was it done blind?
End of cable debate - snake oil
- Thread starter anon_bb
- Start date
was it done blind?
RobHolt
Moderator
- Joined
- Oct 23, 2004
- Messages
- 4,126
- Reaction score
- 12
They will be flat to at least 20khz and will sound the same.
I wasn't suggesting it wasn't explicable, but it shows how hard you have to work to get a significant change. Twice as much LCR in an interconnect gives you 0.05dB dulling at 10k. Roughly.
Repeating the exercise with a different 1m interconnect, silver, not coax, about as different as reasonably possible, shows no difference on the graph. More interestingly two runs with the same cable done some hours apart are more different. So given reasonable input and output impedances the two 1m interconnects are 'electrically equivalent'. But most 'believers' would suggest they would sound different.
That's the challenge.
Paul
Repeating the exercise with a different 1m interconnect, silver, not coax, about as different as reasonably possible, shows no difference on the graph. More interestingly two runs with the same cable done some hours apart are more different. So given reasonable input and output impedances the two 1m interconnects are 'electrically equivalent'. But most 'believers' would suggest they would sound different.
That's the challenge.
Paul
RobHolt
Moderator
- Joined
- Oct 23, 2004
- Messages
- 4,126
- Reaction score
- 12
To put that into perspective, just turning your head slightly when listening will change the response up at 10khz by several db.
Even the slight difference in your seated position every time you listen is likely to result in much greater than 0.05db difference between sessions.
- Joined
- Mar 23, 2006
- Messages
- 418
- Reaction score
- 0
BBV. Yes.
- Joined
- Mar 23, 2006
- Messages
- 418
- Reaction score
- 0
BBV, It wasn't me that organised & carried it out. It was Paul Miller for HiFi Choice. He did have a fairly consistent methodogy which he applied whether it was cables, cd players or amplifiers under scrutiny. It was pretty comprehensively documented each time they published a group test he'd undertaken. I'm sure if you can find a copy of Choice from the early 90's you'll get it chapter & verse. They were carried out blind though. The listeners did not know what they were listening to or what price it was. They always re-introduced items again during the test to check for consistency of opinion. I look back with some horror to the tedium of those days!
Other than that, email him & ask him about the exact protocol used. He's the editor of HiFi News now so isn't exactly difficult to get hold of.
They don't do such thorough blind testing now. Its just too expensive to organise.
Other than that, email him & ask him about the exact protocol used. He's the editor of HiFi News now so isn't exactly difficult to get hold of.
They don't do such thorough blind testing now. Its just too expensive to organise.
Last edited by a moderator:
Rmax = minimum loudspeaker impedance * 0.02
If loop resistance for a given length of loudspeaker cable exceeds Rmax then power loss may be audible e.g. greater than 0.1dB. Or, the cable may have an effect on heard sound when compared to another cable where loop resistance is less than Rmax.
Thus two equivalent loudspeaker cables would both have a loop resistance less than or equal to Rmax. The exception to this would be where the reactive components of two cables are sufficiently dissimilar (usually owing to conductor geometry).
The debate, as I see it, is not that all loudspeaker cables sound the same (they don't). It's the stuff about metallurgy, dielectrics and thighs of vestal virgins producing gob-dropping effects.
Paul Ranson wrote:
Pre-fuggin-cisely
- Joined
- Mar 23, 2006
- Messages
- 418
- Reaction score
- 0
That was preciesly the debate .....!
**************************...as mosfet so rightly says various speaker cables don't sound the same. I often feel they should it would make life a lot simpler......as someone suggested its the other bits that make cables sound different ....try mechanically damping a set ....Theres credible reports that cables move when big amps are switched on....also some cables can be demonstrated to be micro phonic though I've not seen the demonstration.
**************************...as mosfet so rightly says various speaker cables don't sound the same. I often feel they should it would make life a lot simpler......as someone suggested its the other bits that make cables sound different ....try mechanically damping a set ....Theres credible reports that cables move when big amps are switched on....also some cables can be demonstrated to be micro phonic though I've not seen the demonstration.
