 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
206.255.204.18
In Reply to: RE: Who decides if "...a few people who think their audio product or tweak makes no audible difference, but decide posted by Richard BassNut Greene on January 30, 2009 at 13:20:04
The challenge with conducting audio based DBTs (as opposed to medical ones which are easy) is that the tester can't know the answer either. So, if we're talking about using a jar of rocks (I confess I don't have the foggiest concept of placement or connection), how could you control the tester's knowledge of the presence or absence of the DUT?
rw
Follow Ups:
You're not alone. You must've inadvertently signed up for the slow class. 8 years should be long enough.
Edits: 01/31/09
do enlighten us with a primer on the correct placement and configuration of a jar of rocks. Surely years of empirical evidence has steered you to what you believe provides the optimum arrangement. Where does one start?
Perhaps you might use my main system and room as an illustration. You'll find that I'm certainly not tweak averse. You'll find high tension hospital grade outlets (which ensure the power cords don't pull themselves out), DIY versions of the VPI Brick (which are placed directly on top of the power trannies on the VTL amps), bass traps (which took days of experimentation in order to arrive at the best placement) and ceramic cable insulators (which are quite intuitive to place). How many jars do I need? Where do I put them?
You'll note I'm not asking for explaining their mechanism - I'm simply asking what advice you would give folks in order to maximize their enjoyment of such a product.
And, apparently since you're *fast* enough to determine that I'm slow, then kindly answer the question I advanced. We've got a system, jar(s) of rocks, a listener and lab assistant(s). What is the procedure for shielding the assistant(s) from knowing whether or not they've placed the jars *in the system* or "out of the system" without themselves knowing that answer? I'm all ears.
rw
Well, we'll see. :-)Everything you never wanted to know about Brilliant Pebbles can be found at:
www.machinadynamica.com/machina17.htm
Customer comments can be perused at:www.machinadynamica.com/machina19.htm
Edits: 01/31/09
How do you conduct the "DB" part of a validation test?
rw
I think you mean how do YOU conduct the DB part of the validation test. Answer is: any way you want to.
Cheers
No, make that a "brilliant" response and I wasn't disappointed. Apparently, you don't understand complex questions like "how could you control the tester's knowledge of the presence or absence of the DUT?"rw
Edits: 01/31/09
You're the big DB advocate, Mr. Complex, suppose you answer your own question.
You're the big DB advocate, Mr. Complex...
You completely missed the point of my question to Mr. El Nutto. At the expense of confusing the issue with facts, I've never been a proponent of audio DBTs. What makes you think that is the case?
Let's review what he said:
The only way to be very confident a tweak makes no audible difference is to listen to it using a double-blind methodology...
So, I'll repeat the question directed to him worded slightly differently. So, how can you ensure the tester(s) are unaware of the changes they must affect for the use of the tweak? Its not like they just hand out bottles of unlettered drugs with unknown contents. This applies equally to VPI Bricks, isolation bases, power conditioners, room treatments of all kinds and yes even pebbles among other things. Obviously, he hasn't thought through the question very deeply.
Comprende?
rw
I confess you lost me, might be just me.... "...unaware of the changes they must affect" is the phrase that throws me - what does that mean?
A true double blind audio experiment requires that neither the listener nor the device or person doing the component switching knows which component is in use. That is to eliminate the possibility of any cues being given as to the active choice. Such is very impractical for most kinds of comparisons unless of course you introduce uncontrolled variables like ABX boxes. Single blind testing is much easier to conduct in that the person making the switches for the listener can know which component is being used. I've done that for cable tests with my wife where I leave the room through one door, she enters from another, makes the changes (without my ever seeing her) and then leaves again. I return to the room and proceed with the test.
