Strangely enough the conditions described above are the same as I advocate...
double blind is not needed if the tester and listener are different persons (as explained).
Trained listeners (wow.. I don't think I brought that up either)
quiet listening room (an eye opener)
loudness matching (match levels < 0.1dB is not loudness matching ?)
well recorded music (wow in this thread and others it was one of the reasons why I said the test was flawed (using vinyl rip))
multiple comparisons (Now why didn't I think of that... statistical relevance)
Ah wait I see the catch this was written by a misguided man who believes in these tests.. dammit I almost fell for it)
I have conducted a few BT's that most certainly apply to at least the requirements you state and are stated above.
yes mostly about different opamps, powre amps, pre-amps, output resistances, capacitors, topologies, CD players, cables even.
I can assure you (like Szoze also can) that you do not have to do a (D)BT evertime you change some small thing you can clearly hear.
It's not needed at all.
Anyway.
It is completely pointless for me to listen to files that are the same in waveform description (1 and 0 content) simply because:
A: my gear is not good enough (really) I do not use computers for audio playback.
B: My expectation bias prevents me from hearing differences that cannot be there theoretically.
C: It has been proven I cannot discern (listened to Alex's files that have been verified to sound different but are the same in contents.
Like the J-Play test and the Italian one it is a waste of time.
I am not the one that must take that test as I fail in at least 3 of the criterea (bias, equipment, hearing capabilities) to test files that are the same in all aspects (except perhaps in an unknown to science way)
The only way to test the claims is for you and Alex to put the money where the mouth is and take the test (using Alex's stick and the test controled by an independent controller, i.e. not me) by clearly outlined rules.
I have stated those rules clearly in the J-Play thread.
Alex (or John) prepares a MS with 11 'bad' files (named so they are clear what they are) and 11 'good' files stored on it so we know there will be no degradation sending files or they are copied with undeserving computers.
The stick is sent to the 'controller' who renames 20 files and leaves 2 files un-renamed for reference. The controller MUST be unquestioned in honesty and have no perticualr affiliation with either party and need to be agreed upon.
He would have to know a bit about computers to rename, delete and change property boxes without 'messing'up the files.
The files need to be identical in content and checksum in THIS case.
A test where files are altered in contents cannot be tested this way...
There is to be no copying on the stick or from the stick.
When the stick has arrived the controller flips a coin which decides testfile 1 will become file A or B and renames one of the A or B files 'testfile 1'
next the controller flips a coin again which decides test 2 will become file A or B and renames one of the A or B files 'testfile 2'
All the way to 'testfile 10'
The 10 files that remain are simply deleted.
The property boxes are cleared or set similar (even those of the control files)
Only file 1 to 10 remain + the 2 original files for reference (to check whether the files still differ and as reference)
These can be listened to in all comfort with all the time you need.
It is even allowed various listeners can communicate about their findings before the submit their entry to the controller.
Not post publically to avoid awkward situations or if confident enough post them BEFORE the results are in.
Because of the test being blind and the files being identical (yet differ magically) they can be listened to, analysed, scrutinised e.t.c. and because of statistical relevance (x out of 10 in random order) the test will be more meaningfull to biased and ignorant EE's that are not in touch with the real world and do not know how to test and believe in measurements only (namely me and the entire scientific community you have been batteling all this time)
IF the differences are SO obvious and the files appear the same you guys (and others that can positively identify those files) should really have NO trouble identifying all files.
9 or 10 out of 10 would be enough to convince the world.
What would be your changes to this test (I have a tendency to miss important points) to make it so you can agree with it.
Again this should really be a breeze to take as the sonic differences are substantial.
YOU guys can end this debate (ears, gears, knowledge) ... I can NOT and spout only old codswallop or have 2-bobs each way and have my own flawed religion and biases.
be CONSTRUCTIVE guys...