Audio Critic

[IXP]

Orinally posted by PianoTeacher
I claim he cannot pass his proposed protocol, whether or not the phenomenon exists, because it is wrong: it can only be done theoretically on paper but not actually by ANY human test subject.

This is blatantly false.

Assuming the GSIC does something to the disc, and assuming that you do not know what it does, you cannot possibly know that the proposed protocol could not be passed.

Let us say that the GSIC superposes a 1000 Hz steady tone at the maximum amplitude representable by the binary coding on the disc. You have asserted that WellFed, indeed that no human being, could not possibly distinguish a treated from an unteated disc 10 times out of 10. This is simply absurd.

IXP

 

[alfaniner]

Re: Diogenes, hold up your lantern a bit higher! A rough informal speculative protocol

...
Once again: apologies for my verbosity as I continue to have trouble in refining these complexities to the *irreducibly minimal number of words*. ...Best,
PianoTeacher

Just wondering, have you ever heard your piano students grumble "[SIZE=-2]Geez! Just shut up and let me play!![/SIZE]"

 

[Skeptical Greg]

Claim:


" I can tell the difference between black signs and white signs. "


Test:

" Present the claimant with black and white signs ( ten of each to be more specific ) .. "


Analyze results:

" How many did the claimant correctly identify? "


Someone please explain to me under what conditions, or for what reason/s the claimant would not be able to score 100%?




PianoTeacher:

" We must subject these signs to spectrographic analysis to determine, if in fact, they are black and white . "

" We must also analyse the containers the paint came in, the composition of the base material of the signs, and the manufacturing process of all materials involved."

" Anything less, and it would not be possible to ascertain whether or not the claimant can actually do what they claim to be able to do. "

 

[jmercer]

Nonsense, PT. Either he can tell the difference as he claims, or he can't. THAT IS THE FULL AND COMPLETE TEST PASS/FAIL CRITERIA.

99% of these ridiculous gyrations is to prevent cheating by JREF or Wellfed. Fine. There's a simple and practical method for preventing cheating.

Have 20 disks sent from the store(s) to a unbiased trusted third-party (such as an independent major accounting firm, or local police, or any neutral trusted third-party) by pre-arrangement. Have a GSIC chip shipped to the same third-party direct from the manufacturer as well.

The day of the test, have this third party OPEN the chip package and treat and mark the first 10 disks in private (without JREF or Wellfed). Have the second group of 10 disks opened (again, without the presence of the JREF witness or Wellfed), and have those disks marked without being treated.

Then the third party will discard the chip, place all 20 disks into a package, seal it with tamper-proof tape, and transport the package directly to Wellfed's house where Wellfed and JREF participants are waiting. The third party then opens the package in front of EVERYONE, the test begins, and both the third party and JREF witness tally the marks chosen.

If both tallies match, the results are then considered accurate representations of the choices made. Any discrepancies in the tally would cause the test to be redone with the exact same disks. Once we have an acceptable tally, the third party will reveal which mark reflects the treated disks... and the tally will tell the tale.

Stop making this more than it is.

(edited to add "treat and mark the" in the fourth paragraph, and removed redundant sentence)

 

[Beleth]

Re: Response to Beleth about his personal test speculation

I suspect that the differences that Michael tries to perceive are much more subtle than the ones in your theoretical test.

That's what I was getting at. But where should the line be drawn? What qualitatively differentiates an "obvious" difference from a "subtle" one? And what differentiates a subtle difference from no difference?

The reason Michael is held to a 10-out-of-10 standard is that it's a binary test (is it treated, yes or no), and that the standard for the challenge is to have a minimum of a 1000:1 chance of passing the test through luck. Nine binary choices don't cut it; that's only 512:1. Ten choices do cut it at 1024:1.

We could build some sort of margin for error, but to keep it at the 1000:1 odds, we'd have to increase the number of trials. 10 out of 11 won't do. 11 out of 12 or 12 out of 13 might; I don't have my combinatoric hat on this morning.

