" |
#246
Posted to rec.audio.high-end
|
|||
|
|||
Vinyl's Comeback - featured NYTimes article
On Sat, 27 Feb 2010 14:13:49 -0800, Scott wrote
(in article ): On Feb 27, 7:25=A0am, Audio Empire wrote: On Fri, 26 Feb 2010 17:52:11 -0800, Scott wrote (in article ): Many tests have been "published" the problem is most of them are anecdotal and some who advocate blind testing as needed for "proof" have been caught cherry picking from the anecdotes. One can see clear as day hopw easily one particular anecdote was attacked due to the undesirable results. If you notice, I was speaking SPECIFICALLY about DBTs I was speaking about blind tests too. and was answering a poster who thinks that expecting the average audiophile to conduct DBTs i= s "silly". He is right. It is silly to expect anything of other audiophiles. DBTs are useful for removing bias effects. That can be applied to any test where bias effects are in play. There is nothing unique in audio about telling differences between two aleged similar items. Fact is bias is in play and has an affect on preferences even when gross differences in sound are present. That some audiophiles would limit their use of bias controls to try to prove a point they already beleive about differences seems futile. If you think about it. removing bias from the audition process is far more important when there are audible differences than when there are not audible differences. Think about it. I have. Double blind tests show that difference either exist or do not WH= EN biases are removed. Seems to me we are saying the same thing. Remove the sighted or expectational biases, and even "true believers" can see the tr= uth of these so-called "differences" (although many still won't admit it). No we are not saying the same thing. what I am saying is remove the bias and one can get an unbiased opinion on sound quality even when the differences are gross. My point being that if two things sound the same and someone prefers one over the other because of biases that person does not get objectively inferior sound. He gets the same sound but enjoys it more. OTOH when differences are real but preferences are swayed by bias then one runs the very real risk of choosing objectively inferior sound because of bias effects. This audiophile really does loose out due to bias effects. So while some audiophiles are obsessing over who is right about what sounds the same they are missing the boat when bias effects can actually have a real detrimental effect on one's choice of components. I find that extremely ironic. Yes we are saying the same thing and that's my final word on that aspect of this subject. We just express it differently. I believe that you are thinking about DBT testing in general and I'm thinking about DBT as applies to cables, amplifiers, CD players, speakers, audio tweaks, etc. Like I said, DB or ABX tests are really for finding differences, not f= or determining which is better. That is simply not true. DBTs are really for removing bias effects. Bias effects are in play always when we are talking about subjective evaluations of perceptions. I dunno, when sighted tests find differences that DBTs show not to exist, then I would say that it's good at revealing whether or not the differenc= es are real or imagined. In other words, we're saying the same thing, you ju= st like the way you word it better 8^) That wasn't my point. But it was mine. You asserted that DBTs were limited in use to finding differences. They have a much broader range of use which includes determining unbiased preferences. No, I said that they aren't very good for making value judgements about the quality of those differences. I never said that they were limited to finding differences. But I was also thinking solely about audio DBTs when I wrote that. And for sure, the usefulness in audio is pretty much to finding differences. But in other fields, they are far more useful. For instance, DBT drug tests can find out if a drug is really effective or not by giving some testers in the drug a placebo and other the actual drug. Those administering the test don't know which is which, and the testers don't know which they are taking. When the results are tabulated, and it is found that a statistically significant sampling of the group actually given the drug noticed real improvement of their condition, then it is generally clear that the drug exhibits SOME effectiveness. This test, however is unlikely to uncover the extent of the effectiveness, or any side effects. These require other DBTs designed to uncover those characteristics. For most products, all one needs to know about a product can be gleane= d f=3D rom a spec sheet or a simple demonstration. That will not eliminate bias effects at all. Who said it did? You said all one needs for *most* products can be gleaned from a spec sheet or simple demo. Apparently in *most* cases you are not concerned with bias effects in one's process of choosing. Why the limited concern for bias effects when they are always in play if not controlled? Because in most products, selection comes down to suitability to a need, specifications, and personal taste. For instance a man with six kids is likely not looking for a sports car to haul them around in. A guy who commutes 100 miles a day to work, is unlikely to purchase a 14 MPG Dodge Hemi Charger for that commute. A person who doesn't like brussels sprouts isn't likely to purchase a frozen entre containing them. But audio is different. People "think" they hear things (or get told that they hear things) that simply do not exist (such as a difference in "sound" between a 1-meter throw-away Radio Shack interconnect and a $4000 1-meter Nordost Valhalla interconnect or other obscenely priced piece of wire). When subjected to a double blind test between these two, whereby sighted and expectational biases are removed, and they hear no difference between the two, then the totally honest audiophile will admit that he is wrong. On the other hand, a man who says that two amps sound different and that his super expensive Whiz-Bang 2000 sounds better than another amp can be vindicated when the two are compared via a correctly set-up DBT. If he picks his Whiz-Bang 2000 a statistically meaningful number of times, then we are left with the overwhelming conclusion that that there probably is a difference between the sounds of these two amps, and that he was right. What that DBT fails to reveal, usually, is which one really sounds better. That's often a question of taste, and DBT's aren't very good at determining that aspect of the differences in the sound of two DUTs from a taste standpoint. |
