|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
65.19.76.104
In Reply to: RE: "Accuracy": not simple, alas posted by Phelonious Ponk on June 28, 2010 at 20:14:04
If you found my explanation difficult, then I think I made my point. The question of transparency is not simple. :-)
This concerns preference theory. I see there is quite a bit of literature among the marketing theorists on this subject, but most of the interesting looking articles are not worth the $30 that seem to be the cost to read them. A simple example of how objects that have multiple attributes are difficult to rank in a consistent order is the game of "rock, scissors, paper". Which one of these is better? It depends on how you compare them, the basis of the game. The important point to understand is that preferences (as manifest in specific choices, be they answers to questionnaires or purchase decisions) are not just a property of the objects being compared. They are also affected by how the decisions are approached, the order of presentation, etc.
If a collection of audio components differed solely along a single technical dimension, e.g. had the same frequency response and distortion properties but differed only in the level of white noise, then it would be easy for subjects to rank these components according to transparency, and the results would likely be consistent (for most subjects) with the measured results. With multiple dimensions as actually exist with imperfect audio components, the situation is much more complex, even potentially paradoxical.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Follow Ups:
It's not the concepts I'm struggling with, Tony, it's the language. It's not at all clear what you're referring to as "distance relationship." What is meant by "physical dimensions" is no more clear, nor is "consistent ordering." At least they're not clear to me.
P
So as to make things concrete, I will talk about objective measurements, although I believe that subjective perceptions are ultimately more important. I will also confine myself to electronics, indeed only mono electronics, just to keep things simple. Even simpler, a line level amplifier.
One can characterize the performance of this device in three basic ways, noise, frequency response, and distortion. Oversimplifying, these separate concepts do not interact. Noise corresponds to output when there is no input. Frequency response corresponds to differences between input and output that are proportional to the input (linear), and distortion corresponds to differences between input and output that are not linear. All three of these categories can be further broken down since the effects will depend on frequency. You can not characterize the distortion of this amplifier by a single number. You will need a table of dozens of numbers. You find this in over simplified detail on a spec sheet or in more detail in the "measurement" section of articles in magazines. Now if you are comparing a few amplifiers and you are working with a few dimensions, say two to keep it very simple, you can plot the measurements on a piece of graph paper. At the origin you would have the "perfect" amplifier. I think most would agree that such a device, were it to exist, would be what you called "transparent". You could plot other devices as points on the paper. Now if you took a ruler you could measure the distance on the paper between two points and it would provide a measure of similarity. If you measured the distance between the point corresponding to the device and the origin you would have a "distance" that characterized how transparent the device was.
Here you see already the difficulty. If there is only a single dimension all the devices will plot on a line and it will be clear which device is more transparent than the other. But as soon as you have two axis a question of scale comes up. If you expand one axis and contract the other the ruler will place more emphasis on the attributes associated with the first axis and less on the attributes associated with the second axis. There is no obvious way to compare a given amount of noise vs. a roll-off of high frequencies above 10 kHz, for example.
Consistent ordering is simple. If I rank devices and always come up with the same order of preference, no matter what tests I perform, my ordering is consistent. This is possible, at least for some audiophiles who aren't suffering from audiophile neurvosa and if a small range of reference recordings are used for comparison. If many recordings are used or if many listeners get to express their preferences, it is unlikely that consistent results will be observed.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Tony I agree that it is not as simple as a few measurements at a few points along the path. I agree that those snapshots, taken for convenience and often abused, don't tell the whole story. I even agree that subjective listening is the most important measurement. So let's just say it's all subjective. My active rig in which the amplifiers, with great, though limited, imperfect numbers and loads of excess headroom are matched by design to the individual drivers in my speakers, is no more accurate than a turntable into a SET tube amp into speakers with an efficiency rating in the high 80s. Trust me, there are plenty who would say that such a thing. Never mind the wow, the flutter, the compression, the clipping, the very measurable and high harmonic distortion. If the lover of that vinyl/valve system thinks its best, it is.
And my little active monitors are better than the most elaborate custom midfield system, because I think they are. The very essence of subjectivism. Measurements don't tell the whole story, so they don't matter. Let's throw them out. Let's abandon the goal of accurate reproduction because accuracy is in the ear of the beholder. If it sounds good it is good.
Cool. Close down the discussion boards, brother. We have absolutely nothing left to discuss.
P
Enjoy the Chaos, that's what an artisan, interpreting the artistic experience is all about. More choices aren't bad, (although it may all seem overwhelming), follow your own path and enjoy the diversity.
Why does everything need to be so black and white?