Other factors lead me to conclude this, as well. The GSIC is being marketed as a tweak for high-enders. Your test is not relevant to the fine discriminations of tonal differences of extremely superb high quality audio; low quality BUT NOT SILLY audio (i. e. Mapleson cylinders) could suffice.

And as I and others have said before: what the rest of the system contains is irrelevant. What the GSIC device physically does to the disc or any other thing in the universe is irrelevant. All that is being tested is Michael's ability to discern the difference between a treated disc and an untreated disc. If he can demonstrate that he can do this, he wins a million dollars!

I do often hear remarkable beauty of sound from high end systems. I have no objections to the existence of the social phenomenon of "high end audio", nor to its practitioners.

And neither do we, as long as the components do what the manufacturers say they can do. Michael's CD player plays CDs. His speakers emit sound. His speaker wires carry electical impulses from his CD player to his speakers. None of this is in question; these components have been shown to work as advertised. But the GSIC device has not. Not only has it not been shown to work as advertised, there are reasons to believe that it cannot work as advertised. Michael thinks otherwise. Fine; the Challenge is here so he can put his money where his mouth is.

I am not biased against hi end audio; though I am biased against hi end nonsense, and am EXTREMELY biased against hi end fraud and exploitation.

Then we are on the same team.

Does this answer, and amplification of my biases, assist you in an informative manner?

Yes. Thank you.

 

[Bodhi Dharma Zen]

Forget the GSIC, if this test were about something that sounds different, like two sets of speakers, the claimant would be able to pass easily. The sound changes would be evident time after time.

I dont see the need for a complex methodology, at least for Wellfed to knowif he can or cant listen a difference. Two CDs, one individual (and an observer) "treat" one of them.

Another individual (and the same observer) will mark one, without knowing if it was "treated" or not.

This person can change both discs in the CDP, and the claimant will annotate CD1 or CD2, according to his perceptions.

If he can pick up, consistently, the difference between both CDs the GSIC works. What can be simplier than this?

 

[rwguinn]

PT--
The applicants don't really need your help in providing weasel-words and excuses pertaining to their failure. They do perfectly fine on their own.
But I am sure Wellfed appreciates the assist.

 

[Beleth]

In addition to the possibility of "gassy tubes" and improper room temperature invalidating the test, Piano Teacher failed to point out that the earth could spin off axis and go crashing into the moon, right in the middle of the procedure! That could happen you know. We want to make sure we cover every possible angle.

I was going to make a similar comment, but then I realized that PT was right. About the tubes, at least.

There are certain risks that need assessing. Failure of one of the components of Michael's system is certainly worth addressing, as is the chance that Michael might come down with a cold or an ear infection during the test. (I had a doozy of an ear infection in March. Clogged up my phone ear for almost a month. But I digress.)

It's better to deal with escape clauses ahead of time than to cackle at Michael's failure due to a technicality later.

Certain risks are worth taking into consideration. Maybe not down to the tube level, though. "If at any time during the test, a mishap or act of God significantly changes the characteristics of Michael's listening experience, the trial will be reset to the beginning, and all data collected up until that event will be discarded", say.

 

[TjW]

Re: Re: Response to Beleth about his personal test speculation

Not only has it not been shown to work as advertised, there are reasons to believe that it cannot work as advertised.
<substantial snippage>
[/B]

It might be worthwhile to reiterate that:
The GSIC claims to make a permanent change to the disc, not the player.
The GSIC claims to change the timing on the disc.
Since the disc does not contain the timing, only the data, it can't work as advertised.

Someone over on Audio Asylum has already done a bit-by-bit comparison of the same disc before and after GSIC treatment. Unsurprisingly, the data was identical.
I believe Wellfed accepted that the data were identical.

