Poll: Sonic differences between low/mid/high end CD Players are... - Page 8  

It's kind of like point/counterpoint. Actually, it's more like point/Jane, you ignorant slut...

Scott

Craig (to whom you were replying) is being disingenuous. Speakers are considered likely to sound different for good technical reasons (mainly that as electromechanical transducers, they exhibit 'gross' measurable differences, compared to, say, solid stae devices like CD players) -- and this is born out in blind tests. An assumption that two different speakers probably will sound different, is warranted.

Toole and Olive used blind protocols primarily to prevent preference bias when evaluating speakers which really do sound different. And indeed, speaker 'auditioning' as done by 99.999% of consumers is quite prone to such biases, based on visual cuues like size, shape, brand names, and prior information such as price. When Olive's subjects compared speakers blind, their preferences tended towards certain models (with similar measurable characteristics); when they compared them sighted, there was no trend, the preferences were all over the map, as might be expected where different people have different pre-loaded ideas about what a 'good' speaker is.

No, Krabapple, I am not being disingenuous. In fact, for you to say that about ANYONE else is pathetic. I approach this from the view point of helping people to get the most for their money. Blind speaker auditions do just that.

That seems to be more than your small mind can handle, so why don't you just move along ?

I have done ABX tests to compare audio files (e.g.. various levels of mp3 compression , compared to .wav source), which are quite easy to set up with free software.

I haven't done ABX tests of components, nor have I done randomized DBTs of stuff I buy from the pharmacy, for the simple reason is that I'm not set up to do them. That, of course, doesn't mean that controlled testing is pointless for the consumer. In an ideal world, someone reputable would be doing these tests on audio gear, as the medical establishment does for drugs and treatments, but since making a fool of yourself as regards an imaginary 'difference' between CD players is not health-threatening, I doubt the impetus is there. Heaven knows the audiophile magazines won't touch it, since the results might render their existence largely pointless.

The statistics are much more involved than that. For example, a difference *expected* to be small, you'd want a rather large number of trials to establish whether it was heard or not, and you'd want to use a p value of 0.05 or perhaps even smaller. Without consulting a p-value table, I would guess at least 15 trials for each person, with a success rate of 13 or better, to conclude someone heard a difference...assuming proper level-matching was done and the it was a double-blind test.

I know what you mean, but I'd put it a different way -- step one is to design an experiment that can answer the question you're asking...whether the answer is the one you are looking for , or not.



The Audiogon 'study' being cited isn't even worth citing, except as an example of how good intentions can be foiled by poor execution.

And where, my fellow molecular biologist, is the relevant literature to be found? JAES recommends blind testing but they don't do product comparison. The psychoacoustic literature is out there and available to those truly interested, but it too won't feature much int he way of product comparisons. So as with most products, one must extrapolate from the scientific data -- science for example has made some pretty secure predictions about the conservation of mass and energy, and thus one can make some pretty secure predictions about a product that claims to achieve perpetual motion.

Yup. It's hardly 'nonsense', and it's ignorant or shady to claim such. Subject independence is another 'given' in a properly controlled test. Subjects can and do influence each other in ways that aren't necessarily conscious; this is an uncontroversial psychological fact. (And ironically, the Audiogon protocl also fails a nonscientific 'golden ear' criterion, that comparisons should be done in a 'natural' setting, as if the person were listening at home...not in a group.)


Bingo. And this goes back to notion that CD players CAN sound different; the real question is how likely any two players are to sound different. A non-oversampling plaeyr stands of decent chance of actually distinguiishing itself sonically from a standard player. But how many non-oversampling (i.e., 'broken', IMO) players are there on the market? How likely are they to turn up in any given comparison?

The funny thing is that Audiogon's boutqiue player comparison might well have some audibly different units in it, and they could have bolstered their lame blind test with some good measurements to back that up. But noooo......

The 'golden ear' criteria for listening comparisons is subject to change according to whim. You may recall when Serinus and company came up with a null result on the power cord test he and others suddenly recovered their ears in a test of power cords at some member's home. Once they could see, it was patently obvious that there were audible differences.

Don't know. I'm admittedly new to this whole debate. I really don't have much experience beyond the occasional links that are brought up around here. Of the stuff that I've seen, I couldn't say that I felt that anything was a perfectly executed study. That said, I'm not sure that one is necessarily required for something as subjective as the hobby of home audio.

Truthfully, from one day to the next I'm not sure if the debate is that:

1) There are no differences between ANY CDPs
2) There CAN be differences between CDPs, it's just not likely
3) Anyone who spends more than $200 on a CDP knows nothing about audio and just has too much money
4) Anyone who hasn't spent more than $500 on a CDP is just whining about not being given their props in the audio community and looking to bash the financial "have's"

Hey...I'm just a confused mol. biologist looking to navigate his way through the always controversial world of the audio enthusiast. Had I known that a PhD in EE was required to enjoy a good CD, well...I guess I would have been an engineer. I just could never get past the overalls and checkered cap...

