• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

GSIC Blinded Testing

BPSCG

Penultimate Amazing
Joined
Mar 27, 2002
Messages
17,539
I promised to have the first results of my GSIC testing up by August 27. I'm a little late; I was running tests on my first subject last night, the charming and intelligent Mrs. BPSCG, when my sister and her family showed up at our house earlier than we expected and I had to interrupt the test. I hope to complete it this morning after Mrs. BPSCG has finished reading the Sunday Washington Pest.

For those who are interested in the protocols, such as they are, for my GSIC testing, here they are:

Album: “Gordon Lightfoot: Complete Greatest Hits”
Label: Warner Bros./Rhino
Catalogue Number: R2 78287
Two identical copies purchased 8/26/2005
I treated one copy of the album with the GSIC chip, following exactly the instructions that came with the GSIC. I performed this on our single-CD player, because the instructions say to put the GSIC on top of the CD player, above the transport, and the position of the transport appeared to me to be more apparent on the single-CD player than on the 200-CD jukebox being used to conduct the tests.

I then marked the GSIC-treated copy, clearly, but unobtrusively, on the label side of the disc.

I then generated ten series of random numbers in a spreadsheet by dividing one three or four-digit number by another three or four-digit number and carrying the result out to ten decimal places. These numbers are to be used to determine the sequence in which the two CDs are to be played for comparison purposes.

I put both CDs back into their jewel cases, gave the jewel cases to Mrs. BPSCG, and told her to insert one of the CD's into slot #197 in the jukebox, and the other into slot #198. I did not tell Mrs. BPSCG how to identify the GSIC-treated CD, and unless she compared the two CD labels side-by-side, it is unlikely she would have noticed the mark I used to distinguish them. She then shut the jukebox door, so I could not see which CD was in which slot.

I then had her sit in a chair facing directly away from the stereo speakers and the CD jukebox. I explained to her what she should listen for, by reading from the instructions that came with the GSIC, quoting, "less congestion, more information, greater dynamic range and more air." I did not explain what any of these terms meant, except for dynamic range, which I described as the apparent difference between the softest and the loudest sounds. I did not ask her to identify the GSIC-treated chip, but simply to state which one she thought sounded better. I said the differences might be quite subtle, and she agreed to give her best guess, even if she didn't think she could tell any difference.

I then selected one of the ten groups of ten random numbers. If the first digit in the number was odd, I would then play the first thirty seconds of the first track of the disc in jukebox slot #197; if the number was even, I played the corresponding thirty seconds of the CD in slot #198.

After playing the first disc, I would then play the first corresponding thirty seconds of the other disc for comparison purposes, then asked Mrs. BPSCG which one she thought sounded better. I recorded the result, then had her plug her ears so she could not hear if I was punching up a different CD for the next trial. When I had the starting CD in place for the next trial, I tapped her on the shoulder and she unplugged her ears, and we went on to trial #2.

We had completed seven of the ten trials when the doorbell rang. I hope to finish Mrs. BPSCG's trials this morning. Mrs. BPSCG says we could use a million bucks.
 
GSIC Test Results #1

Results of the first test (see OP for protocol followed):

Mrs. BPSCG was unable to distinguish between the treated CD and the unttreated one. In the ten trials, she selected the treated CD five times, and the untreated one five times, as sounding better.

After the test, I asked her if she ever thought she'd heard a difference between the CDs, and she said for the most part, she could not tell any difference at all. I asked her if she ever heard a difference. She was doubtful; if she'd heard any difference at all, it was very slight.

In any case, if she'd heard, or thought she'd heard, any difference, she was unable to consistently distinguish which CD sounded better.

Note 1: As we were finishing the last trial, my two nephews and the girlfriend of the older one stopped by. I explained to them what we were doing, and they asked to be tested. They didn't have time to spend the half-hour or so it would have taken to conduct a complete test, but I did allow them to try one very uncontrolled trial, where I did not block their view of the CD player and where they spoke their guesses aloud after the trial, rather than writing them down for later review. They all agreed that CD #197 sounded better. The boys said it had better bass, while the girl said it sounded clearer. Since they are all in their late teens/early 20's, and Mrs. BPSCG and I are (ahem) older, I'd be interested to hear what happens when they take a complete, controlled DBT.