"I'm not locked in here with you, You're locked in here with ME"
It doesn't have to be black and white; it really CAN'T be black and white. If the guy with the vinyl/valve rig tells me he LIKES it better, cool. Enjoy the music. THAT'S subjective. If he tells me it IS better, more natural, more musical or some other poetry devised to imply superiority (an objective value), I'm going to tell him he's wrong, by my ears AND by the numbers, and discussion will ensue.That's why we call them discussion boards.
P
Edits: 06/29/10
then those terms are very subjective....
"Natural," "musical," "transparent" are subjective terms.
Those flowery terms are not defining, nor are they implying (universal) superiority, - they are terms that describe an individual's preference.
Individual preferences are outside objectivity, - always.
You can say that the Manley Stingray is a tubed, integrated, amplifier. That is an objective fact.
You can also say that the Manley Stingray presents an absolutely lifelike transparency and accurate reflection of the recording: well, you're stating a subjective interpretation at that point. Wherein one can plug in an even better Manley NeoClassic amplifier and someone else will assert that it is lovely sounding. Or, someone could plug in a pair Audio Physic speakers and warm up the whole sound of the whole system, making it "less wonderful"
"I'm not locked in here with you, You're locked in here with ME"
"You can also say that the Manley Stingray presents an absolutely lifelike transparency and accurate reflection of the recording: well, you're stating a subjective interpretation at that point."
-- We'll have to disagree on that point. "accurate reflection of the recording" is not a subjective interpretation in my view, it is saying that the Manley Stingray puts out exactly what goes into it, with gain. And as I just told you in the other thread, that is also exactly what "transparency" means in this context. "Absolutely lifelike?" You got me there. That one is pretty subjective, and extremely vague.
Say it is lovely, warm, smooth..use any purely subjective term you choose that implies no more than that you like the sound, and I'll just congratulate you on enjoying the music. The problem, as I see it, is when audiophile manufacturers, and audiophiles, use quasi objective language in the description of purely subjective attributes. And it is the rule, not the exception. Is it terribly serious? Nope. But it gives us something to talk about.
P
> Say it is lovely, warm, smooth..use any purely subjective term you choose that implies no more than that you like the sound, and I'll just congratulate you on enjoying the music <
I'm with you. Unfortunately, when those comments are used, the replies are usually of the "You simply love the sound of distortion and prefer colored, inaccurate components" or something to that effect. Then the responses are "I prefer the lifelike sound of real instruments in real space rather then incomplete measurements"... and the arguments that have been going on for decades continue. :)
Sordidman and Tony seem to think that accuracy and transparency are moving targets. Perhaps that's the essence of the schism between so-called subjectivists and objectivists.
Nice thread!
Cheers,
"I'm not locked in here with you, You're locked in here with ME"
"I'm with you. Unfortunately, when those comments are used, the replies are usually of the "You simply love the sound of distortion and prefer colored, inaccurate components" or something to that effect. "
Yes, an unconstructive response, and one I've used more than once, I'm afraid. Unfortunately it is often the correct answer, even though "enjoy the music" works much better. If someone loves their out-dated technology, there is really no talking them out of it and no point in trying.
P
> Unfortunately it is often the correct answer, <
That's the whole point; no one knows if it's truly correct at the same time they're using it.
> If someone loves their out-dated technology, there is really no talking them out of it and no point in trying. <
True. And if someone loves their new-and-not-improved technology, the same goes. Again, the problem is the frame of reference i.e the subjectivity of it all. The bottom line is that whatever sounds the most "real" or "accurate" or "lifelike" to the listener is what IS the most lifelike. Unfortunately, chasing accuracy to music means I would have had to be there when it was recorded. Aside from a couple of occasions, I haven't had that pleasure. In the end, it's all subjective.
Even in its heyday as something similar to a science, back when reviewers who used measurement instruments were considered valuable contributors to the hobby, not fringe fanatics, hifi didn't chase the accuracy you're talking about. Studio monitoring systems don't even chase that accuracy directly. Microphones don't hear the way people do, so it is compromised before it even gets to the monitoring system. The accuracy you CAN chase, however, is fidelity to the source (high fidelity ring a bell? Not the excellent John Cusack movie...). That's a goal well worth chasing, a goal we'll have to count on the pros to pursue, because audiophiles, their vendors and their press have mostly abandoned it to subjectivism.
P
> That's a goal well worth chasing, a goal we'll have to count on the pros to pursue <
Unfortunately, they'll never know when and if they get there. It'll simply be yet another belief system. Best of luck to the pros.