Last edited by a moderator:
bottleneck
talks a load of rubbish
It preciesly, or precisely wasn't.....!
It was that electrically similar cables will sound the same (given they are not badly made...e.g. no shielding etc)
I don't think anyone has ever argued that it is impossible to have two cables sounding different to each other when they measure differently.
It was that electrically similar cables will sound the same (given they are not badly made...e.g. no shielding etc)
I don't think anyone has ever argued that it is impossible to have two cables sounding different to each other when they measure differently.
lordsummit
moderate mod
Oh for ****'s sake.
Can we not use language guaranteed to wind other people up. See the other thread on speaker cables.
I'm either going to throw all the cable threads together and let you fight it out Lord of the Flies style, or ban discussion of cables altogether.
FFS have you lot got nothing better to do
Can we not use language guaranteed to wind other people up. See the other thread on speaker cables.
I'm either going to throw all the cable threads together and let you fight it out Lord of the Flies style, or ban discussion of cables altogether.
FFS have you lot got nothing better to do
- Joined
- Mar 23, 2006
- Messages
- 418
- Reaction score
- 0
Does this mean that if I take a 5m set of QED 79 strand cable and then take another 5m set of QED 79 strand cable (that measures exactly the same) and listen to them both, then they'll both sound the same?
I'm amazed.
I'm amazed.
You're beginning to get it Murray.
Now follow the thought through. If a cable doesn't sound the same as QED79 strand then it is probably wrong.
Paul
Now follow the thought through. If a cable doesn't sound the same as QED79 strand then it is probably wrong.
Paul
- Joined
- Mar 23, 2006
- Messages
- 418
- Reaction score
- 0
Paul. If a cable doesn't sound the same as QED79 strand then it almost certainly isn't QED 79 strand.
This is the tech panel that Paul Miller wrote accompanying the cable test mentioned above. It does suggest that the issue is a little more complex than some believe.
By the way anyone who knows Paul Miller will confirm that he's a fairly serious minded individual (understatement) not given to accommodating notions of Foo. I think he originally studied Chemistry at ICL. Can't recall if he has a PhD.
Do also bear in mind that this was written for publication in a consumer magazine and not as an AES paper. Magazines wouldn't sell very well if written in that way. You may take issue with some of what he says & if so you should email him & pick him up on these things. I'm sure you'll get a forthright response! At least he has done the measurements and has conducted published blind tests which few, if any, here can claim to have done.
Interconnects (as reviewed in last month's supplement) can be either symmetric or asymmetric in geometry, then there's certainly less variety in the design of speaker cables, almost all of which are symmetric. Many adopt the classic figure-of -eight construction with spaced conductors to reduce capacitance. However, as the spacing between these parallel conductors increases, so does the cable's series inductance.
The choice of stranding, however, remains a principal distinction between different speaker cables. Conventional thick, multi-stranded cable wil employ many hundreds or thousands of fine strands combined in either a bunched, concentric or rope-lay style weave. The end result is a low resistance cable offering minimal power loss over long runs.
However, such cables are at the mercy of the Skin Effect and non-Iinear inter-strand conduction. Only by separating individual strands within a conductor are these problems resolved. Litz cables are the ultimate example where every fine strand is individually insulated. Capacitance might go through the roof but at least the side-effects of Skin Depth are kept at bay. (If you are unfamiliar with terms such as Skin Effect then please refer to the glossary at the back of this booklet)
Some purists advocate the use of single-strand solid-core cable. This technique adheres most closely to the ideals of pure non-interactive conductIon but can, over progressively longer lengths, be hampered by high resistance. One or two cables use single cores of a substantial size (even hollow cores) as a way of combatting both the Skin Effect and inter-strand conduction.
A sensible compromise between high capacitance, high resistance and Skin-related distortion is achieved by deploying just a few strands of moderate thickness per conductor each of which is insulated or otherwise isolated from its nearest neighbour. As a final touch some of these cables are also equipped with an electrostatic screen, grounded at the amplifier-end.
