ABX proponents are using it all the time to "prove" there is no difference between, amps, cables, you name it....
I have nothing against blind testing. Just make sure it is valid.
dave
Given that i've seen the same implications made by others (practicing statiticians) a number of times (and given that i can follow the statistics -- Honours BSc Statistical Mathematics UofA 1976 -- even thou i have not practised).
And if you follow his reference...
Even SY (out of context it losses a bit, but it is on the same subject):
dave
I'm always making an ash of myself.
Anyway, I'm outta here in a few hours, heading over to ETF. All y'all behave yourselves, OK?
Anyway, I'm outta here in a few hours, heading over to ETF. All y'all behave yourselves, OK?
Agreed, I praised Olive's protocols elsewhere many times. It's the suitability of his targets for making blanket statements about the limits of audibility I question. Or, "what's preference got to do with it?"
Audio / Video Cables Defined
Cables in an audio system definetly change how the system will sound for the good and bad. For an example, we take an audio system consisting of a Sonic Frontiers tube cd player, (tubed with Mullard or other high quality NOS tubes) line preamplifier and finally a very linear power amplifier such as a Mark Levinson solid state design driving ProAc speakers:
with all of these specific components without cables alone, you will obtain a very neutral linear sound (which is a very very nice sound i might add). As soon as you put cables in this system linking these components together you will obtain a new sonic signature which depending on the cables (brands, technology of the dielectric being used, impedance, capacitance, solid core or stranded, shielding, types of metal and so on) is not apart of the original presentation. Having said that, even with some expensive cables there will be an alter of the original signal resulting in a very bright presentation that is very fatigue to the human ear and especially at critical listening volumes or simply can sound very muddy and fat.. bloated mid-range and unsustainable low-frequencies (bass). Now there are 3 major properties and many other minor properties regarding what a good cable will consist of for low-level signals that need shielding and extra care (preamp level):
1) length of the cable
2) shielding
3) impedance
The general rule of thumb when dealing with interconnects is to keep them short and well shielded so little damage is done to the original signal. Alot of people tend to believe that guage is an important factor in cables, it is because when running long lengths the resistance of the cable grows resulting in loss.. but that is mainly for speaker cables which have large amounts of current traveling through them that would require a heavier guage at a certain distance. To be honest, I find RG-6 to work very well in many systems for the low level signal portion because of its mega shielding and solid core wire in a 75 ohm configuration which elimanates the "skin effect" compared to multi stranded cheap cables. For the system above, I would recommend something more expensive in the lines of Audioquest for their excellent design goals and topologies used to construct cabeling with minimal loss as possible. I would not recommend cables such as transparent for this type of audio system simply because they use a network which is located in a shielded box around the cable that is designed to elimanate extra noise which in return adds a dull type of sound that clearly is not apart of the original signal. Using these cables on a system that sounds very bright (screetchy highs, metal dome tweeters, sybulance.. when female vocals sound like a bunch of cutlery hitting the floor... you get the picture) will make it sound more pleasable. But in the end, beware of the marketing portion of audio... it is very DICEIVING. 60% of the time even in specialty high-end shops selling gear that is as much as a down payment on a house may not be all what it is cranked up to be. When looking to buy a new component, make sure that you research the design and carefully plan out what components will be used with that particular device as some components have different sonic signatures which will result in something that was definetly not worth the hype, money or time.
The 2 major factors for speaker cables is to have a sufficiently large guage and to have good conductors. There are many more minor factors but in most cases only make a difference when being used in a very detailed high end system where every slight change in cables, type of tubes being used, biasing.. etc has a completely new effect on the final outcome of what it will sound like. Think of it as a fine tuned instrument such as a Yamaha Grand Piano where every fine change in that piano will make a difference.
One last note about cables, do not be fooled when dealing with DIGITAL AUDIO or VIDEO as expensive cables have no effect on what the outcome is under any circumstance (that is why DIGITAL was invented 🙂. Spending alot of money on an HDMI cable may be necessary because you might not get any output at all because of it being digital. There are some cases though when running 1080p through 50 feet of cheap HDMI cable will show sparkles on the screen due to the receiver (TV) indicating borderline complete loss, so basically where digital is concerned it either works or it simply does NOT. There is no "in between" with it. Just make sure that the cables are well shielded and maintain the proper impedance for the application. Use RG-6 for digital audio and analog video.