So, how can you possibly conduct a double blind experiment on any number of types of devices that must necessarily be put in place or removed - and the person making that change isn't aware of that which they are doing? The obvious answer is that you cannot unless they immediately use a genuine Men In Black Neuralyzer on themselves to clear their memory. Tony's response was equally funny in its approach. The point of my original post was to point out the glaring fault with Mr. El Nutto's "modest" proposal.
rw
It is entirely possible to conduct a double blind test (double blind to some standard, at least) for some "proximity tweaks" according to the following procedure:
1. X is the tweak in question. Assume it is of a certain size and weight. Assistant #1 places it in an appropriately sized cardboard box taped shut.
Y is a similarly sized neutral object of the same weight. Assistant #2 places it in an identical cardboard box taped shut. Assistant #1 then randomly permutes the two boxes and leaves the scene.
2. Assistant #2 then takes each of the two boxes, places them in two identical boxes that are slightly larger than those used in the first round. Assistant #2 then permutes the two boxes and leaves the scene.
3. [repeat as desired with Assistants 3, 4 , 5...]
4. The experimenter then selects one of the two boxes and places it in the appropriate location, according to the experimental protocol. The other box is removed to a remote location.
5. The listening test then commences.
This method should be applicable to audiophile tweaks if the following assumptions apply: (1) wrapping the device in several nested cardboard boxes does not impair their effectiveness, (2) there is a practical removal distance that neutralizes the effect of unused tweaks, (3) the neutral object of similar size and mass has no effect on the sound. If these assumptions do not fully apply then this procedure will lose detection power. However, should positive results be obtained, one can be confident they are caused by the tweaks and not by the "Clever Hans" effect. I make no claims as to the practicality of this method. That will depend on the available funding.
Note: I have participated in a blind wine tasting exercise that involved several assistants wrapping bottles in nested paper bags. There were three wines that the group had tasted several months prior. We were given their description and each poured three glasses. Then we were invited to identify the wines by name and vintage, the first one to raise his hand had the opportunity to demonstrate his expertise or make a fool of himself. As it turn out the test was quite easy.
If only DBT audio tests could be as pleasurable. [The wines were two vintages of B.V. Private Reserve and one of Heintz Martha's Vineyard. They were kindly provided for the test subjects by our host.]
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
shines a light on the incorrect assumptions being made with similar approaches. One ex-poster here asserted that his '74 Citation 11 preamp was *perfect* because he could pipe the output through the monitor loop of his Marantz 3800 and not hear any difference. The circular reasoning used here is that because each preamp is perfect and the addition of the other doesn't make any audible difference, then both are audibly perfect. Right.
He used a similar "shunt* test to demonstrate to him that all interconnects are audibly perfect. How? You insert your $.99 cable into the tape monitor loop of your preamp and switch it in and out of the circuit. Given that the other cables in the system are also of the $.99 variety and that the buffered monitor loop fails to acknowledge the real world of devices driving other ones, we have another example of circular reasoning going on here.
Brilliant.
rw
well, if one had a lot of bp in the room (say 10-15) it would be difficult. if 4 bp in room corners only it might be possible to perform blind tests. they're kind of small. I did a test a/b ces with 8 bp in room for sr. reviewer, no problem.
Edits: 01/31/09
regards the difference between single and double blind tests. Yours were of the single variety.
rw
I realize that. Actually don't think they were anything more than sighted a/b. Don't see any advantage to db, everyone's free to do whatever, obviously. Knock yourself out.
Edits: 01/31/09
"So, if we're talking about using a jar of rocks (I confess I don't have the foggiest concept of placement or connection), how could you control the tester's knowledge of the presence or absence of the DUT?"
I have a solution.
Connect the output of an ABX box to a tinfoil moving device. Randomly shield or don't shield the rocks by the tin foil, according to the choices made by the box. This should work, according to MIT proven science. :-)
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
The foil actually amplifies the effect of the rocks. It takes years of training to handle both at the same time.
unless of course, Mr. Kait provides a better answer. That is to say if he provides an answer at all. :)
rw
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:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