#247
Posted to rec.audio.high-end
|
|||
|
|||
Vinyl's Comeback - featured NYTimes article
On Feb 28, 4:04=A0am, Audio Empire wrote:
On Sat, 27 Feb 2010 14:13:49 -0800, Scott wrote (in article ): Yes we are saying the same thing and that's my final word on that aspect = of this subject. And my final word is that we are not. There you have it. =A0You asserted that DBTs were limited in use to finding differences. They have a much broader range of use which includes determining unbiased preferences. No, Here is what you actually said. "Like I said, DB or ABX tests are really for finding differences, not for determining which is better " So it is what you said. I said that they aren't very good for making value judgements about the quality of those differences. I will accept that amended version. It is still completely wrong though. Bias controlled preference tests are actually very very good for making value judgements. And they are widely used for that purpose. I never said that they were limited to finding differences. Your words. " "Like I said, DB or ABX tests are really for finding differences, not for determining which is better " But I was also thinking solely about audio DBTs when I wrote that. Irrelevant. Bias controlled tests can be single, double or even triple blind. they are essentially all bias controlled tests just that double blind is better at removing sublime biases than single blind. DBTs are not only suitable for determining preferences they are ideal if one is concerned about bias effects which are present in any and all subjective preferences made without bias controls. And for sure, the usefulness in audio is pretty much to finding differences. I'm thinkin Floyd Toole and Sean Olive might disagree given the multimillion dollar facility they have at HK for doing blind "preference" comparisons for speaker development. You said all one needs for *most* products can be gleaned from a spec sheet or simple demo. Apparently in *most* cases you are not concerned with bias effects in one's process of choosing. Why the limited concern for bias effects when they are always in play if not controlled? Because in most products, selection comes down to suitability to a need, specifications, and personal taste. And audio is different? For instance a man with six kids is likely not looking for a sports car to haul them around in. A guy who commutes 100 miles a day to work, is unlikely to purchase a 14 MPG Dodge = Hemi Charger for that commute. A person who doesn't like brussels sprouts isn'= t likely to purchase a frozen entre containing them. But audio is different. Audio actually is no different. Bias effects are present in all subjective perceptual evaluations. People "think" they hear things (or get told that they hear things) that simply do not exist (such as a difference in "soun= d" between a 1-meter throw-away Radio Shack interconnect and a $4000 1-meter Nordost Valhalla interconnect or other obscenely priced piece of wire). W= hen subjected to a double blind test between these two, whereby sighted and expectational biases are removed, and they hear no difference between the two, then the totally honest audiophile will admit that he is wrong. It has nothing to do with honesty and the fact is in many cases the perception of a difference returns as soon as the bias controls are removed. It is human nature not a matter of honesty. On the other hand, a man who says that two amps sound different and that his sup= er expensive Whiz-Bang 2000 sounds better than another amp can be vindicated when the two are compared via a correctly set-up DBT. If he picks his Whiz-Bang 2000 a statistically meaningful number of times, then we are le= ft with the overwhelming conclusion that that there probably is a difference between the sounds of these two amps, and that he was right. If "vindication" is the issue then we are talking about egos here not audio. What that DBT fails to reveal, usually, is which one really sounds better. It all depends on how a DBT is designed. why anyone would bother with anything that is not geared towards preference is beyond me. well, actually it isn't. It is the test of choice for those consumed with debating these things. That's often a question of taste, and DBT's aren't very good at determining that aspect = of the differences in the sound of two DUTs from a taste standpoint As said before, you are plainly mistaken on that matter.DBTs are very very good for making bias controlled determinations of subjective preferences. Further more, logic dictates that the best use of blind protocols in the auditioning process would be when making a choice between components that actually sound different since there is no wrong choice sonically between components that sound the same. Ironic that the vast majority of DBTs don't actually make use of them in their own choices when they are of the most use. Clearly, this is where we disagree. |
#248
Posted to rec.audio.high-end
|
|||
|
|||
Vinyl's Comeback - featured NYTimes article
"C. Leeds" wrote in message
I recognize that you agree that blind testing is often of little use to the typical audiophile. As has been pointed out many times, audiophiles don't have to do blind tests themselves to benefit from blind tests. Blind tests have been crucial parts of the development of many different products that are employed by audiophiles these days. Since high end manufacturers rarely if ever develop basic audio technologies, neither high end audiophiles, nor high end manufacturers need to employ blind testing for high end audiphiles to benefit from them. |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
NYTimes article - Stereo Sanctuaries | High End Audio | |||
NYTimes is despicable | Audio Opinions | |||
MIX featured in "Soul Plane" | Pro Audio | |||
MIX featured in "Soul Plane" | Pro Audio | |||
MIX featured in "Soul Plane" | Pro Audio |