Head-Fi.org › Forums › Equipment Forums › Sound Science › My cable test enterprise
New Posts  All Forums:Forum Nav:

My cable test enterprise - Page 5

post #61 of 438
Quote:
Originally Posted by nick_charles View Post
Transient response might be a factor , I will scope some square waves but differences between two 1m lengths of cable cannot have any significant effect on phase, even doubling the cable length would only add 0.021 degrees of phase shift even at 20K.
«Significant» already implies interpretation. Also <0.1dB is considered insignificant by conservative standards and nonetheless of importance in some cases. Just translate the same amount to phase deviations! You never really know.
post #62 of 438
Thread Starter 
Quote:
Originally Posted by JaZZ View Post
I And as to the latter, unfortunately your intended digitization will inevitably corrupt it (considering the tough filtering done by ADC and DAC).
.
1.Will that corruption be audible ?, tests suggest probably not

2.In any case the condition is the same for all test candidates so even if some mild corruption takes place the differences between cables will still be preserved.
post #63 of 438
Quote:
Originally Posted by nick_charles View Post
1.Will that corruption be audible ?, tests suggest probably not

2.In any case the condition is the same for all test candidates so even if some mild corruption takes place the differences between cables will still be preserved.
Have you taken notice of my corresponding report a few posts back?
post #64 of 438
Thread Starter 
Quote:
Originally Posted by JaZZ View Post
«Significant» already implies interpretation. Also <0.1dB is considered insignificant by conservative standards and nonetheless of importance in some cases. Just translate the same amount to phase deviations! You never really know.
Sorry but as far as publsihed research goes a phase shift of well under 1 degree is just not audible under any circumstances.
post #65 of 438
Quote:
Originally Posted by nick_charles View Post
Sorry but as far as publsihed research goes a phase shift of well under 1 degree is just not audible under any circumstances.
Phase and transient response are just two different aspects of the same thing: the time axis. If a deviation of <0.1 dB is audible, the inevitable phase/transient distortion that comes with it can just as well be audible or is even the cause for audibility.

If you do the ABX test, please do it in an unbiased, non-ideologic manner that makes it as easy as possible to find audible differences, not the other way round!
.
post #66 of 438
Thread Starter 
Quote:
Originally Posted by JaZZ View Post
Have you taken notice of my corresponding report a few posts back?
Yes but you admit it was done sighted, thus you cannot exclude human bias.

Also you are talking about altering the frequency balance not just the attenuation, now so far none of the cables I have tested have done this.

While thay have all attenuated more or less they have all left the FR at the same shape, none have rolled-off or sucked-out more than any other.
post #67 of 438
Thread Starter 
Quote:
Originally Posted by JaZZ View Post
Phase and transient response are just two different aspects of the same thing: the time axis. If a deviation of <0.1 dB is audible, the inevitable phase/transient distortion that comes with it can just as well be audible or is even the cause for audibility.

If you do the ABX test, please do it in an unbiased, non-ideologic manner that makes it as easy as possible to find audible differences, not the other way round!
.
So far no cable has been on average more than 0.009db worse than any other, in the most extreme case one cable was for one specific frequency 0.02 db worse than the best cable.

My ABX tests wll be blind. Blind tests are actually extremely sensitive. I have use blind tests successfully to detect the difference between 16/44.1 and 24/96 recording artifacts and to detect the presence of different low pass filters and level differences between samples or distortion products.

If I hear differences it will not be due to my knowledge of what I am listening to. I think this is the only fair way to do it.

Sighted listening tests when testing for really small differences are I am afraid to say just too unreliable. As mentioned before I hate the Blue Jeans Cable design with a vengeance. A sighted test on this would be fatally flawed.
post #68 of 438
I think Nick really needs to compare some better cables to the ones he has. go silver, signal cable silver resolution cable with bullet can be had for $100 from audiogon. homegrownaudio also has nice silver cables. then come back and tell us your opinions.

below certain price point, or after certain price point, probably it is hard to tell a difference. but I think there is a notable improvement going from $30 IC to $200 IC. This is my opinion after comparing my BJC IC with signal silver and another DIY silver cable.
post #69 of 438
Quote:
Originally Posted by nick_charles View Post
Yes but you admit it was done sighted, thus you cannot exclude human bias.

Also you are talking about altering the frequency balance not just the attenuation, now so far none of the cables I have tested have done this.

While thay have all attenuated more or less they have all left the FR at the same shape, none have rolled-off or sucked-out more than any other.
My claim stands: Do the test as unbiased an non-ideologic as possible -- and in a way that excludes all variables which could affect the perceptibility by introducing further distortion and thus minimizing the differences to insignificance! I must say, though, that your latest statements make me quite pessimistic in this regard.