 

[alfaniner]

Re: Re: Response to Beleth about his personal test speculation

...
The reason Michael is held to a 10-out-of-10 standard is that it's a binary test (is it treated, yes or no), and that the standard for the challenge is to have a minimum of a 1000:1 chance of passing the test through luck. Nine binary choices don't cut it; that's only 512:1. Ten choices do cut it at 1024:1.

...

I learned something useful and interesting today. Thank you.

Certain risks are worth taking into consideration. Maybe not down to the tube level, though. "If at any time during the test, a mishap or act of God significantly changes the characteristics of Michael's listening experience, the trial will be reset to the beginning, and all data collected up until that event will be discarded", say.

I highly doubt that specific wording would make it into a protocol, but I understand your meaning.

 

[KRAMER]

Good One !!!

I have considered being my own observer...

That's a good one, and an entirely different paranormal claim. You'll have to send in a new application...especially since I haven't heard a peep from you since last week when you promised a new protocol within hours.

A most curious, deafening silence.

Oh, and by the way, Piano Teacher is who he claims to be.
Did anyone here really accuse him of being something else?
If so, I must have missed it.

 

[PianoTeacher]

Four or More Differing "test gestalts"

After my prior post, I had time to refine an arguably irreducible series of sets that would represent "Four Different Test Gestalts" (using the definition of test gestalt offered in my previous post: i. e., "how to simply describe a test for a Wellfed-like audio perception claim".)

Set No. 1: Dioigenes' set.

Concept to be verified = "black box no. 1". We need to know nothing about the black box. The test is summarized by him as being black/white "simple".

Set No. 2: Beleth's set:

Concept to be verified = 'audible differences between two CDs, one copied with a VERY lossy process' = "black box no. 2".

Set No. 3: Wellfed's set:

Concept to be verified = 'ill-defined generalities and convictions and conclusions, derived from sloppy test with fatal internal inconsistencies' = black box no. 3.

Set No. 4: PianoTeacher's suggested set:

Concept to be verified: 'narrowly defined "audible differences" determined by highly controlled narrow and repeatable test' = black box no. 4

------

This is the situation as I see it presently.

Diogenes' set 1 is not my preferred set 4. Set 1 is arguably too simple, and incorrectly reduced since we have to define something about what is being examined and tested, by Wellfed, in black box no. 1.

Beleth's set 2 is tautological since her lossy process is always audible to all potential test subjects.

Wellfed's set 3 is simply out of the question; it has a a forgone conclusion built in because of its internal logical inconsistencies and lack of controls, in conflict with the need to define limits as well as Wellfed's predictable loss of concentration.

PianoTeacher's set 4 is something that I am not going to "judge". I leave that to peer review. All I am prepared to do is (a) try to define it as I've done, and reduce it as given above; (b) offer it for falsification; and (c) explain that it is different from the other sets. My own intrinsic sense of this is that it MAY arguably be more scientifically practical.

OTHER SETS

Among further sets of test gestalts we may reduce and propose, there is one I've described before, which also intrigued one other poster:

Carrier Test Set No. 5: data comparison of two CDs, one treated, the other untreated, by means of an automated algorithm independent of human aesthetic judgment and perception

[The description below is complex because it has never been discussed before by me, nor introduced into this forum.]

Analogue Difference Set No. 6: PianoTeacher's old "analogue era" stand-by: subtracting the electrical analogue differences.

The treated CD has asserted "added value" and if true, must contain more information than the unreated CD.

For the purposes of simplicity, I will only use 1 pair of CDs in order to make this explanation intelligible. We will have "CD A" - treated; and "CD B" - untreated.

Baseline No. 1 -- equipment baseline -- is established that controls for equipment noise and jitter.

Baseline Nos. 2 -- CD silence -- is established for each CD pair: a median value of silence determined by examining the noise levels of disks CD A and CD B.

Using an audio matrix circuit, the difference signal between CD B and CD A, is extracted by means of subtraction. Electrically, this is expressed as:

primary difference = CD A - CD B.