I don't believe that accuracy is a moving target. It is an unattainable target, although an excellent goal. If it appears moving, it is because each time we attempt to corral it we catch only a partial glimpse of the situation, limited by the particular circumstances of our observations, be they subjective or objective.
Too many times have I compared components (or alternate versions of a recording) and concluded they sounded "the same" only to discover upon further listening that there was a distinct and repeatable difference. The same thing happens with purely objective measurements. With a new test or new test equipment two devices that had previously measured identically can measure differently.
There is a deep truth behind all of this and it goes far beyond the reproduction of music through the use of audio technology.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
agree with this assessment....FWIW
"I'm not locked in here with you, You're locked in here with ME"
nt
Nothing to argue with, there, Tony. Data gathering is imperfect. But there is a sizeable gap between what you're saying and the usual, "it's all subjective, you can't prove anything" argument which, upon further discussion, is too often revealed to mean "you can't demonstrate anything, you can't show valid evidence of anything, there is nothing in the world of science and engineering to show that any component might be more accurate than the ones I've chosen." This argument, made by manufacturers, reviewers and users, has resulted in the abandonment of that excellent goal to a sloppy subjectivism that broadly imagines, and then believes, a plethora of audiophile absurdities, many much more widely-adopted than shakti stones and black CD discs.
Proof is evasive. Evolution is a "theory." That doesn't make my mother a monkey. And it doesn't make clipping valves, just to use one common example, more "natural" than a well-designed solid state amp with more headroom than the load will ever require. I don't need proof to understand that is nonsense.
P
more meaningless words "well designed" another moving target with no real meaning....
"Evolution is a "theory." That doesn't make my mother a monkey"
Of course not: do you think that those are the only two choices? Do you think that Evolution theory is the best explanation that we have available?
Does it have a high probability of being "accurate?" Is it closer to the "truth"
Do you need to have a higher authority than your own listening comparisons to determine what is more accurate?
Whenever there are undefined, "moving targets" like "well designed," and "good sound," and "accurate" sound; how do you resolve a potential conflict or dispute when someone asserts something to be accurate?
Take the LAMM vs Halcro example. Are you saying that because the LAMM has tubes/valves it is less accurate? Are you saying that the LAMM is poorly designed?
"I'm not locked in here with you, You're locked in here with ME"
Not sure what you're referring to. Is there a classic battle of subjectivity between the Gods LAMM and Halcro? I'm blissfully unaware.
P
in this thread.
Don't know if you're worn out by it all. But I did want to thank you for helping me to think about a number of things, as well as bringing some pretty good arguments and reason.
Thanks,
As I mentioned, I used the Halcro/LAMM stuff as an example. As I read it, and I don't know if you agree, - but your definition of ACCURACY has to do with correspondence: like the correspondence theory of truth. (We listen to the trumpet live, then we listen to the trumpet through the audio playback system. The closer the audio system sounds to the real trumpet, the more accurate it is. As I see it, this can never, ever, be objective because human hearing is poor, varies (sometimes widely) from person to person, depends on other things like training, moods, level of distraction, and is by it's very nature, inaccurate).
It's like having several witnesses at a traffic accident each telling a slightly different story through slight variances in their perceptions: (like getting the color of one of the cars slightly wrong because they saw the accident from inside Starbucks).
"I'm not locked in here with you, You're locked in here with ME"
Thanks, but sorry, that's your definition of accuracy, not mine. Mine is much simpler and much more quantifiable. Let's say the component you're evaluating is an amplifier. Take a comprehensive set of measurements of the signal going into the amplifier and a comprehensive set of measurements of the signal coming out of the amplifier. The amplifier that doesn't change the signal at all is absolutely transparent. Pretty rare, I'd guess. What's not all that rare is the one that doesn't change the signal audibly, when compared to the reference amp, assuming the reference amp itself is designed for high fidelity to the source, not a "house sound," and is of very high quality, and assuming the two amps both generate enough current and wattage to properly drive the transducers without driving them into clipping.
I would love it if "transparency" could be defined relative to real instruments, but there are far too many compromises, beginning with the very first component in the signal chain, the microphone. Measuring and listening for the transparency of components relative to the source signal and a reference system, however, is not impossible and is done effectively all of the time. Belief in measurement is not popular in audiophile circles, and a lot of mythology and mystery has been built up over the last couple of decades to dismiss it, but it is, actually, a pretty reliable indicator of how both trained and untrained listeners will respond in listening tests. Go to Sean Olive's (Research director for Harman International) blog, Audio Musings. It hasn't been running that long; you can probably read through the entire archives in a couple of hours. You will be enlightened.
P
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: