The isolation pad may not be improving the digital data from the CD itself, but all of the other components (power supply, etc) between the laser and the audio output.

Which I'd grant, if the components were something like a tube amplifier. I could see how transient vibrations could affect a vacuum tube. But everything inside a CD player is solid-state.

I could also see how mechanical isolation could prevent ground loops between pieces of equipment. Or how additional physical distance might prevent electromagnetic interference. But in both of those cases, there would have to be something wrong with components in order for an isolation system to be of any help at all.

I also found a url of a review of some isolation pads

I read those, and in neither case did the reviewer do a blind comparison.

Here is a fairly straightforward article on the need for blind testing in subjective tests. Without a blind test, you're opening yourself up to the possibility of subjective validation, wishful/selective thinking, etc. No matter how hard you try to be objective during such a test, if you know the expected outcome then you're going to have a tendency to fool yourself into thinking the evidence is there.

The soundstage of an audio recording is a very subtle thing. Almost as subtle as the dowsing example given in that article. The difference is that there are known factors which affect the soundstage of an audio recording and they can be reproduced in controlled tests.

It would be fairly easy to design a protocol that would blind-test the isolation hypothesis:

1) The protocol requires two people: an audiophile willing to take the test, and a tester willing to administer the test. (This is a simplified protocol which assumes both the tester and the testee want to be honest about it. If things were more serious, you'd need to verify that the tester and the testee aren't in cahoots, you'd need independent judges, you'd need to make it double-blind, etc.)

2) Agree on a reference recording- a specific passage of a specific recording where the soundstage is assumed to improve with isolation. Make sure the audiophile agrees that the passage is a correct one with which to determine the soundstage. It should be a fairly short passage, since you're going to have to run through it a bunch of times. I'd say no more than 30 seconds or so, unless the audiophile says he really needs a longer reference passage. Ideally, it would help if there were an easy way to cue up the exact beginning and end of the passage on the test equipment.

3) Visually isolate the equipment rack from the audiophile. Such as placing a partition between the audiophile and the rack. Make sure no mirrors, glass, chrome reflections, or shadows can give away what the tester is doing behind the partition.

4) Test the speakers and the listening position to make sure the audiophile agrees that they accurately represent the soundstage.

5) Agree on the number of tests to run. I know little about statistics, but for argument's sake, let's say that thirty runs would be enough. You would also agree on a number of "hits" that would constitute a positive outcome. If the difference between padding and not padding is as striking and unmistakable as you say it is, there should be an incredibly high number of hits: I'd say at least 25 out of 30. I don't know the formula for determining what would be expected due to chance, though, you'd have to look that up in a statistic textbook. I'd guess that anything under 22 would be chance, and 23-24 would be inconclusive.

6) The tester and the audiophile both have a piece of paper and a pen. The tester stands behind the partition where the audiophile cannot see. The tester calls out the number of which test run it's going to be, for instance, "This will be test number five". Both the tester and the audiophile write down the number. From this point on, the tester does not say a word or make any other noises.

7) The tester flips a coin. Heads=padded, Tails=unpadded. It is important that this coin flip happen AFTER the tester calls out the test number so that no voice inflection cues are given. Before doing anything else, the tester writes down next to the number whether it's supposed to be padded or unpadded.

8) Regardless of the outcome, the tester fidgets with the equipment so that the audiophile can't tell whether the tester is adding or removing the pad. (Here is a place where the tester could give unconscious clues if they weren't careful or if there were some kind of collusion between tester and testee- I don't know how to solve this one without significantly complicating the protocol.) The tester either places the pad under the CD player or removes it, as dictated by the coin flip. Absolutely no other changes are made to the equipment except the removal or addition of the isolation pad. Not even volume changes.

9) Without saying another word, the tester plays the test passage. After the test passage is done playing, the audiophile writes down his guess next to the number on his piece of paper. He calls out to the tester "Ready" once he's done recording his answer.

10) The tester calls out the next test number, flips the coin, etc. etc. for thirty runs.

11) At the end of thirty runs, the two pieces of paper are compared. Total the number of "hits" with "misses".


Now, to get this thread back on topic: A nearly identical protocol could be used for testing the Empeg's sound stage in comparison with a CD player. You would have to be extremely careful to compensate for differences in equalization between the two units. For instance, when I hooked up my Empeg to my home stereo, it's EQ curve when "flat" was significantly different than my CD player's. You'd probably need some frequency-test equipment to sort that out. But once it was done, then the test itself would be even easier since you could just use an A/B switchbox, or two different input channels on the amplifier.


___________
Tony Fabris