How is this possible? (A hi fi review. Review)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Just browsing a bit tonight, and came across an amplifier review that was rather ecstatic in it's fulsome praise. And that is fair enough.

http://www.stereotimes.com/amp052008.shtml ''The simple truth is that the better the speaker (and of course, the source) the better the Lavardin IT sounds, taking on near-magical neutrality, subtlety, resolution, and clarity. ''

His approach on the correct way to build a sound system '''over the years varieties of system-building strategies have been advocated. The dumbest of these was the advice to spend most of one’s money on the loudspeaker''' does not coincide with mine, but that's another discussion for another place.

However, just how good is this amp??

""The old Linn approach – buy the best front-end you can – proved far more practically successful, even if one used a $200 amp and $200 a pair speakers. After listening to the Lavardin IT integrated amp, however, I’m convinced that the most rational, logical, and effective approach is to build one’s system around this superb amplifier. ""

Ok leaving aside whether or not I'd follow his system building approach, we are getting a picture of how good this amp is (and for all I know it may be the best amp ever built).

But what puzzles me, and led to my question, is the following exposition.

Mating a $7500 integrated amplifier to $289 or $400 a pair loudspeakers might seem insane, but there was method in my madness. Take for example the Celestion F30 loudspeakers (photo left). Part of the new Celestion “design-it-in England, manufacture-it-in-China” regime, the $400 F30, when coupled with conventional amplifiers, sounds like a sorry ghost of the grand old Celestion tradition. An upper-midrange/low-treble peak, which I had assumed was a sonic result of the metal tweeter resonance point, ruined an otherwise fairly decent speaker

To say that its performance with the Lavardin IT was transmogrified is a supreme understatement. The peaky treble was gone, transformed into an extended and artful accuracy and delicacy. The bass articulation and bandwidth flowered, the speaker took on an articulate rhythmic suavity that it had never showed signs of previously. The resolution increased to where a casual listener actually thought the speaker cost $10,000.

Just how can an accurate amp with near-magical neutrality, subtlety, resolution, and clarity alter and correct a poor speaker response?

surely rather than being an accurate amp it is instead a powered tone control?
 
of course, I agree. But it's a sad state of affairs when the poor old consumer (who knows no better and treats audio reviews as gospel) comes across something like this.

He establishes his credentials and auditory prowess by showing his ears are sufficiently good to pick the anomalies in the FR (as if that is the only outpoint in the FR of a $400 pair of speakers!) (the reader goes WOW he must know what he's saying) and then he comes up with a statement like that?:bigeyes:
 
Hi,

It contains more magic and less artistic distortion than lesser amplifiers
so of course it can do things lesser amplifiers cannot. If you have not
got one, of course you are wasting your time. Just do not ruin it with
any old cables, of course they must have some magic too.

:)/sreten.
 
"He establishes his credentials and auditory prowess by showing his ears are sufficiently good to pick the anomalies in the FR (as if that is the only outpoint in the FR of a $400 pair of speakers!) (the reader goes WOW he must know what he's saying) and then he comes up with a statement like that?"

What credentials?? What makes you think his ears are any better than yours?? If you think a $400 pair of speakers sounds good you should hear what it does for my free computer speakers that came with my Dell.

Rob;)
 
True - however the post really addressed the amplifier, specifically the amplifier making a loudspeaker significantly audibly better, which means the amplifier itself has to be significantly audibly better - it cannot have some "active" effect on a loudspeaker that somehow changes its performance characteristics.

Sheer and utter nonsense.
 
Until you have heard something similar occur..:devilr: :D

..and yes, I've heard a similar effect. (..early 90's Krell amp with the equivalent "best buy" mass-produced loudspeaker.)

As to *why* it can occur.. well I suspect it is in large part due to a massive (LOTs of current) and well regulated power supply - almost no cheap amps have this. The more expensive Lavardin's not only have such a power supply, but I also believe they have a chunky inductor right after rectifier - seems to make a big difference (aka "choke input") and is something that I don't think even the 90's Krell's had.
 
How to Conduct A Scientific Test

Now that this subject has been brought up, and because it touches on some of the endlessly debated issues in audio - namely subjective vs. objective evaluations, audible differences, etc., I would like to describe how a real scientific test is, or should be, performed, be it on a candidate for a new diabetes drug, or a new type of amplifier, or anything else.

If you look at how clinical drug trials are performed - those which determine the efficacy and safety of drug candidates prior to FDA approval (yes I'm in pharmaceuticals) - they are performed in a way which eliminates all possibilities of bias or subjectivity, and of course to weed out the so-called "placebo effect." These are called "double blind" trials, and it simply means that neither the trial subjects NOR the investigators conducting the trial know who is taking the actual drug and who is taking the placebo. At the end of the trial, after results have been recorded, it is unblinded to see who actually took the drug.

This sort of testing can be applied to just about anything, including audio, though in audio the ability to evaluate perception in a quantitative way is considerably more challenging. But many of the tests I've read about here fall fall short of their potential to be scientifically objective, and meaningful.

For example, if you wish to test an amplifier to see if it makes any audible difference in the sound a loudspeaker makes (i.e., use human test subjects), you could/should try it this way:

First, you use MORE than one person - statistically it takes a minimum of 30 test subjects for results to be meaningful. Then, you remove all variables - no one should be able to see the system they are listening to, because they will see the changes being made to the system during the test, and this can create bias, or the placebo effect - this is a blind study in the most literal sense.