I'm out.
.
post #70 of 438
Thread Starter 
Quote:
Originally Posted by andychen View Post
I think Nick really needs to compare some better cables to the ones he has. go silver, signal cable silver resolution cable with bullet can be had for $100 from audiogon. homegrownaudio also has nice silver cables. then come back and tell us your opinions.

below certain price point, or after certain price point, probably it is hard to tell a difference. but I think there is a notable improvement going from $30 IC to $200 IC. This is my opinion after comparing my BJC IC with signal silver and another DIY silver cable.
Lend me your better cable and I will happily compare it against the cables I currently have.
post #71 of 438
Quote:
Originally Posted by nick_charles View Post
Lend me your better cable and I will happily compare it against the cables I currently have.
Do you only need one cable, as in not both cables of an RCA set? If so I may be able to send over one of the "granddaddy" Cardas Golden References that I received when doing my test. I have a pair of paper clip ICs I may be able to send along with that. That should provide enough variation (both in price - 10$ compared to 1200$) and in conductor material ( steel or aluminum - whatever paper clips are made of - versus golden ratio stranded copper constant q blah). I also have a little white POS cable. The largest measured difference I came up with was between the Cardas Golden Reference and this little white POS. Heres a picture with white noise showing electrical differences -


Send me a pm if you are interested. I need to talk it over with my folks though, even though its only half of a set it still holds some type of value I guess.

Dave
post #72 of 438
Thread Starter 

Statistical significance

I just loaded my dataset up into SPSS and ran a repeated measures ANOVA with cable type as a covariate and some ANOVAs between cables.

The results show that for these cables there was no significant effect on changing cable, no cable was significantly different from any other. Frequency also did not alter the result, no cable was significantly different at any frequency.

The other result was that there was much more variation between trials than between cables, this is entirely predictable as a cable is a passive device and and ADC is not, so my decision to average 10 trials was justified. Amusingly the repeated measures variation was significant. Remember this was a variation of no more than 0.056db between any two trials for any given cable at any frequency, but when you have 4600+ data points little differences become statistically significant.
post #73 of 438
Thread Starter 
Quote:
Originally Posted by JaZZ View Post
My claim stands: Do the test as unbiased an non-ideologic as possible -- and in a way that excludes all variables which could affect the perceptibility by introducing further distortion and thus minimizing the differences to insignificance! I must say, though, that your latest statements make me quite pessimistic in this regard.

I'm out.
.
Please explain to me how I can do it in a less biased way. I could try sighted listening tests but I just do not see how I could possibly avoid bias here. As I mentioned repeated blind cable swaps are just not viable.

As for AD distortion, well I can test this by doing an ABX between a reference music sample and a DA/AD if the gain adjusted files are detectably different than I will drop that idea.
post #74 of 438
Quote:
Originally Posted by nick_charles View Post
Please explain to me how I can do it in a less biased way. I could try sighted listening tests but I just do not see how I could possibly avoid bias here. As I mentioned repeated blind cable swaps are just not viable.
Well, then this cable test will be of little worth, nothing than an entertaining exercise.


Quote:
As for AD distortion, well I can test this by doing an ABX between a reference music sample and a DA/AD if the gain adjusted files are detectably different then I will drop that idea.
That sounds fine, but even if you come to the conclusion that the difference is non-existent or negligible (which I would be surprised about), there's still a considerable probability that the essential signal content responsible for the cable differences has gone lost.

I'm sure it's no surprise to you that I count myself to the «cable believers», and as such I have made the experience that cable differences are gone after digitization. Honestly, I was a bit baffled about your invalidation of my test result just because it wasn't gained from a blind test. Do I need blinded eyes to not hear a difference? Just so much: I also couldn't see the cables in the other tests where I was able to detect the difference -- of course still knowing which was which.

Now you have so much fear of being biased by the look of the cables -- maybe justified under the circumstances you're describing. At the same time you measure them thoroughly with the conviction to not find a significant measuring difference, since cables are «passive devices». BTW, I know many passive devices that alter the sound, such as resistors, capacitors, headphones, loudspeakers...

For an unbiased approach it would be important to do listening tests before the measurings. Otherwise the test person is prone to be biased by the latter. That's why the more serious test magazines consequently stick to this order. You don't seem to care about that. That's another reason why -- contrary to my initial goodwill statement -- I don't believe anymore that your approach is entirely unbiased and non-ideological, also because of some of your interim reactions. You rather seem decided to make your test an «objectivist statement» of the well-known kind (maybe not even consciously). Sorry if I am completely wrong, that's just my now impression.

BTW, to discuss DBT, we're obviously in the wrong forum. Anyway, some more thoughts about DBT: To enable the highest likelihood for detecting audible differences, it is extremely important to know which sample you listen to at the very moment. Of course not the type of cable, but let's name it cable A and cable B. The other (often used) variant, random music samples, is virtually doomed to produce negative results. Without anchor points within the chaotic pattern of sample series the ears are overstrained.

It's not like two pictures you can compare each time you change your line of sight and with which you always perceive the same details and after a certain time are certainly able to detect the (possible) differences. With random music samples you get a constant change of signal shape, never will there be two identical sequences you can hear any time you like and compare any time you like. But the former method takes even more time than the latter, so I don't think it's a real option for you in the case you renounce digitization. And that's my main objection against your (planned) test array: It's by far not designed for maximum likelihood of positive results.
.
post #75 of 438
Cool picture myinitialsaredac.

Quote:
Originally Posted by JaZZ
BTW, to discuss DBT, we're obviously in the wrong forum.
At this point this thread is as scientific as the threads in sound science forum where DBT is allowed, a simple pm to a mod can get it moved.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › My cable test enterprise