[Harry Manback]

Interesting read.
 
http://arstechnica.com/gadgets/2015/07/even-vegas-strangers-agree-340-audiophile-cables-make-no-difference/
 
Love me some ArsTechnica.com

 

[wnmnkh]

I don't make criticisms often on ABX test, but this is such poorly done test in many regards.
 
1) The whole setup is basically Grado RS-2e directly connected to a Dell laptop. Seriously? Not even using a cheap, decent external amp/dac? (and this is rather subjective, but I wouldn't say Grado headphone is suitable for such test as well)
 
2) The listening environment is not even properly isolated from outside noise. No, "a listening station" does not cut it. And Grado headphone is open-aire.
 
3) Even worse is that the laptop itself was crapping itself throughout the test. Once again the setup was far from stable to use for audio playback.
 
4) The playback software was WMP. WMP uses Direct Sound mixer, and its sound quality is meh-ok at best. There is a reason why people use ASIO, Kennel Streaming or WASAPI which bypasses Direct Sound and communicate directly with soundcards/DACs. 
 
 
With such conditions, even the most trained person would have a hard time comparing a 128k mp3 created by a crappy original Fraunhofer converter and original wav file.
 
 
I think Ethernet cable making audible difference is ridiculous, but claiming this test to be valid is even more absurd.

 

[TheGrumpyOldMan]

I'm surprised it took this long to show up here, thought about starting a thread. It's not that often that our audiophile interests end up in a more "mainstream" publication. The comments section can be a fun read, as always with such a subject

 
I was sure when I read the equipment list, the above would come up. But keep in mind that one reason they did so because the claims by some were that differences were "not subtle or slight" and "as plain as day.", so they should very well be discernible under those circumstances by a measurable majority of people.
 
They also include links to articles critical of the whole thing, mostly from the Audiostream site, if that's more to one's liking.
 
But more interestingly, they also have another article with the dissection of the cable, at least it seems to be a decent one, though still with some quality issues... and at least it's not filled with magic sand

. Way over-priced though, either way.
 
And today a follow-up article with actual measurements with a very expensive tester, courtesy of Blue Jeans Cable. Of course BJC also provide a measurement of their own product, so some might object to that. Either way, based on the measurements and sensible response of Kurt Denke, I will look into their cables in the future. And that super-cheap ethernet cable was actually pretty crap, failing some tests, so you do get what you pay for, and I will keep that in mind for my next such purchase, even non-audio related...

 

[arnyk]

  I don't make criticisms often on ABX test, but this is such poorly done test in many regards.
 
1) The whole setup is basically Grado RS-2e directly connected to a Dell laptop. Seriously? Not even using a cheap, decent external amp/dac? (and this is rather subjective, but I wouldn't say Grado headphone is suitable for such test as well)
 
2) The listening environment is not even properly isolated from outside noise. No, "a listening station" does not cut it. And Grado headphone is open-aire.
 
3) Even worse is that the laptop itself was crapping itself throughout the test. Once again the setup was far from stable to use for audio playback.
 
4) The playback software was WMP. WMP uses Direct Sound mixer, and its sound quality is meh-ok at best. There is a reason why people use ASIO, Kennel Streaming or WASAPI which bypasses Direct Sound and communicate directly with soundcards/DACs. 
 
 
With such conditions, even the most trained person would have a hard time comparing a 128k mp3 created by a crappy original Fraunhofer converter and original wav file.
 
 
I think Ethernet cable making audible difference is ridiculous, but claiming this test to be valid is even more absurd.

 
The above appears to the usual kind of nit-picking that follows any test whose outcome people don't appreciate.
 
It can be summarized as follows:
 
(1) The equipment used in the test are not my personal favorites or what I own, so the test is invalid.
 
(2) The equipment used in the test are not my personal favorites or what I own, so the test is invalid (repeated as needed to allow me to dismiss and deny)
 
(3) I can pick nits with some of the things that happened during the test but were corrected.
 
(3) I can pick nits with some of the things that happened during the test for which there is reliable evidence that it is not actually a problem, but I still don't like it.
 
The test was not rocket science. If you wish to allegedly refine it, you probably could.

 

[Exacoustatowner]

   
The above appears to the usual kind of nit-picking that follows any test whose outcome people don't appreciate.
 
It can be summarized as follows:
 
(1) The equipment used in the test are not my personal favorites or what I own, so the test is invalid.
 
(2) The equipment used in the test are not my personal favorites or what I own, so the test is invalid (repeated as needed to allow me to dismiss and deny)
 
(3) I can pick nits with some of the things that happened during the test but were corrected.
 
(3) I can pick nits with some of the things that happened during the test for which there is reliable evidence that it is not actually a problem, but I still don't like it.
 
The test was not rocket science. If you wish to allegedly refine it, you probably could.