Note 2: I asked Mrs. BPSCG after the test if she had noticed, when she was putting the CDs into the jukebox, any difference in the labels. She told me she had not, that she had avoided looking at the labels as much as possible when inserting them, knowing I had marked one of them.

Note 3: I couldn't hear any difference, but then again, a) I wasn't listening carefully, was paying more attention to watching the CD timer, and, b) my hearing isn't very good. But I can say for certain that if there was any improvement in the sound, I didn't find it readily detectable.

Comments?

*Revealed after the test as having been the one in jukebox slot #197.
 
Did two more tests today. Tests conducted separately.

Test 1. Subject Pat , female, early 50's, claims good hearing, "but not as good as it used to be." Used the same protocol as described in the OP, using the same album.

In seven of the ten repetitions of the test, Pat identified the GSIC-treated CD as the one whose sound she preferred. She said she thought one of the CDs sounded "sharper", though she wasn't sure. She thought she detected a change in syllables, and that one CD sounded louder, but she didn't know why. Note that for all that, she said the CD that sounded louder and sharper was not the one she preferred.

Test 2: Subject Brian, male, mid-50s. I know his hearing isn't very good because his wife is always complaining he never listens to her :D In six of the ten repetitions of the test, Brian identified the GSIC-treated CD as the one whose sound he preferred. He said after the first test, he was pretty sure of his choice, but volunteered that it might well have been because he'd expected to hear a difference. On subsequent tests, he said if he heard any difference at all it was a difference of "nuance."

Brian's results are clearly within the boundaries of chance; Pat's are less so, but I would think that if the GSIC did what its purveyors claimed, a listener with decent ears would be able to pick it out nine times out of ten or better. The odds are certainly against randomly guessing correctly seven times out of ten, but not at all implausible - you wouldn't think it remarkable if you threw ten pennies in the air and seven of them landed heads.
 
not me

I thought you wanted to see the damn thing tested...

It was, by LostAngeles. She didn't offer us any photos of the session in LA, unfortunately. That would have been nice.

I'm perfectly satisfied it is a crock of ◊◊◊◊.
Have fun... who cares. Not me.
 
Re: not me

webfusion said:
It was, by LostAngeles. She didn't offer us any photos of the session in LA, unfortunately. That would have been nice.

I'm perfectly satisfied it is a crock of ◊◊◊◊.
Have fun... who cares. Not me.

I was satisfied even with the binary tests.... ... ... The CD is identical or it is not. One of the two.

What annoys me - a little - about testing is that a GSIC is being purchased from the 'manufacturer' (and I use that term losely). And then purchasing the higher number of "uses" one for the number of tests.... arrrrrrrrggggh....

Brain pain.
 
Wouldn't it be worth running a parallel test?

Play two identical CD's (or even the same one twice) and ask each of your listeners if they can detect any difference.
This is designed to see if the human mind finds 'differences' even when there aren't any...
 
BPSCG - no real comment as it is as all as I would expect it to be but I do find it interesting don’t assume my lack of significant feedback is that your results aren't interesting.

(Of course we can’t trust you, for all we know you could be league with the big record companies to deny us users this astonishing device.)
 
glee said:
Wouldn't it be worth running a parallel test?

Play two identical CD's (or even the same one twice) and ask each of your listeners if they can detect any difference.
This is designed to see if the human mind finds 'differences' even when there aren't any...
In theory, I suppose so. In practice... Between the time it takes to explain the process to the testee and the actual testing, the whole thing takes about a half hour. About twenty minutes of that half hour is taken up in listening to the first thirty seconds or so of the same song, twenty times.

I prefer not to make my subjects sick of the whole thing. But it might be a good test to run separately, on separate subjects.