Cheers!
Cables in an audio system definetly change how the system will sound for the good and bad. For an example, we take an audio system consisting of a Sonic Frontiers tube cd player, (tubed with Mullard or other high quality NOS tubes) line preamplifier and finally a very linear power amplifier such as a Mark Levinson solid state design driving ProAc speakers:
with all of these specific components without cables alone, you will obtain a very neutral linear sound (which is a very very nice sound i might add). As soon as you put cables in this system linking these components together you will obtain a new sonic signature which depending on the cables (brands, technology of the dielectric being used, impedance, capacitance, solid core or stranded, shielding, types of metal and so on) is not apart of the original presentation. Having said that, even with some expensive cables there will be an alter of the original signal resulting in a very bright presentation that is very fatigue to the human ear and especially at critical listening volumes or simply can sound very muddy and fat.. bloated mid-range and unsustainable low-frequencies (bass). Now there are 3 major properties and many other minor properties regarding what a good cable will consist of for low-level signals that need shielding and extra care (preamp level):
1) length of the cable
2) shielding
3) impedance
The general rule of thumb when dealing with interconnects is to keep them short and well shielded so little damage is done to the original signal. Alot of people tend to believe that guage is an important factor in cables, it is because when running long lengths the resistance of the cable grows resulting in loss.. but that is mainly for speaker cables which have large amounts of current traveling through them that would require a heavier guage at a certain distance. To be honest, I find RG-6 to work very well in many systems for the low level signal portion because of its mega shielding and solid core wire in a 75 ohm configuration which elimanates the "skin effect" compared to multi stranded cheap cables. For the system above, I would recommend something more expensive in the lines of Audioquest for their excellent design goals and topologies used to construct cabeling with minimal loss as possible. I would not recommend cables such as transparent for this type of audio system simply because they use a network which is located in a shielded box around the cable that is designed to elimanate extra noise which in return adds a dull type of sound that clearly is not apart of the original signal. Using these cables on a system that sounds very bright (screetchy highs, metal dome tweeters, sybulance.. when female vocals sound like a bunch of cutlery hitting the floor... you get the picture) will make it sound more pleasable. But in the end, beware of the marketing portion of audio... it is very DICEIVING. 60% of the time even in specialty high-end shops selling gear that is as much as a down payment on a house may not be all what it is cranked up to be. When looking to buy a new component, make sure that you research the design and carefully plan out what components will be used with that particular device as some components have different sonic signatures which will result in something that was definetly not worth the hype, money or time.
The 2 major factors for speaker cables is to have a sufficiently large guage and to have good conductors. There are many more minor factors but in most cases only make a difference when being used in a very detailed high end system where every slight change in cables, type of tubes being used, biasing.. etc has a completely new effect on the final outcome of what it will sound like. Think of it as a fine tuned instrument such as a Yamaha Grand Piano where every fine change in that piano will make a difference.
One last note about cables, do not be fooled when dealing with DIGITAL AUDIO or VIDEO as expensive cables have no effect on what the outcome is under any circumstance (that is why DIGITAL was invented 🙂. Spending alot of money on an HDMI cable may be necessary because you might not get any output at all because of it being digital. There are some cases though when running 1080p through 50 feet of cheap HDMI cable will show sparkles on the screen due to the receiver (TV) indicating borderline complete loss, so basically where digital is concerned it either works or it simply does NOT. There is no "in between" with it. Just make sure that the cables are well shielded and maintain the proper impedance for the application. Use RG-6 for digital audio and analog video.
Cheers!
Cables in an audio system definetly change how the system will sound for the good and bad. For an example, we take an audio system consisting of a Sonic Frontiers tube cd player, (tubed with Mullard or other high quality NOS tubes) line preamplifier and finally a very linear power amplifier such as a Mark Levinson solid state design driving ProAc speakers:[snip]Cheers!
Of course with many tubed components cables may influence the sound, as much tube stuff has a rather high Zout which is sensitive to capacitive and other loading effects. So if you want to demo cables differences select your equipment with care.😉
But it then gets messy to find a component and cable that work good enough together to enjoy it. I'd prefer a component that is sensibly engineered not to be affected by any but the most pathological cables....
jd
Yeah, a tube power amp (for example) with a damping factor of 0.5 will be hugely influenced by 100mR of cable Z.