The performance of any cable is dictated by the choice of materials and geometry decided upon by the manufacturer. Subjective and technical performance are both influenced, though the link between the two is still pretty tenuous. At the most elementary Level it's possible to link the construction of the cable with its Resistance (R), Capacitance (C) and Inductance (L). Nevertheless you'll achieve very little correlation with sound quality by matching the RCL characteristics of the cable with, say, the output impedance of the source or load impedance of the speaker. In practice the nature and purity of the conductors, conductor stranding, dielectric stability of the insulator, physical geometry, characteristic impedance, RF performance, screening and mechanical rigidity of the finished cable all play vital roles. If all those terms sound like greek to you then don't panic, as these key topics crop-up time and again in the cable reviews, we thought it might be nice to explain what they all mean.
Glossary
Resistance
This term is measured in ohms or mohms (thousandths of an ohm) and is a measure of how, quite literally, the cable 'resists' the flow of direct electrical current. Consequently the resistance of a cable is linked to the choice of metal, thickness and overall length of the conductors rather than their exact geometry. Interconnect cables that have a high-ish resistance can actually degrade treble quality, particularly if earth-currents are circulating between one poorly- grounded piece of equipment and another. The higher its series-resistance then the higher will be the voltage generated along its length. And obviously this spurious voltage simply contributes to the overall (predominantly high frequency) level of noise. Furthermore a high series- resistance will exacerbate dielectric absorption, a nonlinear storage of charge in the adjacent insulating layer. This is particularly relevant if the cable employs a relatively poor dilectric such as PVC. An insulator with a low surface energy, like PTFE will be less susceptible. Otherwise a high series resistance actually helps damp-down the Q at resonance and encourages a wider bandpass.
High resistance in speaker cables acts in tandem with the amplifier's output impedance to reduce loudspeaker damping. This, in turn, can modify the response of the speaker in line with its varying impedance curve. Remember, although a speaker is nominally rated at either eight or 4ohms, in reality its impedance can swing up and down with frequency, as you can see from the plots we provide along with our loudspeaker reviews.
Impedance and Characteristic Impedance
This is the cable's opposition to the flow of alternating electrical current (aka music) and is a composite of the cable's resistance, capacitance and inductance.
The impedance of a cable will increase at higher frequencies as its reactance (a further term introduced by the cable's series inductance) is added to the overall resistance.
At higher frequencies still a cable should be viewed as a transmission line with resistance, capacitance and inductance terms distributed along its length and contributing to a so-called characteristic impedance.
Capacitance and Leakage
Measured in Farads or pF (one UK billionth of a Farad), the parallel capacitance of a cable is a measure of the impedance between signal and return conductors at very high frequencies. The higher the capacitance then the lower the parallel impedance (often measured as leakage) which, partIcularly at very high frequencies, might indicate an instability or breakdown in the dielectric insulator) separating both conductors. In our tests leakage is measured as the DC resistance between unterminated signal and return conductors and is expressed in Mohm (millions of ohms). Cable with a low figure for leakage will either have its signal and return cores in very close proximity or will be employing a poor dielectric (or, in the worst possible case, both).
Inductance
The series inductance of a cable adds to its impedance at high frequencies. lnterconnect cables with a high capacitance and inductance can be more sensitive to the RF garbage that's emitted from many CD players, leading to a more fatiguing sound that's both coarse and grainy as a result of extended RF intermodulation.
Skin Depth and Inter-Strand Conduction
Due to electromagnetic induction, signal current is not evenly distributed through thc entire cross-sectional area of the conductor. Instead a progressively higher current density is encountered towards the surface (the Skin) of the conductor. This effect, called the Skin Effect, increases in significance with increasing frequency.
The Skin Depth, meanwhile, refers to the point (measured from the surface) where the current density falls by 63 per cent. This is relevant to us because the impedance of the conductor increases where the current distribution is lower, that is towards the centre of the conductor. Hence the distribution of impurities and crystal grain boundanes through the conductor will have a varying influence depending on the frequency of the signal. For example, the Skin Depth in a copper conductor is some
0.467mm at 20kHz but only 0.039mm at 2.8224MHz (a frequency often seen in a CD's digital output).