I noticed that there were frequently 30 second delays between unplugging one cable, and another and getting the music playing. Even if you discount the quality of the audio-and I've NO idea how good these Grado Headphones are-for example- you can't discount the delays. Given that I would be skeptical of the results one way or another-and I don't believe the claims for the magic cable. It's a data cable! A DIRECTIONAL data cable???? All that said, your points are valid- even given the best testing set up the claims would fly in that if they had used the RIGHT amps and dads then THEN the differences would be obvious. Did I mention DIRECTIONAL cables? 

 

[arnyk]

  I noticed that there were frequently 30 second delays between unplugging one cable, and another and getting the music playing. Even if you discount the quality of the audio-and I've NO idea how good these Grado Headphones are-for example- you can't discount the delays. 

 
Now that you have brought up the 20 second delay, I have to accept it as being a potentially serious problem.
 
This appears to be the mention of it:
 
"I had several moments of panic throughout the actual test when the laptop balked—usually after yanking out one cable for the other and trying to re-establish contact. Things went smoothly almost every time, but there were a few instances where the laptop just hung, and I had to kill the Windows Media Player and Explorer processes. Fortunately, this had no impact on the test other than the occasional 20- or 30-second delay."
 
So the 20-30 second delay was an occasional event and not necessarily a part of the entire experiment?
 
I wonder what the normal delay time was - probably still several seconds and too long for peak listener sensitivity.

 

[wnmnkh]

 

  I don't make criticisms often on ABX test, but this is such poorly done test in many regards.
 
1) The whole setup is basically Grado RS-2e directly connected to a Dell laptop. Seriously? Not even using a cheap, decent external amp/dac? (and this is rather subjective, but I wouldn't say Grado headphone is suitable for such test as well)
 
2) The listening environment is not even properly isolated from outside noise. No, "a listening station" does not cut it. And Grado headphone is open-aire.
 
3) Even worse is that the laptop itself was crapping itself throughout the test. Once again the setup was far from stable to use for audio playback.
 
4) The playback software was WMP. WMP uses Direct Sound mixer, and its sound quality is meh-ok at best. There is a reason why people use ASIO, Kennel Streaming or WASAPI which bypasses Direct Sound and communicate directly with soundcards/DACs. 
 
 
With such conditions, even the most trained person would have a hard time comparing a 128k mp3 created by a crappy original Fraunhofer converter and original wav file.
 
 
I think Ethernet cable making audible difference is ridiculous, but claiming this test to be valid is even more absurd.

 
The above appears to the usual kind of nit-picking that follows any test whose outcome people don't appreciate.
 
It can be summarized as follows:
 
(1) The equipment used in the test are not my personal favorites or what I own, so the test is invalid.
 
(2) The equipment used in the test are not my personal favorites or what I own, so the test is invalid (repeated as needed to allow me to dismiss and deny)
 
(3) I can pick nits with some of the things that happened during the test but were corrected.
 
(3) I can pick nits with some of the things that happened during the test for which there is reliable evidence that it is not actually a problem, but I still don't like it.
 
The test was not rocket science. If you wish to allegedly refine it, you probably could.

 
Please, Arny, you can do MUCH better than this.
 

 
Most of the cases, onboard soundcard on laptop is simply sub-standard, with audible noise floor and distortions...  worse than my 30 bucks Sansa Clip.
 
I can't find the measurements for Dell M2800 onboard soundcard, but I suspect it is as miserable as this one above. It is very unwise to disregard this concern as playing 'favorite'.
 
Now, I also want to point out these guys did not even measure performance of the test equipment. We have no way of knowing whether the computer/soundcard was performed in a way that would not affect the result (one of the reasons why Direct Sound has disadvantages due to interaction with OS). This could be at least sorta tested with some cheap but reliable ADC.
 
My concern about Grado is mostly technical one (yes, I dislike Grado in general but it is indeed subjective opinion, as I mentioned already). It is open-aire headphone, which is very vulnerable to outside noise.
 
 
This is not a rocket science, but the test is so poorly done that my middle school science teacher would given F on both effort and accuracy.

 

[arnyk]

   
Please, Arny, you can do MUCH better than this.
 

 
I'd try to do much better, such as show a measurement for a Dell M2800 laptop that was actually used (ALC3226 audio chipset), and not try to palm off a measuerment the Thinkpad 301 (Conexant Audio Chip) shown above.   Besides, if I was going to make a point about noise, I'd use a noise test, not a frequency response test.

 

[wnmnkh]

 

   
Please, Arny, you can do MUCH better than this.
 

 
I'd try to do much better, such as show a measurement for a Dell M2800 laptop that was actually used (ALC3226 audio chipset), and not try to palm off a measuerment the Thinkpad 301 (Conexant Audio Chip) shown above.   Besides, if I was going to make a point about noise, I'd use a noise test, not a frequency response test.

 
Now, this is much better response. Thank you.
 
Yes, this is not a graph for M2800, and I did not use the graph for M2800 because... there isn't one. Wish folks at Ars and Randi organization could had done measuring for me. And noise is just one of the problems laptops' soundcards usually have. I have no intention to focus solely on noise alone. As the graph above shows, quite a lot of laptops suffer poor frequency response too.
 