Scott

Yeah, Krabapple. The way I said it, it sounds like "set up your experiment to prove your hypothesis true" but the way you said it (the way I meant to say it) its "set up your experiment so that it can either prove your hypothesis true or it can prove it false." If your experiment doesn't pose the right questions, you won't get relevant answers. And "did you like A or B better?" doesn't answer the question of "can you prove that you can distinguish A and B from each other in a blind test?"

I just watched the movie "Sideways" last week, and it reminds me how just about every time the guy from Wings took a sip of wine, he's like "Oh, thats pretty good" with no further comment. I suspect the same thing may have been going on with the test above.

I think it's more like you've got to ensure the things that can affect perception such as output levels across the audio band are fairly close between units and then use objective means to evaluate a subjective phenomenon. If then there is no audible difference then perhaps you should factor other criteria into your decision.

For the record, I don't even use a dedicated CD player right now. I'm fairly confident that I would get a better result using one though, since I do subscribe to the theory that simpler is usually better.

Call it psychoacoustics or whatever you want, but I guarantee you that 10 times out of 10 I could tell the difference between using my Denon 1600 DVD as a stand alone player (i.e., analog connections) and using it as a transport coupled with my B&K processor (i.e., digital connection).

I guess that's where I'm coming from. In my system, I believe I hear a difference. So, I'm interested in hearing other components to see if I hear a difference there too. Will any two given CDPs sound different? I don't know...If there is a difference, does it justify the price tag? Again, don't know, but that's ultimately a much more situational consideration.

Scott

No doubt there was some serious whistling involved too.

And ssteel01, you of course don't have to have a phd in EE to enjoy a CD. You DO have to have some reasonable experience in experimentational science to judge whether a study concludes something properly or if there were errors in the study.

Your third and fourth options (the ones where anybody who spends more than $200 is a fool, or anybody who spends less than $500 is a cheapskate) doesn't take into account the fact that there are other reasons than sound quality to buy well built hardware. I don't care if somebody buys a cd player because it looks cool and is well built, because those are purely issues of taste. But "are there any sonic differences between two players" is not a matter of taste and has nothing to do with your ability to enjoy music.

If I were to sum up the two sides as I see it, one side is saying "There STILL hasn't been any good experiment that shows that even one person can reliably determine the difference between two cd players in a blind, level matched test. Merely asking them what they prefer in a blind test does not answer whether they could REALLY HONESTLY distinguish them at all." All in all, I'd compare the whole situation to "The Emperor's New Clothes."

The other side says "I can personally hear the difference, so you must be deaf or unwilling or unexperienced enough to hear it." The only problem with this statement is that it doesn't defend their own abilities. Someone somewhere may have better listening abilities than me, and those who claim to shouldn't be afraid to prove it. Except the response to "prove it" is usually "Why should I have to prove it? All that matters is my own preference." Except that it isn't even a question of preference, its a question of whether your preference is based on real phenomona or pure imagination.

You want to first define what you mean by 'differences'. Do you mean measureable differences , or audible ones? Being careless about such words is what fuels a large part of these debates.

FWIW, as regards audible differences, for me (2) reflects the state of affairs accurately, based on the best data we have on audibility and CDP construction. And since the poll as written fails to include that as an option, I won't be voting, but will continue criticizing it



Cute. It might even be to the point, if anyone had said or implied that an advanced degree is required to 'enjoy a good CD'. But no one has. I hope you don't rely on such strawmen in your arguments about biology.

I guarantee you that such a comparison is deeply flawed as a means of determining whether there is a real audible difference.


Suppose you do hear a real difference -- one that would be borne out in a blind test. Are you interested in knowing where that difference comes from -- whether it's something intrinsic to your two setups (meaning, they will always sound different), or whether it's just a matter of adjusting settings so they're giving audibly equal output?

If you aren't interested, or if you're somewhat interested but not interested enough to do the work to find out, that's fine. But then your claims about the 'difference' should be qualified to that extent, right?

Why is this no surprise ? You constantly barge in and whine about other people's tests, yet you have never done one of your own.

That is consistent with your personality.

Well, we could all be like you: Do nothing constructive, then bitch about anyone else's tests.

Actually, I'd say instead that 'There STILL hasn't been any good evidence that CD players tend to sound instrinsically different, or that we should expect them to." This leaves open the chance that rare pairs of CDPs might sound different, but evidence beyond 'I hear it' is always required to show that one or both 'buck the trend', whether it's information about the engineering of the CDPs, or bench measurements, or results of blind tests (concordant evidence from the latter two would be pretty much a slam-dunk; in fact, a correllary view of the 'objectivist' side would be that 'any CDP that passes a blind difference test, must show some measurable explanation').

Fair enough. I would say audible difference and I would tend to agree with you that (2) is where the interesting discussion lies.


Come on now. Of course nobody said that. It's just that a lot of times around here it comes off like you can't possibly enjoy your gear unless it was arrived at via extensive blind comparisons at some predetermined price point that is acceptable for what the product is.