That difference signal is then further treated by subtracting the Baseline No. 1 (equipment noise) and Baseline 2 (CD silence), from the difference signal derived by subtraction of CD B from CD A.

secondary difference(1) = primary difference - Baseline 1 . [equipment noise control]

secondary difference(2) = secondary difference(1) - Baseline 2. [actual audio difference signal, minus common equipment noise]

For simplicity, we then define secondary difference(2) to be "evidence".

Here's the sticky part: a judgment must be made about the amplitude of the difference signal, "evidence", to determine if it is "noise" or "signal" and whether it is random or has musical coherency; and if it is "significant" and sufficient to PROVE a difference between CD A and CD B.

However, if the test is well controlled using ONE control CD player and ONE set of electrical circuits and measurement instruments, and two CDs, controlling for limits of resolution, the "judgment" may be quantified, and argued against standards that may be agreed on.

As you can see, this protocol is hard to reduce to a satisfactory simplicity for achieving a "black/white" simple test; but it IS, electrically and realistically, with respect to human cognitions, able to reify a practical "difference". I doubt that JREF will bet a million bucks on this, because of the irremediable conflicts in debating "significant distinctions" in deciding the "content" of the evidence.

Set 6 is, however, the kind of test that "objective rationalist skeptical audio engineers" would like to do, in a defined context.

-----

There are other sets, too. Some of the ones I have described, above, tend to produce useful results; others produce no useful results; and I see no identical and comparable results in a meta-comparison.

I do hope that tends to give substance to my argument that Diogenes' set is improperly reduced. His "undefined black box 1" -- irrespective of either audio or Wellfed's neural response and reliability -- has no significance in a test of Wellfed's claim.

PianoTeacher

 

[Wellfed]

What I am after is the replication of my normal listening environment. I know that this isn't going to happen to perfection, but I wish to avoid any difference in my environment that is not a necessity in protecting the interests of JREF.

ABX testing is not part of my normal listening environment. AB testing is seldom part of my normal listening methodology.

 

[BPSCG]

I've been reading this thread in fits and starts, and was curious about the claims made for the GSIC by the manufacturer. So I went to their website to see what they had to say and found "customer comments.":

I've used the IC 5xs and heard a clear, consistent improvement each time -- reduction of brashness and increased sense of solidity to the sound...

and

...[t]he thing that I particularly notice about the GSIC-30 is that the sound is more "musical." Also, there is less of a "digital edge" to the music... All this makes listening a more enjoyable experience.

and

... [t]he following seems consistent: the window is much cleaner, so if the CD is well-recorded the improvement is significant, the stage much larger, deeper and wider, and the music comes well out into the room. The pace, rhythm, and flow of the music is much better (especially appreciated with large-scale music).

and

I had the opportunity to try the intelligent chip on 3 CDs last night and another this morning....On all of the first 3 discs the improvement was quite obvious. I found that the full effect didn't set in until a few seconds after treatment -- and in all 3 cases the sound became much more relaxed and a sense of space and dimensionality became more apparent. The best way I can describe it is that the "decay" of the sound is greatly improved. Notes flow from one to the next instead of "start-and-stop." The forth disc didn't fare as well as the others; it is an older CD that obviously wasn't recorded or mastered as well as the others, so there's probably less detail to extract from this disc.

I listened to entire discs after treatment instead of just a couple of tracks for demonstration purposes. This is, I believe, the greatest testament to the effectiveness of the chip. I treated one disc this morning by placing the chip beneath the player. It worked that way too!

I love those last two sentences. You'd think putting the chip under your CD player would degrade the sound, but I guess I don't understand physics and acoustics as well as I should...

Just a few rhetorical questions:

If the difference in sound is as obvious as the above correspondents rave, then what is paranormal about the claim to be able to tell the difference between a treated and an untreated disc?

If everyone can hear the difference, then this is not a paranormal claim.

If nobody - including Mr. Anda - can hear the difference, then the thing is a fraud.

