Just let's not forget this:
Bottom line: sonic differences during playback of multiple instances of the same file are possible as there are mechanisms that allow for such differences that don't violate laws of physics etc but the difference is never caused by some hidden information in the files. Multiple instances of the same file, no matter how these instances were created, do not contain persistent information about *how* they were created, let alone they would somehow "inherit" that information to another instance (which just shows the locigal nonsense of the claim, how could only the first copy of a file contain any information about its creation while subsequent copies wouldn't?)
The point is, a different instance of the same file does not and cannot contain any persistent but hidden information about how it was created that would manifest itself in a persistent different sound, let alone when the file is copied multiple times into a new instance again (via Dropbox and from there to the local playback medium etc). But the local playback process may result in a sonic difference depending on where the file is located on the file system etc, by some very weak form of logic induced modulation. Say, if one instance is located on a RAMdisk and another instance is located on SDcard there is a chance that the jitter on the USB port transferring the data to the DAC might have another pattern that would be (very) weakly correlated and repetable, at least this is not entirely impossible. But this correlation is *not* linked to how the original instances of the files were created. It simply doesn't matter, one could use one and the same instance of the file (created by one single copy or rip process) to begin with and would still get a similar difference after creating a new instance.
Bottom line: sonic differences during playback of multiple instances of the same file are possible as there are mechanisms that allow for such differences that don't violate laws of physics etc but the difference is never caused by some hidden information in the files. Multiple instances of the same file, no matter how these instances were created, do not contain persistent information about *how* they were created, let alone they would somehow "inherit" that information to another instance (which just shows the locigal nonsense of the claim, how could only the first copy of a file contain any information about its creation while subsequent copies wouldn't?)
One of the rules of civilised debating is that participants always declare any interest, whether commercial, reputational, relational etc. It is sad that this courtesy is often omitted here. As SY used to say, it is hard to teach someone something when his income depends on him not knowing it.
This is nothing new. Disguised and sitting among the crowd is the accomplice of traveling snake oil salesman who volunteers to try it. Then he acts surprised how well it works.
We've all heard / read about it or even seen different versions of sketch.
I have found that there are two markers which may indicate undeclared commercial interest in audio on this site:
1. dogged persistence in asserting 'facts' which are known to be false i.e. contrary to known science and mathematics
2. continual spreading of FUD (fear, uncertainty, doubt) about generally accepted audio knowledge and hinting that there is more to be known by others (but not, of course, the person spreading the FUD) sometimes combined with a belief in 'proof by absence of counterproof' (i.e. "you can't convince me I am wrong, therefore I must be right").
Both are not always present together, but either should serve as a flag.
1. dogged persistence in asserting 'facts' which are known to be false i.e. contrary to known science and mathematics
2. continual spreading of FUD (fear, uncertainty, doubt) about generally accepted audio knowledge and hinting that there is more to be known by others (but not, of course, the person spreading the FUD) sometimes combined with a belief in 'proof by absence of counterproof' (i.e. "you can't convince me I am wrong, therefore I must be right").
Both are not always present together, but either should serve as a flag.
All good points, my intent was to address the claims like the files even retain their properties when emailed or downloaded onto another computer.
Yep, exactly the same discussion that we had in that german forum. Like it did there, this discussion is going to split participants in two camps: those who believe or pretend to believe, for whatever reasons, that the files do retain the magic property and those who know this is entirely impossible because it is, well, a physical, logical and philosophical impossibility, now and forever.
@KSTR,
Yep that was an interesting one.........
@scott_wurcer,
Only judging from Max Headroom´s last post in this thread wrt the topic it imo didn´t seem that he believes to the "ghost in the machine" but just on different physical processes during the reproduction that could lead to audible differences, but maybe i´m wrong.
Yep that was an interesting one.........
@scott_wurcer,
Only judging from Max Headroom´s last post in this thread wrt the topic it imo didn´t seem that he believes to the "ghost in the machine" but just on different physical processes during the reproduction that could lead to audible differences, but maybe i´m wrong.
The hypothesized physical processes appear contrary to modern well known physics. Maybe the difference between science and a subtle type of scientism.