So a digital interconnect cable would require conductors less than 4Oum in thickness for the RF current to be distributed evenly throughout! It is thinking along these lines that has prompted the solid-core aficionados to opt for 0.6-0.7mm strands in their audio cables. However, the repercussions of the Skin Effect are more insidious and complex in a conventional multi-strand cable.
This is the tech panel that Paul Miller wrote accompanying the cable test mentioned above. It does suggest that the issue is a little more complex than some believe.
By the way anyone who knows Paul Miller will confirm that he's a fairly serious minded individual (understatement) not given to accommodating notions of Foo. I think he originally studied Chemistry at ICL. Can't recall if he has a PhD.
Do also bear in mind that this was written for publication in a consumer magazine and not as an AES paper. Magazines wouldn't sell very well if written in that way. You may take issue with some of what he says & if so you should email him & pick him up on these things. I'm sure you'll get a forthright response! At least he has done the measurements and has conducted published blind tests which few, if any, here can claim to have done.
Interconnects (as reviewed in last month's supplement) can be either symmetric or asymmetric in geometry, then there's certainly less variety in the design of speaker cables, almost all of which are symmetric. Many adopt the classic figure-of -eight construction with spaced conductors to reduce capacitance. However, as the spacing between these parallel conductors increases, so does the cable's series inductance.
The choice of stranding, however, remains a principal distinction between different speaker cables. Conventional thick, multi-stranded cable wil employ many hundreds or thousands of fine strands combined in either a bunched, concentric or rope-lay style weave. The end result is a low resistance cable offering minimal power loss over long runs.
However, such cables are at the mercy of the Skin Effect and non-Iinear inter-strand conduction. Only by separating individual strands within a conductor are these problems resolved. Litz cables are the ultimate example where every fine strand is individually insulated. Capacitance might go through the roof but at least the side-effects of Skin Depth are kept at bay. (If you are unfamiliar with terms such as Skin Effect then please refer to the glossary at the back of this booklet)
Some purists advocate the use of single-strand solid-core cable. This technique adheres most closely to the ideals of pure non-interactive conductIon but can, over progressively longer lengths, be hampered by high resistance. One or two cables use single cores of a substantial size (even hollow cores) as a way of combatting both the Skin Effect and inter-strand conduction.
A sensible compromise between high capacitance, high resistance and Skin-related distortion is achieved by deploying just a few strands of moderate thickness per conductor each of which is insulated or otherwise isolated from its nearest neighbour. As a final touch some of these cables are also equipped with an electrostatic screen, grounded at the amplifier-end.
The performance of any cable is dictated by the choice of materials and geometry decided upon by the manufacturer. Subjective and technical performance are both influenced, though the link between the two is still pretty tenuous. At the most elementary Level it's possible to link the construction of the cable with its Resistance (R), Capacitance (C) and Inductance (L). Nevertheless you'll achieve very little correlation with sound quality by matching the RCL characteristics of the cable with, say, the output impedance of the source or load impedance of the speaker. In practice the nature and purity of the conductors, conductor stranding, dielectric stability of the insulator, physical geometry, characteristic impedance, RF performance, screening and mechanical rigidity of the finished cable all play vital roles. If all those terms sound like greek to you then don't panic, as these key topics crop-up time and again in the cable reviews, we thought it might be nice to explain what they all mean.
Glossary
Resistance
This term is measured in ohms or mohms (thousandths of an ohm) and is a measure of how, quite literally, the cable 'resists' the flow of direct electrical current. Consequently the resistance of a cable is linked to the choice of metal, thickness and overall length of the conductors rather than their exact geometry. Interconnect cables that have a high-ish resistance can actually degrade treble quality, particularly if earth-currents are circulating between one poorly- grounded piece of equipment and another. The higher its series-resistance then the higher will be the voltage generated along its length. And obviously this spurious voltage simply contributes to the overall (predominantly high frequency) level of noise. Furthermore a high series- resistance will exacerbate dielectric absorption, a nonlinear storage of charge in the adjacent insulating layer. This is particularly relevant if the cable employs a relatively poor dilectric such as PVC. An insulator with a low surface energy, like PTFE will be less susceptible. Otherwise a high series resistance actually helps damp-down the Q at resonance and encourages a wider bandpass.
