Yes, that is also true. It is one way to distinguish the scientist from the quack: the scientist welcomes genuine correction, especially if delivered in the right spirit.Jakob2 said:
I guess one man's FUD is another man's attempt at clarity.
What I have seen on this forum on a number of occasions is that someone persists in defending an indefensible position or attacking a reasonable position, but not by offering good arguments. Eventually the mask slips and he says something like "well I must be right because of all the sales/articles/money I have accumulated". Then we say "Ah! That explains it.".
The obvious answer seems to be "Yes" .
Imo this thread shows meandering to a lesser degree than many others and tends to stay mainly on the original thread topic and related ones as well.
Btw, i´m quite confident that at least some of the material that i contributed in this and the other "DAC sound all the same" thread wasn´t posted already by other members.
Further we know that it need much repetition before facts are acknowledged and remembered.
I stated repeatedly I don't propose ABX use A/B if you want or any other protocol but propose something.
My comment was meant to say, for instance, if you come up with a protocol that suddenly gets dramatically more correct results would you not question why or just say that it is better?
EDIT - I see another wire thread has started, any bets?
Last edited:
If lies are repeated enough times, (some) people start to think of it as truth. That's how propaganda is spread.
Did i said something to the contrary?
As repeated slightly less than 148 times, what protocol to choose depends on the objective of an experiment.
So i guess this is the rephrasing?
Actually it was the other way round; ABX was the new kid on the block and therefore experimenters were interested about comparisons to the established protocols.
As said before, it´s not only a matter with ABX; in general the triangle and Duo-Trio (to which the ABX variant belongs) share similar problems. The so-called operational power of these protocols is way smaller than of 2AFC or 3AFC protocols.
Which means if you want to reach the same statistical power, you´ll need a lot more trials than you´ve expected. The difference can be quite dramatic, for example that you´d need 21 trials in the 3-AFC test while the triangle would require 276.
When comparing A/B and duo-trio for the same conditions,you´d need 27 trials in the A/B but 310 for the duo-trio tests.
I thought this was a spin-off of the DAC test, otherwise pick one (please not whether silver wire sounds bright) and an example of how to proceed.
BTW do you have any links to tests for audibilty of impairments that you think were properly conducted?
Mmmm. I suppose. Fascination turns to frustration turns to weariness.
But it does keep to ink forum flowing, so have at it.
I'm stunned to this day that folks still quote Eddington's issues around trying to prove Einstein's prediction of light defected in a gravitational field and then simply end the conversation. Maybe he had problems but in the last eclipse even a serious amateur had access to the equipment and technology to verify the predictions to an extreme accuracy.
Once you look closer, you'll find that audio gear more often than not is a pathologically bad implementation. It's almost impossible to find a device that fully adheres the principles and best practises of signal integrity and EMC. This is a cost factor so it's clear that budget devices try to cut corners here but shockingly bad implementation is often found in the boutique class as well. For example it really is no wonder that power and interconnect cables do make a difference given the design flaws found in almost all HiFi gear (and some studio gear as well, even the rather trivial "Pin1 problem" is still present in many new products).
Back to the example with two identical files sounding different (which actually means: two playback processes of the same file sounding different). We know that when the DAC-chip receives the identical bitstream this alone does not suffice to produce the same output, it's only a prerequisite. The output is the same only to first order. Factors that can influence the output are:
- noise on the reference voltages
- jitter on the clock and data lines
- noise on the supplies (including GND planes etc) in general
- capacitive or inductive coupling into the analog output section
(list in no particular order and not neccessarily complete)
As soon as any of these factors isn't truly random but correlated to the incoming digital signal (when viewed with all its analog properties -- all signals are analog after all) there is a chance to have a signature. Now, is there a possibility to have different signatures coming from different playback processes? Yes there is. It is far fetched but not impossible, computers are analog devices and we'll find the analog properties of the digital audio interface being ever so slightly modulated by the internal states.
Do I personally believe in these very very subtle effects could ever be audible, given the mechanisms are really weak and also very remote to the actual DA-conversion? No, I have neither heard the effect myself nor have I measured it. Then again, did I believe in audible logic-induced-modulation down in the device before I knew anything about it, many years ago (and before I was able to reliably measure it like I am now)? No, I didn't. Now I do....
No it didn't, amongst other things it meant 2 identical files, repeatedly played in the same hardware sounding repeatedly and - easily identified as - different. One file had been "treated" in a way to make is sound reliably different (though bit identical). You are thinking of very hard to notice differences, that should be randomly distributed across the files when played many times.
For those wanting to hear what MAX GOOP can do to (audio) files when copied over USB.
Funniest snake oil theories
Funniest snake oil theories
Two bit-identical files no matter how they were created are still the *same* file by sheer definition (at least in the IT world) but *two* physical instances of it. Differences at playback time are possible just because it's two instances of the same thing, not two different things to begin with.
I do not doubt the reported listening experience but the difference must have been caused by another yet unidentified mechanism (that still must follow the laws of physics) that was mistakingly attributed to the files themselves, that is, details of their creation. "A file, like an gentleman, has no memory" ;-)
I do not doubt the reported listening experience but the difference must have been caused by another yet unidentified mechanism (that still must follow the laws of physics) that was mistakingly attributed to the files themselves, that is, details of their creation. "A file, like an gentleman, has no memory" ;-)
Scott, if we both download Dan's file from his "standard" folder we both have the same file but there are now two instances of it, one in Wales and one in Germany, agreed?
And if I downloaded the file from "treated 1" instead I still have the same file that you have (as it is bit-identical). If Dan were clever he could use one single identical file, secretly uploaded again, to all three folders and have his supporters repeat the listening test with the same outcome as before, wanna bet? Actually no one could even tell when he also manipulated the time-stamps of the files.
And if I downloaded the file from "treated 1" instead I still have the same file that you have (as it is bit-identical). If Dan were clever he could use one single identical file, secretly uploaded again, to all three folders and have his supporters repeat the listening test with the same outcome as before, wanna bet? Actually no one could even tell when he also manipulated the time-stamps of the files.
Last edited:
- Status
- Not open for further replies.