If you want to send me a CD, I can "treat" it for you (quotation marks used advisedly, since I'm 99.99999% certain the "treatment" does nothing) and send it back so you can do the parallel test if you like; my GSIC still allegedly has nine uses left, since I've only treated one disk so far. Your proposed test might very well be the more interesting one, since just about everyone here is convinced the thing is a hoax, whereas you'd be testing not the device, but your subjects - the hypothesis that expectation can influence perception. PM me if you're interested.
 
Darat said:
BPSCG - no real comment as it is as all as I would expect it to be but I do find it interesting don’t assume my lack of significant feedback is that your results aren't interesting.

(Of course we can’t trust you, for all we know you could be league with the big record companies to deny us users this astonishing device.)
I have a more devious plan than that. I'm searching for that one guy with the "golden ears" who can actually pass. Then he or she will take the JREF challenge, and we will win the million bucks..

Hey, here's a question: Suppose I found someone who could pass the test, and he wins the million. Is there any rule in the challenge that says the same person can't take the challenge over and over again? Yes, it's extraordinarily unlikely anyone could win the challenge legitimately, but what if someone found a way to cheat that JREF couldn't detect and couldn't figure out?
 
BPSCG said:

Hey, here's a question: Suppose I found someone who could pass the test, and he wins the million. Is there any rule in the challenge that says the same person can't take the challenge over and over again?

I suspect that when/if the JREF challenge was won, the JREF will cease to offer the challenge. Put really bluntly, they will no longer have the money to pay out. Even if they can find another angel to support the prize, they will likely adjust the rules somewhat -- and, of course, the same claim would no longer be considered "paranormal" as it can now be experimentally confirmed.
 
Re: not me

webfusion said:
Have fun... who cares. Not me.
Then why are you the second most prolific poster on this thread, after me? :p
 
BPSCG said:
I have a more devious plan than that. I'm searching for that one guy with the "golden ears" who can actually pass. Then he or she will take the JREF challenge, and we will win the million bucks..

Hey, here's a question: Suppose I found someone who could pass the test, and he wins the million. Is there any rule in the challenge that says the same person can't take the challenge over and over again? Yes, it's extraordinarily unlikely anyone could win the challenge legitimately, but what if someone found a way to cheat that JREF couldn't detect and couldn't figure out?

Don’t think there is any rule but since the million dollars would have gone it may be something like "The JREF $2.27 and a half chewed pencil from Randi's desk Challenge".
 
Re: not me

webfusion said:
It was, by LostAngeles. She didn't offer us any photos of the session in LA, unfortunately. That would have been nice.

I'm perfectly satisfied it is a crock of ◊◊◊◊.
Have fun... who cares. Not me.

I haven't because IIG took and therefore has the photos (and video!)

Nevertheless, I totally think it's cool that BPSCG and his alphabet pals are testing it out too. I'm pretty much satisified on it, sadly.

I'm silently following the thread BPSCG. Have fun, really.
 
Test 1. Subject Pat , female, early 50's, claims good hearing, "but not as good as it used to be." Used the same protocol as described in the OP, using the same album.

In seven of the ten repetitions of the test, Pat identified the GSIC-treated CD as the one whose sound she preferred. She said she thought one of the CDs sounded "sharper", though she wasn't sure. She thought she detected a change in syllables, and that one CD sounded louder, but she didn't know why. Note that for all that, she said the CD that sounded louder and sharper was not the one she preferred.
I have a more devious plan than that. I'm searching for that one guy with the "golden ears" who can actually pass. Then he or she will take the JREF challenge, and we will win the million bucks..

BPSCG

I think you are on to something. 7 out of 10 identifications has a probaility of 0.172. Repeat this on her next set of 10 and it goes down to 0.029, enough for a Scientific paper (p < 0.05). And just two more repetitions (without failing of course) and she is at 0.0008, good enough for a preliminary test!

IXP
 

Back
Top Bottom