USB cable group test in HFN

[Whaleblue]

And thus implying that there are Einstein laws for audio just around the corner, waiting to be discovered? Newton breaks down in extreme circumstances, either very small thingies or very fast thingies, which have no consequence in daily physics. Is audio operating at the edge of some such extremity that might warrant an entirely new view on things?

I don't know (though I doubt it). But that is the point I'm trying to make.

To eschew all subjective evidence to the nth degree because it does not fit the facts would be a mistake.

In the cases discussed here there are, of course, holes that are so quickly found it is not worth the effort of giving them any credence. That, I accept.

But thinking about it, electricity is a very fast thingy, isn't it

 

[Mescalito]

I don't know (though I doubt it). But that is the point I'm trying to make.

To eschew all subjective evidence to the nth degree because it does not fit the facts would be a mistake.

In the cases discussed here there are, of course, holes that are so quickly found it is not worth the effort of giving them any credence. That, I accept.

But thinking about it, electricity is a very fast thingy, isn't it

So let's have a working hypothesis which accounts for the perceived differences, then. One that accouints for the fact that the same datastream passing through different USB cables produces differen sounds when decoded by the same DAC. Anything plausible we can test will do.

Chris

 

[clivem2]

So let's have a working hypothesis which accounts for the perceived differences, then. One that accouints for the fact that the same datastream passing through different USB cables produces differen sounds when decoded by the same DAC. Anything plausible we can test will do.

Chris

You could try something along the lines of post #448.

 

[adamdea]

Werner, I'm surprised at you I chose my words carefully "disproved as being the be all and end all"

Although Newton's laws are 99.999% correct they have been found offer a slightly incomplete picture. That was the point I was trying to make.

Yes but why are those laws so powerfully predictive? why did they last so long? Because they fitted consistently with the observable phenemona. They were neither based on nor disproved by anecdotal reminiscences. They couldn't be really because pretty much everything we ever observe in our lives will conform to the Newtonian "special case".None of this has anything to do with the belief systems of audiophiles.



The key to all of this is to consider what constitutes evidence in the context, as well as to consider the important and anything but absolutist principle that extraordinary claims require extraordinary proof. Do you think that it is "absolutist" not to believe in homeopathy? Loads and loads and loads of people do. Is there mounting subjective evidence?

 

[hammeredklavier]

(referring to the winning cable at £6.5k)

No typo, 9 out of the 10 tested were priced from £18 - £139, with the £6.5k included as the tenth as a 'cost-no-object comparison'.

Yeah, I just read some of the other posts. I was only skipping through this thread, but that did catch my eye. Blimey, that makes Naim's crackpot vibration resistant kettle lead look like the hi-fi bargain of the century!

 

[arthur]

Presumably you've also sent PM that email. Any reply yet?

Why are you banging on about it?

If you want to ensure PM sees it, send it to him yourself.

I imagine he'll be too ashamed to reply.

 

[arthur]

I don't know (though I doubt it). But that is the point I'm trying to make.

To eschew all subjective evidence to the nth degree because it does not fit the facts would be a mistake.

In the cases discussed here there are, of course, holes that are so quickly found it is not worth the effort of giving them any credence. That, I accept.

But thinking about it, electricity is a very fast thingy, isn't it

Not in cables. Not fast enough to trouble big Al.

 

[pure sound]

Why are you banging on about it?

If you want to ensure PM sees it, send it to him yourself.

I imagine he'll be too ashamed to reply.

Another warrior.

 

[arthur]

Another warrior.

Another Mary

 

[Mescalito]

You could try something along the lines of post #448.

Ok,

Let us assume that something USB cable related IS changing the output of the DAC.

The first thing to do is to analyse the output in the analog domain. Which we do with the 2 USB cables which have been reported as being the most different.

And we find that the DAC output is within the specified bounds for the DAC in qustion. This immediately throws very strong doubt on the original assumption. It aslo effectively shows that if the same sequence of 0's & 1's is fed through the same DAC, you get the same output. Not suprising.

So lets try re-digitising the output from the DAC using the 2 USB Cables and do a bit by bit comparison. They will almost certainly be slightly different due to the random nature of noise in the analog output stage of the DAC. But we can do further analysis to determine if the differences are more significant than that.

If this analysis shows that the output has changed to a greater extent than random noise variation would account for, something is happening. If random noise can be shown statistically to account for the differences, something still may be happening, but it's at a level lower than the random noise level of the DAC.

The question then has to be asked if differences of this magnitude are audible.

We also, at this stage have to start questioning the original assumption.

And here, we are up against the old double blind test again. To do this properly is going to be very slow & tedious, and a sufficient no. of test subjects will have to take part to make the results statistically robust. Not cheap.

Or we can apply Occam's razor and weigh the possibility of unknown physics existing in a field as well studied & established as physics being at play to such an extent that the relatively crude discrimination of the human hearing mechanism can dedtect it but that far more sensitive instruments cannot, or that we are dealing with very well researched and documented psychoacoustical effects. Which one does a reasonable & rational person favour?

Chris





O

 

[clivem2]

Ok,

Let us assume that something USB cable related IS changing the output of the DAC.

The first thing to do is to analyse the output in the analog domain. Which we do with the 2 USB cables which have been reported as being the most different.

And we find that the DAC output is within the specified bounds for the DAC in qustion. This immediately throws very strong doubt on the original assumption. It aslo effectively shows that if the same sequence of 0's & 1's is fed through the same DAC, you get the same output. Not suprising.

So lets try re-digitising the output from the DAC using the 2 USB Cables and do a bit by bit comparison. They will almost certainly be slightly different due to the random nature of noise in the analog output stage of the DAC. But we can do further analysis to determine if the differences are more significant than that.