And if only people with years of training in listening and excellent hearing can tell the difference in a double-blind test, then why would anyone over the age of about 35, whose hearing is likely not nearly as good as it was when he was 5, want to waste his money on the thing?

 

[Wellfed]

Re: Good One !!!

That's a good one, and an entirely different paranormal claim. You'll have to send in a new application...especially since I haven't heard a peep from you since last week when you promised a new protocol within hours.

A most curious, deafening silence.

...

I didn't think you'd get it, and yes, it IS hard to respect dumbness. Let's just pick this up again in October, perhaps you will have grown up by then.

 

[Beleth]

ABX testing is not part of my normal listening environment. AB testing is seldom part of my normal listening methodology.

This sounds eminintly fair and reasonable to me. Besides, it's an implied part of your claim - that you can hear the difference on your current setup.

If you aren't claiming that you can hear the difference through an ABX switch, I see no reason to force you to do so. But then of course I have sod all to do with the official negotiation; I'm just being a Buttinski.

 

[Z]

Re: Re: Re: Response to Beleth about his personal test speculation

It might be worthwhile to reiterate that:
The GSIC claims to make a permanent change to the disc, not the player.
The GSIC claims to change the timing on the disc.
Since the disc does not contain the timing, only the data, it can't work as advertised.

Someone over on Audio Asylum has already done a bit-by-bit comparison of the same disc before and after GSIC treatment. Unsurprisingly, the data was identical.
I believe Wellfed accepted that the data were identical.

If the disc remains unchanged, what is there to test? "I can hear a difference when I wave fairy dust in the same room as my 'Imelda Marcos Sings the Blues' CD - The bass becomes more robust, and the rear acoustic arena expands in space!"

Yeah, right.

There's no change - so anything you hear that's different is called wishful thinking. So what are we testing, again? Nothing, apparently. Oh, no, wait, we're not testing the chip at all - we KNOW that doesn't work. We're just testing your psychic skills at determining if this thing was sitting over the disc at some time in the past.

Yeesh - Kramer, you're embarrassing the JREF if you let this farce continue much longer. Either he proposes a simplified protocol with none of this ego-feeding nonsense to test his psychic powers, or case closed....

Man, too bad My opinions don't mean squat, huh? Oh well...

 

[Z]

What I am after is the replication of my normal listening environment. I know that this isn't going to happen to perfection, but I wish to avoid any difference in my environment that is not a necessity in protecting the interests of JREF.

ABX testing is not part of my normal listening environment. AB testing is seldom part of my normal listening methodology.

So you can't tell a difference if a disc is played in an environment other than your own normal listening environment? Is the same true of other slight variations in discs?

If someone walks in with a boombox and plays a song, first off an album, and then off a CD-R, I can usually tell a difference, regardless of the capabilities of the box - unless, of course, it's an extremely crappy box. But you're claiming that you have to have, evidently, perfect, perfect listening conditions to tell the difference between an untreated and a treated CD.

Man, why would anyone want to buy such a thing?

Wellfed, man, you're embarrassing the Golden Sound folks pretty badly right now. I imagine, after this debacle, they'll want very little to do with your testimonies.

"After carefully manipulating all conditions to absolute optimal perfection, I was able to tell a slight difference in the treated and untreated disc! What amazing results!! Go right out and spend hundreds on these chips to optimize your entire collection! Just remember to mark the cases, though - there's no earthly way to tell a treated disc from an untreated disc!!! Oh, I mean, except under the perfect listening conditions for you..."

Glowing endorsement, indeed.

 

[Skeptical Greg]

Set No. 2: Beleth's set:

Concept to be verified = 'audible differences between two CDs, one copied with a VERY lossy process' = "black box no.

Whoa Nellie!


Will you explain the losses involved when copying CD's?

 

[alfaniner]

Re: Re: Good One !!!

I didn't think you'd get it, and yes, it IS hard to respect dumbness. Let's just pick this up again in October, perhaps you will have grown up by then.

So, I guess tonight's double-blind self test is a no go then...