Les Leventhal has published several articles in the JAES:
Type 1 and Type 2 Errors in the Statistical Anaylsis of Listening Tests
Les Leventhal, JAES, Vol. 34, No. 6, 1986 June
Statistically Significant Poor Performance in Listening Tests
Les Leventhal, JAES, Vol. 42, No.7/8, 1994 July/August
Analyzing Listening Tests with the Directional Two-Tailed Test
Les Leventhal, Cam-Loi Huynh, JAES, Vol.44, No. 10, Oktober
A shorter letter from Leventhal, some discussions and comments from Clark and others could be find at the stereophile site:
Stereophile: The Highs & Lows of Double-Blind Testing
That is a truely honest point of view.
But it is really surprising that reasonable attempts to refine test protocols, attempts to avoid common traps in testing were not appreciated by this "entity" .
Normally it should be very clear, that a positive test result doesn´t mean anything, if it can´t be shown, that it was only the EUT that led to the positive result.
On the other hand, a negative test result doesn´t mean anything if no positive controls (on sufficient sensitivity levels) were used.
The whole point in testing is that the results could be generalized in some ways, otherwise the effort is useless.
Wishes
Last edited:
That´s why i find the approach of researchers like Julian Dunn so refreshing.
A lot of people were complaining about the sound quality of the interfaces, someone tried to investigate it and voilá there it is....
I´m sure among the group of listeners complaining were people in whose hearing (and other abilities) Dunn had some confidence and that´s probably the reason why he don´t insisted in a proper dbt on ´interface sound´ before starting any research.
As posted before, it would be much better, if both ´camps´ would work together in refining methods and tests instead of disparaging each other.
Wishes
I couldn't read those papers in detail (time is short!), but in skimming through, it didn't appear that he actually experimentally verified that a) differences were actually heard, and b) eliminating the measurable differences caused the sonic differences to go away. Is that correct?
That is correct.
As posted before audibility of jitter effects is still controversial....
Up to now, i´ve read only three papers trying to do some listening tests on this subject.
Wishes
As posted before audibility of jitter effects is still controversial....
Up to now, i´ve read only three papers trying to do some listening tests on this subject.
Wishes
Jacob, I am sure you also read some of the comments on those articles, e.g.
Comments on "Type 1 and Type 2 Errors in the Statistical Analysis of Listening Tests" and Author's Replies
Authors: Shanefield, Daniel; Clark, David; Nousaine, Tom; Leventhal, Les
JAES Volume 35 Issue 7/8 pp. 567-572; July 1987
I'm quoting this only because I couldn't put it better myself. The core of these comments is:
It's essentially a very interesting (and mathematically solid) way to refine ABX test results, by including a probability of validity of the tested hypothesis.
Out of this approach, prof. Meyer et. al. had probably the best late approach on such an issue, while ABX testing the differences between CD and SACD formats. After a negative result, he never said they are identical; all he stated is that if there are any differences, they are to small to be identified/heared. And that from now on, the burden of proving any differences is on those making the claim.
This approach is valid for a hypothesis that could be (theoretically) investigated, if the results were positive. But when the tested hypothesis is (if true) on a direct collision course with the fundamental laws of physics (and here's my favourite example, the ByBeee device) then the upfront burden to provide positive testing results is on those making the extraordinary claims. Why do you think nobody is testing today pepetuum mobile designs?
Otherwise, I have begged about a million times for a valid/good/workable example of what you call a "positive control" in ABX audio testing. All I got was silence.
Theoretical and Audible Effects of Jitter on Digital Audio Quality
Authors: Benjamin, Eric; Gannon, Benjamin
Affiliation: Dolby Laboratories Inc., San Francisco, CA
AES Convention:105 (September 1998) Paper Number:4826
Kaoru Ashihara1 et al., Detection threshold for distortions due to jitter on digital audio
Acoust. Sci. & Tech. 26, 1 (2005)
Tomoharu Ishikawa, Yukio Kobayashi, Makoto Miyahara, “Improving the Transfer Function of a Sound
System to Constant and its Effect on the Reproduction of High Order Sensations ”, WESTPRAC VII, 175-1,
Wishes
pp.393-396, 2000.
- Status
- Not open for further replies.