I am not certain really. Sure, M2800 is a workstation laptop and -hopefully- should had been manufactured at tighter tolerance. But unless we are talking about very early generations of Sony VAIO laptops which were tailored for audio/video production, most laptops' sound quality is truly afterthought. Even Apple's 2008 Macbook Pro cannot do true 16-bit processing, for instance; you just cannot trust laptop's soundcards to do CD-quality sound even.
 
You probably understand this better than me... but the audio chipset is a small part of equations for sound quality. I bet majority of chipsets would provide very good performance in ideal situations. It is implementation that cripples the sound quality.
And such poor implementation due to cost-cutting and following poor performance are the reasons why I consider typical laptop's sound out as an unacceptable mean for critical listening tests.

 

[Exacoustatowner]

   
Please, Arny, you can do MUCH better than this.
 

 
Most of the cases, onboard soundcard on laptop is simply sub-standard, with audible noise floor and distortions...  worse than my 30 bucks Sansa Clip.
 
I can't find the measurements for Dell M2800 onboard soundcard, but I suspect it is as miserable as this one above. It is very unwise to disregard this concern as playing 'favorite'.
 
Now, I also want to point out these guys did not even measure performance of the test equipment. We have no way of knowing whether the computer/soundcard was performed in a way that would not affect the result (one of the reasons why Direct Sound has disadvantages due to interaction with OS). This could be at least sorta tested with some cheap but reliable ADC.
 
My concern about Grado is mostly technical one (yes, I dislike Grado in general but it is indeed subjective opinion, as I mentioned already). It is open-aire headphone, which is very vulnerable to outside noise.
 
 
This is not a rocket science, but the test is so poorly done that my middle school science teacher would given F on both effort and accuracy.

You can do better-but can you actually HEAR the difference? It does look like the bass output could be a little weak- can we perceive a gradual rolloff maxing at -0.5 db at 20Khz? I doubt it. Perhaps the distortion ratings are high enough to be audible?

 

[Exacoustatowner]

Now, this is much better response. Thank you.

Yes, this is not a graph for M2800, and I did not use the graph for M2800 because... there isn't one. Wish folks at Ars and Randi organization could had done measuring for me. And noise is just one of the problems laptops' soundcards usually have. I have no intention to focus solely on noise alone. As the graph above shows, quite a lot of laptops suffer poor frequency response too.

I am not certain really. Sure, M2800 is a workstation laptop and -hopefully- should had been manufactured at tighter tolerance. But unless we are talking about very early generations of Sony VAIO laptops which were tailored for audio/video production, most laptops' sound quality is truly afterthought. Even Apple's 2008 Macbook Pro cannot do true 16-bit processing, for instance; you just cannot trust laptop's soundcards to do CD-quality sound even.

You probably understand this better than me... but the audio chipset is a small part of equations for sound quality. I bet majority of chipsets would provide very good performance in ideal situations. It is implementation that cripples the sound quality.
And such poor implementation due to cost-cutting and following poor performance are the reasons why I consider typical laptop's sound out as an unacceptable mean for critical listening tests.

I appreciate your points and it looks like the bass response is audibly deficient- which suggests the other performance measurements would be compromised- even if a less than 0.5 db drop in the treble would not be audible

 

[dprimary]

  I'm surprised it took this long to show up here, thought about starting a thread. It's not that often that our audiophile interests end up in a more "mainstream" publication. The comments section can be a fun read, as always with such a subject

 
I was sure when I read the equipment list, the above would come up. But keep in mind that one reason they did so because the claims by some were that differences were "not subtle or slight" and "as plain as day.", so they should very well be discernible under those circumstances by a measurable majority of people.
 
They also include links to articles critical of the whole thing, mostly from the Audiostream site, if that's more to one's liking.
 
But more interestingly, they also have another article with the dissection of the cable, at least it seems to be a decent one, though still with some quality issues... and at least it's not filled with magic sand

. Way over-priced though, either way.
 
And today a follow-up article with actual measurements with a very expensive tester, courtesy of Blue Jeans Cable. Of course BJC also provide a measurement of their own product, so some might object to that. Either way, based on the measurements and sensible response of Kurt Denke, I will look into their cables in the future. And that super-cheap ethernet cable was actually pretty crap, failing some tests, so you do get what you pay for, and I will keep that in mind for my next such purchase, even non-audio related...

At least they actually measured the cables to the industry standards. The Fluke DSX is common tester about every structured cabling company has at least one of not a half dozen of them or equal. It is standard practice to measure and document very single network cable installed in a building the results document is often reviewed client or an outside consultant that all the cabling meets the standards. I was hoping to see them use some high end scopes and other test equipment. I would love to see Belden's labs.
 
While telegaertner is respected I do not trust long RJ45 type connectors. The long connectors tend to act as levers and the RJ45 type connector was not really designed to withstand the added pressure, I have had problems with intermittent connection cause by the additional leverage.
 
Most patch cables are stranded, solid ones tend to fail quickly.