You're right. I've never done a voltage matched comparison (I have done a simple correction for differences in SPL though...I know, I know...the accuracy...yadda yadda), but I know that I have a preference. I think it would be interesting to know if/where that preference comes from (e.g., DACs, power supplies, etc.), which is why I come here in the first place. I know it's just a preference possibly not predicated on any real measurable difference, but I know I like one better than the other. Pick any of the usual analogies used here (cars, cigars, wine, etc.). Even if there is a clear cut, measurable difference, I may not like the "better" one better. I'm just advocating the position that in the presence or absence of an actual measurable difference, you might bring something home and say, yeah...I like that better.

Scott

Well, in time, if one is interested in expanding their frontier of audio and testing knowledge, you will start procuring the cited publications and read them thoroughly. One thing leads to another

No one is asking perfection. After all, even peer papers are not perfect. Some are even fraudulent. But, gross issues such as claiming that a 2 trial event establishes anything is pure nonsense and must be pointed out over and over, or some will accept it as truth.

No one is requiring DBT for persona choices. But, when people post and make testable claims, then, that is another story and is the duty for others to scrutinize such claims and comment if warranted, or, again. mythology and voodoo becomes facts in some minds.


No one can point to a post that indicates ALL CDP are audibly the same. Some do read what is not written.
If one is after determining audible differences, there is only one way, bias controlled with some technical testing issues, such as level matching, to find answers.
One may spend as much as they wish on a component; that is an individual choice. But, again, some make testable claims for their purchases and rarely an issue of preference.


Then you may know about human bias and the need for controls when you seek reality

You cannot guarantee this as you have no real basis for it. You may be right, just maybe. But, that is a far cry from a guarantee. However, your chances are not so good realistically.
One way is to participate in a proper DBT, statistically sufficient correct guesses that you can then guarantee it.
Or, you know that one player has broken sections in it?

At least you admitting to having a preference. More than others do
I don't see a need for further discovery unless you really want to know. But, that needs measurements and listening comparisons, bias controlled, etc.
Would it bother you afterwards that you discovered that your preference is just based on other factors and not sound? If yes, don't seek the answers. If not, go for it.

Occasionally throw in a post on other threads in other areas - and surprisingly enough, they're not always argumentative .

No, it wouldn't bother me at all. I know that I have preferences for other things that may or may not be grounded in hard numbers and I'm okay with that.


Okay, I'll rephrase. I'd really be surprised if I couldn't. I think the chances of me hearing a difference between two CDPs in my own setup are the most favorable conditions. Certainly better than my expectation that I could hear the difference in someone else's setup, but I imagine nobody would disagree with me on that.

Just as an aside, I wonder what would happen if positive/negative controls were built into these experiments. For example, have a case where levels (or something like that) were intentionally biased and see if the listener(s) heard the difference. To me, if they couldn't, it would invalidate all other data. Sorry...just thinking like a life scientist again.

Scott

You should make it as favorable and conducive as possible.


Not necessarily. You could try something though, provided you've got the means to level match. Get two CD players that are the same, and using a test tone, make one something like 6 dB louder. Now, run some blind tests and see how you do. Gradually, make the difference smaller each time making sure you've got a reasonable statistical sampling. If you find you've hit a brick wall after practicing (let's say the difference is 1 or 2 dB), then that may be your particular discriminatory ability. It might though, not be mine. In any test, you've got to practice.

Hmmm. If you were really putting together a statistical survey, that would be an interesting baseline. Might be an interesting comparison to see if there's a particularly level of sensitivity that correlates with hearing a difference. I don't suppose that it would be a particularly surprising finding. Not saying that I have that ability at all though.

Just to throw out a tidbit so people don't think I'm one of the cd's sound better when they're colored green audio enthusiasts...

There's a switch on the Denon to select "audio only" which supposedly shuts off (bypasses?) the video circuitry in the DVD player. I suppose this should yield some sort of audible improvement, but I can't say that I've ever heard a difference when that is applied. Maybe it's broke...maybe I don't have the hearing/equipment required to appreciate the difference. Dunno...

Scott

Just to add another though on top of Chu's.
People do pick and determine audible differences to one component even when the same component is presented over and over, nothing really changed. Human nature to do this. We are wired to look for differences even when there are none we still do it at a high rate.

Just two papers on this, using test tones and the best conditions, lab, sound booth and headsets:

Florentine, Mary, et al 'Level Discrimination as a Function of Level for Tones from .25 to 16kHz. Journal of Acoustic Society of America, 81(5) May 1987, pg 1528-1541.

Zwislocki, J and Jordan H. 'On the Relations of Intensity JND's to Loudness and Neural Noise,' Journal of Acoustics Society of America, 79(3), Mar 86, pg 772-780.


There is a more recent AES Conference paper that shows how well we can differentiate other sources, speech, different music, etc, not test tones. So, psychoacoustic field has been very active over the past 100 years discovering




this is testable as well.