The only thing the test subjects should be told is that they will be asked to tell if they can perceive any differences in the sound when a particular component is changed, but they must not be told when or even if a component has actually been changed (again, eliminates the placebo effect). The people conducting the test will also not be told when or if changes have been made, so they will not somehow unintentionally drop hints to the test subjects or ask the evaluating questions any differently at any time.

The only people who know what changes are being made to the system will be those exchanging the components, but they will have no contact with the test subjects nor with the people recording the results.

After all this is done, the test can be unblinded and the results collated and evaluated. There are still the variables regarding evaluating perception, and how to describe perceived differences, so the test questions have to be carefully constructed, and I have a feeling that is the truly challenging part, more challenging than in a drug trial asking people to describe if their symptoms have changed and in what way.

But the methodology should be the same. If someone has you listening to sound tracks and tells you "Now we are going to increase the harmonic distortion by a factor of ten on the track you just listened to, and we want you to tell us if you hear a difference," you have a totally bogus test that has been rigged and "front loaded" with all kinds of bias and the placebo effect built in.

It seems to me that this type of testing in audio is rarely done, very difficult and logistically as well as financially prohibitive, and at the end of the day it's the reason I ignore subjective opinions that thing A sounds better than thing B.

I find the hard data from actual measurements, done without human subjects, more useful and reliable.
 
$7500 for an amp and he doesn't advocate at least another $7500 for speakers?

Am i missing something in this article or is it just another example of marketing hype aimed at the gullible?

I'm with soft dome, sheer and utter hooey. Any amp having that much an affect on a set of speakers is as terry j put it - a powered tone control.

Does this mean all my amps are now near worthless? Or does it mean I will have to fiddle around with all the speakers because this one is going to change things even though I like them the way they are now?

Ack!

Where are those magic stones when you need them?
 
Re: How to Conduct A Scientific Test

sdclc126 said:
meaningful.


Thats entirely up to each person's perspective.;)

As far as the review is concerned - just how many people do you think really take such reviews as "gospel"?

I personally don't know of *anyone* that does.

At best what you'll find is a peaking of interest (as we have here), prompting in some a little personal discovery (..i.e. going out and listening for themselves).

If another has a similar experience - is it real, or is it a delusion? A silly question, because it doesn't really matter; the goal isn't about achieving a proven result.
 
sdclc126 said:
"Thats entirely up to each person's perspective."

Meaningful, in the scientific, statistically significant sense. I was referring to scientific study, which strives to avoid just what you mention.

I was well aware of that..;)

But who decides *if* that data is statistically significant?

..this underlies my suggestion: to look a little further down the "road" to see if there is any point in the first place.

The review is *subjective* - accept it as such and apply your own standards. If you find it "totally worthless" - thats cool, but try not to inflate your expectations of what you think it should be. It is what it is: light reading for a reader that enjoys it.

Want, or neigh *requiring* a rigorous double-blind study (as some seem to demand) is just an exercise in "wanking-in-the-wind". Its been tried before and it isn't marketable (..and lets not forget the source is in it to make money).
 
We're in agreement here Scott.

As far as who decides what is "statistically significant" - that is actually a mathematical term within the discipline of Statistics itself - not a subjective term. Think of the "bell curve" - if you go far enough out on it yes you may still have data but it becomes increasingly scarce and thus less "significant."

And I actually was accepting the review as subjective and applying my own standards - my response/opinion was based on those standards. I would not purchase the product (money aside) without further investigation.

I certainly don't "require" a rigorous double-blind study, and I pointed out its impracticalities in the case of audio - which is why I emphasize measurements - in the case of this particular amplifier and what the reviewer says about it that is very testable with measurements. And no, scientific studies are BORING to read and are not a marketable method here - their place is in scientific journals.

What I really just wanted to point out was that I've occaisionally read in this forum about so-called studies and tests that, upon a little scrutiny, are just about as useless as their subjective counterparts.
 
sdclc126 said:
We're in agreement here Scott.

As far as who decides what is "statistically significant" - that is actually a mathematical term within the discipline of Statistics itself - not a subjective term. Think of the "bell curve" - if you go far enough out on it yes you may still have data but it becomes increasingly scarce and thus less "significant."

..

What I really just wanted to point out was that I've occaisionally read in this forum about so-called studies and tests that, upon a little scrutiny, are just about as useless as their subjective counterparts.

Sure, it is an objective term.. but consider how data is manipulated to achieve statistical significance. (..the classical "Lies, Damn Lies, and Statistics".:D)

..

Such studies may be useless in a rigorous application of science.. but I'm pretty sure that isn't the point of the forum, nor does it mean that they are useless. Realistically the forum is VERY similar to a commercial magazine, but with far greater "community" (i.e. its much more interactive). The results from both (magazine and forum) however may have a wide degree of use, but again - depending on the perspective of the reader.


So yes, we are on the same "page" if perhaps slightly different "paragraphs".;)
 
"Sure, it is an objective term.. but consider how data is manipulated to achieve statistical significance. (..the classical 'Lies, Damn Lies, and Statistics' "

Tragically true, as this has even been done in drug trials in order to get a product to market, with in some cases the worst possible outcome - loss of human life.

Fortunately such extremes don't happen in audio - we just end up spending more money than (we later discover) we needed to. Or before we've spent our money we come here and find that not only can we spend less but we can get the kind of performance we thought was forever beyound our reach to boot. That may be one of the BEST functions of this forum - it certainly has been for me.

:)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.