3DSonics
away working hard on "it"
Hi,
Which illustrates (again) the invalidity of poorly implemented tests, however it does not proove thatno differences exist either.
That is the problem with the religious mode of thinking. I prefer to know, not to believe. I have devised my own largely blind tests and carried them out to my personal satisfaction. As a result I know what makes differences to me and some others and what does not and more importantly I get a good idea what people actually like.
Absolutely. But not should the identity be obscured, BUT ANY HINT OF WHAT BEING TESTED must be obscured. And you still need to find people who can listen analytical enough and who will actually seriously attempt to identify differences and you need quite a few of them and ways to convince them to take part.
There where some early Blind Tests conducted (and reported in an AES Paper) which had actually set a sensible level of trials and significance (requiring 5 Trials and and a score of 4/5 for suggesting the audibility). The tests found audible differences in speakercables with thus a chance of only 20% that the results where by chance. A pretty large room of people gave results, at least around 50 I believe, mot got 4/5 or better.
The emperimentor was ridiculed and his statistics where torn apart, inaccuratly btw, and it was pointed out to him that he needed to analyse to a higher significance. So repeats experiments with 9/10 (.1 significance). Still the results showed a disturbing tendency to suggest audible differences in better than 1/3rd of the participants, so he gets some more heat. Finally, when demanding 19/20 (.05 Significance) only very few people got positives, these where branded "lucky coins" and discarded.
So, finally there was proof that all speaker cables sound the same.
Except there isn't. Actually, if you took from each test ste not the single individuals data but the whole set, you could easily suggest that analysed to .05 Significance in ALL tests differences between the cables where heard, even if virtually no single individual could tell in the tests with the larger number of trials. What does it suggest?
There are lies, damn lies, godamn lies, cable advertisments and finally ABX Test statistics.
I am well tired of hearing the same old harping on from people who merely wish to believe, instead of attempting to understand what is going on.
May I suggest you wake up, leave your stupified state of belief and look at things as they are in reality?
That depends on many factors, including (reverse) psychology, statistical evaluation methodes and test setup. I notice that in not a single case of published ABX/DB Tests relating "Audio Voodoo" was the test setup and listener group benchmarked against known audible (but subtle) changes and thus the test itself validated.
Neither Blind nor ABX Tests are inherently flawed. However, given that the vast majority of people who advocate and seemingly all who actually conduct them try to use them with evangelical zeal in support of the "All equipment sounds the same" hyphothesis with the coroloary "Everyone who sells gear that costs more than the cheapest stuff from major electronic companies and who sells tweaks, cables etc. et all is a Charlatan and Fraudster and should be jailed" it is hardly surprising they get the results they want.
Now their behaviour by definition places them right up with the "Charlatans" all too common on the other side of the argument.
Hence I usually suggest to people simple experiments that can be done even blind and to stop listening to the Charlatans of either side. Make up your own mind, make your own experiments and draw your own conclusions, become at least the leader of yourself instead of sheepishly running after one group or another in blind believe, which your senses will usually and readily confirm.
Ciao T
notaclue said:
Which illustrates (again) the invalidity of poorly implemented tests, however it does not proove thatno differences exist either.
notaclue said:
That is the problem with the religious mode of thinking. I prefer to know, not to believe. I have devised my own largely blind tests and carried them out to my personal satisfaction. As a result I know what makes differences to me and some others and what does not and more importantly I get a good idea what people actually like.
notaclue said:
Absolutely. But not should the identity be obscured, BUT ANY HINT OF WHAT BEING TESTED must be obscured. And you still need to find people who can listen analytical enough and who will actually seriously attempt to identify differences and you need quite a few of them and ways to convince them to take part.
notaclue said:
There where some early Blind Tests conducted (and reported in an AES Paper
The emperimentor was ridiculed and his statistics where torn apart, inaccuratly btw, and it was pointed out to him that he needed to analyse to a higher significance. So repeats experiments with 9/10 (.1 significance). Still the results showed a disturbing tendency to suggest audible differences in better than 1/3rd of the participants, so he gets some more heat. Finally, when demanding 19/20 (.05 Significance) only very few people got positives, these where branded "lucky coins" and discarded.
So, finally there was proof that all speaker cables sound the same.
Except there isn't. Actually, if you took from each test ste not the single individuals data but the whole set, you could easily suggest that analysed to .05 Significance in ALL tests differences between the cables where heard, even if virtually no single individual could tell in the tests with the larger number of trials. What does it suggest?
There are lies, damn lies, godamn lies, cable advertisments and finally ABX Test statistics.
I am well tired of hearing the same old harping on from people who merely wish to believe, instead of attempting to understand what is going on.
May I suggest you wake up, leave your stupified state of belief and look at things as they are in reality?
notaclue said:
That depends on many factors, including (reverse) psychology, statistical evaluation methodes and test setup. I notice that in not a single case of published ABX/DB Tests relating "Audio Voodoo" was the test setup and listener group benchmarked against known audible (but subtle) changes and thus the test itself validated.
notaclue said:
Neither Blind nor ABX Tests are inherently flawed. However, given that the vast majority of people who advocate and seemingly all who actually conduct them try to use them with evangelical zeal in support of the "All equipment sounds the same" hyphothesis with the coroloary "Everyone who sells gear that costs more than the cheapest stuff from major electronic companies and who sells tweaks, cables etc. et all is a Charlatan and Fraudster and should be jailed" it is hardly surprising they get the results they want.
Now their behaviour by definition places them right up with the "Charlatans" all too common on the other side of the argument.
Hence I usually suggest to people simple experiments that can be done even blind and to stop listening to the Charlatans of either side. Make up your own mind, make your own experiments and draw your own conclusions, become at least the leader of yourself instead of sheepishly running after one group or another in blind believe, which your senses will usually and readily confirm.
Ciao T
