The best plug in filter ever!

@dogberry

For a double-blind test to be statistically valid, there must be at least 25-30 subjects in the test to assume a standard distribution curve so that you can calculate the reliability of the data. One or two people doing a double-blind test is just anecdotal.

Back in 1995 I bought some very lame tweaks that I thought were worthless, and still do think are worthless. I would listen and think, "Am I hearing an improvement? Well, I paid for the thing so I might as well use it because maybe small changes add up." I think that was just optimism. They were not doing anything of real value for my system.

I think if you meet someone who is happy with everything they try, that doesn’t necessarily mean that everything they try is good but could be indicative that they would like anything that they pay money to receive, as long as someone they trust has told them it will be good. This is bias. I love the quote from Mark Twain that "It’s easier to con someone than to convince them that they’ve been conned."

BUT

Logically, this does not preclude any of their tweaks from actually helping. If we are testing diamonds with a broken diamond tester that will say everything tested is genuine, and it tells us that 20 out of 20 tested stones are genuine, we can dismiss the diamond tester, but we cannot logically assume that the stones are all rubbish and that they should all be thrown in the trash.

Likewise, if someone is testing numerous items and everything tested seems to be worthless snake oil, a critical thinker would question this as well. More so someone who does not even listen to the items and proclaims them all to be snake oil without having any statistically valid evidence to support such a claim. Scientists said that reports of a miracle cure that we now know as vitamin C was "witch doctor superstition" for around 200 years before the scientific community had enough knowledge to understand what all of the anecdotal reports were saying.

But statistical analysis practices aside, if one listens to a supposed improvement in one’s system and does not hear the sort of clear (undeniable) audible difference that one experiences with adding acoustic treatments,... and feels the need to test the "improvement" with a double-blind tests,... then the difference was probably negligible and not worth the trouble.

Since a statistically valid scientific test is probably somewhat impractical for most of us, I suggest that if you listen to something and you even ever-so-slightly feel the need to test, just send it back promptly and get a refund.

That being said, I have purchased products, listened, and then thought, "Pfffft. Lame." I have also purchased other products (secretly without telling my wife) that made me say, "Whoa! That was an improvement!" and then my wife, with no knowledge of anything different, walked through and commented, "Your system sounds better what did you do?"

So, I don’t have any bias about tweaks other than to say 1) I expect some of them to be pure marketing rubbish, 2) I expect most of them to be of negligible value of the sort that would need an A/B test to determine if they are an improvement, in which case I would say the aren’t worth it, and 3) I am open to accept that some of them will make an immediate and undeniable improvement that I can enjoy.

I just bought a Puros, and I will report back to the group with my unbiased but anecdotal critique. 

(The other gear in my system includes: Linn LP12, Levinson No.31 transport on PS audio power conditioner fed by Synergistic Research foundation AC cords, Naim Uniti Atom HE DAC & streamer, SR Foundation XLR interconnects & speaker cables,  Levinson 33H monoblock amps, Wilson Maxx Speakers, Nanoflo on most all connections, seven of Krissy's various bizarre devices, 2 bybee AC devices, Townshend pods, Shunyata Alpha NR & Synergistic foundation AC cables, Shunyata Denali power conditioner, Shunyata venom plug-in filter cap, (4) iFi iPurifiers, upgraded fuses, isolated ground, dedicated audio line from breaker panel & isolation transformer, star grounding for all components on a SMD copper grounding block, Skyline diffusors on ceiling in front of listening position, 5 Tube traps, about a dozen large acoustic panels around the room, more home-made diffusors, and I'm sure I probably forgot to list a lot of my gear.)

(NOTE: I'm open to any system recommendations as well!)

Happy listening to all

Halz

@ozzy 

"C'mon man!" LOL 

Yeah this is probably a bad idea... my wife is going to ask me questions about how much various things in my system cost. 

I ran it in G*Power, and for my test to be validly predictive of the entire population of listeners, it would need to include at least 89 men and women who were randomly selected. Otherwise, my test results are only going to prove that my wife can (or cannot) clearly hear the difference that it is making in my system. She will also have to listen to at least 30 different sets for me to be able to calculate the correlation strength. But again, this is all being done for fun. 

A. If my testing says that my wife could not consistently detect any difference in a double-blind test, then all the Puron nay-sayers can post pictures of snake oil in here and all of us who bought will just have to endure it. 

B. If my testing reveals that my wife can consistently detect differences with the Puron in the double-blind test, then the nay-sayers will all have to wear paper cone hats for a day and go and buy a Puron in penance. 

@dogberry

Yeah I think it does depend on the size of the effect. I think it pretty much all boils down to exactly that. So my point is that if the effect is small enough that you even feel the need to test it,... probably not worth wasting any more of your time on it.

Any work published in any scientific journal has to pass muster when it comes to statistical validity.

BUT

If home testing is fun & enjoyable, sure why not! : )

If I like the Puros, I’ll double-blind test it with my wife listening, since she is biased against me spending money and her hearing is not as rolled off as mine. If my tests support the product... it’s still anecdotal but I think you’re right it has less chance of being biased.

PURON DOUBLE BLIND A/B TESTS

I got my Puron in the mail yesterday, and did the first tests. I asked my wife to sit down and listen to tell me what she thinks. I started off without it, and played a song she knows pretty well. I turned my back so she could not see whether I was  adding or removing the Puron, and did my best to prevent her from being able to tell if I was adding or removing it. I played 30 seconds without it, and she said, "OK." Then I played 30 seconds again without it, and she said, "Yeah, OK." Then I plugged it in cold for the first time, with no burn-in at all, and she immediately said, "Oh, Oh! That one is better." In all, with my randomized A or B tests, she consistently picked the Puron for better sound, identifying 5/5 correctly in this session. She commented that my system sounds "more like it is supposed to sound," and "more real."

Guessing five out of five correctly in a double-blind test is not enough to be statistically significant. We will have to do four more sets for a total of 25 samples minimum to get to statistical certainty. Also, the findings studying one person on one stereo are not enough to make a predictive model about the entire population, each with different stereos. However, she only had a 3.125% chance of guessing correctly five out of five times. But, that can happen, so... more tests needed before I can confirm with sufficient certainty.

I let it burn-in for a day and I did my own first listening tonight and could immediately tell that my system has a different sound. I could hear a lot more information, a lot more easily. This was not a subtle change at all. I liked it almost as soon as I sat down to begin listening. It was easier to follow the melody, and easier to hear the instruments.

So, statistically speaking, my wife can say that she is 96.875% certain that she prefers the sound of the Puron in my system. I can say that I like it too. I will report later what the rest of her listening tests show, but I think I will probably text Mark right now and order a second one. Definately worth $250.