Last edited:
What is "another pattern"? Are you claiming there's a possibility to get a file, consistently and repeatable, corrupted when transferring from a SD card?
USB is a packet mode data link protocol that transfers data using synchronous serial communication (always pair the data line with a clock signal). CRCs are performed on the data within the packet payload, all token packets have a 5 bit CRC while data packets have a 16 bit CRC. So the probability to get a bit corrupted is not zero, but infinitesimal, probably of the same order of magnitude as the possible errors from transferring from a RAM disk (which usually doesn't have CRC, BTW). Timing errors (jitter, etc...) are no longer an issue once the frame data is successfully delivered to the target.
Otherwise said, the reconstructed frame at the USB output has no memory of passing through a USB, UART, or whatever other type of serial interface. It's either identical to the original, or it is not (which would be one of those extraordinary claims that needs extraordinary proof).
Last edited:
I wish we could tackle a more compelling problem, but I think you missed comments in the far past. IIRC The files were copied with and without a "treated" USB cable and folks here were invited to download them, the treatment was intended to stay with the file for any subsequent reproduction.
I think what is being described here is undesired analog artifacts from poorly designed digital devices. Crosstalk from motor and focusing action in cheap CD players for instance has been observed in the jitter spectrum of the data.
Ok, but then this is not related to the USB itself, but yet again related to a pathologically bad implementation. Such bad implementations can always be invoked to justify audible effects. Back to the mains cable example, it is very possible that a mains cable used at 25A/mm^2 could create audible artifacts compared to another mains cable at 5A/mm^2. Or a loose contact, etc...
Anyway, to identify such pathological USB issues a listening test is not required. One would need a BERT and some software, it will show data errors in the ppb range or even less.
Last edited:
The psychology and underlying reasoning of those making, rehashing or supporting outrageous claims is fascinating. Except when the stink of money covers the scene.
Could it be that you are now invoking "hidden variables/informations" ?
The question was just the short form, as i´ve cited results from various studies examing/comparing different test protocols. The first one (means that i´m aware of) dated back from ~1952, when experiments noticed that testing DLs for pitch gave lower numbers for A/B than for ABX tests. They speculated already back then that the ABX method was more involving for the participants, especially wrt the internal judgement processes.
In later studies these differences were always confirmed (the A/B was just an example, the same holds true when ABX is compared with 3AFC ) but is of course not restricted to the ABX test. it was found for various test protocols that the results were diverging, so the obvious conclusion is that the test participants not only respond to the sensory difference but to the test question/task as well.
Fortunately there is something like Thurstonian/Signal Detection Theorie modelling available that tries to deliver some additional insight and introduces a more comparable measure by using kind of a signal to noise ratio.
SDT started during WWII and in the meantime there was a lot of progress made, including high number monte carlo simulations based on the Thurstonian model resp. the predictions.
Seem that i don´t understand these last two sentences, could you rephrase ?
FWIW, the USB bit error rate is estimated (under worst case conditions, extreme industrial noise conditions, extreme temperatures, etc...) as 1:10,000,000,000, that is 1 bit error at every 10 billion bits transferred. Under normal conditions, one may not live long enough to see one wrong bit.
If the USB interface radiates noise into the analog side, or ground loops are injecting noise in the DAC, etc... then shield the ******* thing, break the ground loops or use isolation transformers, etc..., don't blame the USB for artifacts.
Last edited:
I didn´t try to follow the assertion (so have no idea if there could be something to it), but being wrong on physical process is surely a lesser sin than believing in spiritual informations mysteriously connected not to bits but to the information sometimes represented by the bits.....
@DF96,
Declaration of interest is common for published scientific information, that it is an accepted generalized rule for debates is a least new to me.
In practice it is my impression that those informations are mainly/often mainly used to cover a lack of arguments by stepping into the land of eristics.
And i´d like to modificate or add to SY´s statement by "it is hard to teach someone something when his ego state depends on him not accepting it" .
Imo your are typically spreading FUD by your vague comments about financial interests or other dubious hidden motivations.......
- Status
- Not open for further replies.