If this analysis shows that the output has changed to a greater extent than random noise variation would account for, something is happening. If random noise can be shown statistically to account for the differences, something still may be happening, but it's at a level lower than the random noise level of the DAC.

The question then has to be asked if differences of this magnitude are audible.

We also, at this stage have to start questioning the original assumption.

And here, we are up against the old double blind test again. To do this properly is going to be very slow & tedious, and a sufficient no. of test subjects will have to take part to make the results statistically robust. Not cheap.

Or we can apply Occam's razor and weigh the possibility of unknown physics existing in a field as well studied & established as physics being at play to such an extent that the relatively crude discrimination of the human hearing mechanism can dedtect it but that far more sensitive instruments cannot, or that we are dealing with very well researched and documented psychoacoustical effects. Which one does a reasonable & rational person favour?

Chris





O

You mention 0s and 1s, maybe these differ when the sound changes but my gut reaction is not. It's more likely something in the analogue domain.

Actually the very first thing to do is to characterise the difference. If for instance it shows up as broader soundstage we first need to take measurements that show what happens when we perceive a broader soundstage (not due to the cable). Without this control we cannot know whether the measurements we are taking are the right ones. Measuring the output of the DAC is fine but FR won't be enough, there may need to be phase measurements, who knows? Until we have the control set up it's all open the question. Someone's PhD perhaps?

 

[Paul R]

We don't have to worry about what you see as 'broader soundstage', just whether the analogue output is significantly different. It's not difficult to do.

Paul

 

[Mescalito]

You mention 0s and 1s, maybe these differ when the sound changes but my gut reaction is not. It's more likely something in the analogue domain.

Actually the very first thing to do is to characterise the difference. If for instance it shows up as broader soundstage we first need to take measurements that show what happens when we perceive a broader soundstage (not due to the cable). Without this control we cannot know whether the measurements we are taking are the right ones. Measuring the output of the DAC is fine but FR won't be enough, there may need to be phase measurements, who knows? Until we have the control set up it's all open the question. Someone's PhD perhaps?

That's why the very first stage of my procedure calls for the DAC outputs to be measured & ananlysed. Admittedly, I did not specify the measurements, but certainly S/N. FR, phase distortion would be part of it.

Chris

 

[clivem2]

We don't have to worry about what you see as 'broader soundstage', just whether the analogue output is significantly different. It's not difficult to do.

Paul

It's one thing measuring speakers in this regard by moving off-axis but working out what's happening when the electronics (not the recording) change the soundstage - this is a whole lot more obscure. I'm not convinced this shows up with traditional measurements. A control metric is required.

 

[Mescalito]

It's one thing measuring speakers in this regard by moving off-axis but working out what's happening when the electronics (not the recording) change the soundstage - this is a whole lot more obscure. I'm not convinced this shows up with traditional measurements. A control metric is required.

But the bottom line is that the output is a continuous waveform which can be compared & analysed. All we can ever hear is that transverse waveform translated into a longitudinal series of continuos pressure variations.

So all that is necessary is to analyse & compare the output waveforms. We are looking for quantitative differences. At this stage of the proceedings, their qualitative nature is not of interest.

Chris

 

[clivem2]

But the bottom line is that the output is a continuous waveform which can be compared & analysed. All we can ever hear is that transverse waveform translated into a longitudinal series of continuos pressure variations.

So all that is necessary is to analyse & compare the output waveforms. We are looking for quantitative differences. At this stage of the proceedings, their qualitative nature is not of interest.

Chris

We need to know that what we measure allows us to reach valid conclusions. If not it might like trying to work out how fast a car is going by measuring the temperature of the engine. Just showing differences or a lack of difference isn't enough until we've proved what eg soundstand changes look like. If might be a really very subtle difference that's all too easy to miss. I've seen people say that 2 outputs are close enough to be considered the same - these outputs tend to never be identical so people say the difference isn't audible - this is known as Conformation Bias. It might be right but it might not be.

 

[Mescalito]

We need to know that what we measure allows us to reach valid conclusions. If not it might like trying to work out how fast a car is going by measuring the temperature of the engine. Just showing differences or a lack of difference isn't enough until we've proved what eg soundstand changes look like. If might be a really very subtle difference that's all too easy to miss. I've seen people say that 2 outputs are close enough to be considered the same - these outputs tend to never be identical so people say the difference isn't audible - this is known as Conformation Bias. It might be right but it might not be.

If a soundstage exists as a physical entity it is in the waveform. We don't need to know what it looks like at this stage. We just need to know if the waveform produced using one cable is in any way different from that using another cable.

Then you identify any differences & only then do you try to correlate those differences to any observations.

Doing it your way introduces an expectation that "sound stage" has a physical existance. This is by no means sure, as it's very existence is predicated upon an aural illusion.

Chris

 

[clivem2]

Doing it your way introduces an expectation that "sound stage" has a physical existance. This is by no means sure, as it's very existence is predicated upon an aural illusion.

There is something that drives soundstage (soundstage is only one example btw) - possibly it's related to increased low-level detail. I don't believe there will be "a soundstage parameter or metric", it'll be something that indicates there's the potential for increased soundstage.

 

[Mescalito]

There is something that drives soundstage (soundstage is only one example btw) - possibly it's related to increased low-level detail. I don't believe there will be "a soundstage parameter or metric", it'll be something that indicates there's the potential for increased soundstage.

And by definition, that something, be it low level detail, phase changes, anything, will be in the waveform. So you cfompare the waveforms first & foremost.

Chris

 

[clivem2]

And by definition, that something, be it low level detail, phase changes, anything, will be in the waveform. So you cfompare the waveforms first & foremost.

Chris

I agree, we just need to know what we're looking for, it might be really quite small.