High resistance in speaker cables acts in tandem with the amplifier's output impedance to reduce loudspeaker damping. This, in turn, can modify the response of the speaker in line with its varying impedance curve. Remember, although a speaker is nominally rated at either eight or 4ohms, in reality its impedance can swing up and down with frequency, as you can see from the plots we provide along with our loudspeaker reviews.
Impedance and Characteristic Impedance
This is the cable's opposition to the flow of alternating electrical current (aka music) and is a composite of the cable's resistance, capacitance and inductance.
The impedance of a cable will increase at higher frequencies as its reactance (a further term introduced by the cable's series inductance) is added to the overall resistance.
At higher frequencies still a cable should be viewed as a transmission line with resistance, capacitance and inductance terms distributed along its length and contributing to a so-called characteristic impedance.
Capacitance and Leakage
Measured in Farads or pF (one UK billionth of a Farad), the parallel capacitance of a cable is a measure of the impedance between signal and return conductors at very high frequencies. The higher the capacitance then the lower the parallel impedance (often measured as leakage) which, partIcularly at very high frequencies, might indicate an instability or breakdown in the dielectric insulator) separating both conductors. In our tests leakage is measured as the DC resistance between unterminated signal and return conductors and is expressed in Mohm (millions of ohms). Cable with a low figure for leakage will either have its signal and return cores in very close proximity or will be employing a poor dielectric (or, in the worst possible case, both).
Inductance
The series inductance of a cable adds to its impedance at high frequencies. lnterconnect cables with a high capacitance and inductance can be more sensitive to the RF garbage that's emitted from many CD players, leading to a more fatiguing sound that's both coarse and grainy as a result of extended RF intermodulation.
Skin Depth and Inter-Strand Conduction
Due to electromagnetic induction, signal current is not evenly distributed through thc entire cross-sectional area of the conductor. Instead a progressively higher current density is encountered towards the surface (the Skin) of the conductor. This effect, called the Skin Effect, increases in significance with increasing frequency.
The Skin Depth, meanwhile, refers to the point (measured from the surface) where the current density falls by 63 per cent. This is relevant to us because the impedance of the conductor increases where the current distribution is lower, that is towards the centre of the conductor. Hence the distribution of impurities and crystal grain boundanes through the conductor will have a varying influence depending on the frequency of the signal. For example, the Skin Depth in a copper conductor is some
0.467mm at 20kHz but only 0.039mm at 2.8224MHz (a frequency often seen in a CD's digital output).
So a digital interconnect cable would require conductors less than 4Oum in thickness for the RF current to be distributed evenly throughout! It is thinking along these lines that has prompted the solid-core aficionados to opt for 0.6-0.7mm strands in their audio cables. However, the repercussions of the Skin Effect are more insidious and complex in a conventional multi-strand cable.
Last edited by a moderator:
What an annoying piece of writing. It's a bit wrong almost everywhere.
I'd be more interested in reading the listening test report.
For what appears to be a thorough analysis of the common audio cable geometries/materials try Jim LeSurf's web pages. I think he even writes for HFN.
http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/audio/Analog.html
Paul
I'd be more interested in reading the listening test report.
For what appears to be a thorough analysis of the common audio cable geometries/materials try Jim LeSurf's web pages. I think he even writes for HFN.
http://www.st-andrews.ac.uk/~www_pa/Scots_Guide/audio/Analog.html
Paul
- Joined
- Mar 23, 2006
- Messages
- 418
- Reaction score
- 0
You ought to let him know Paul! [email protected]
Be sure to post his response here. I shall be waiting to read it with interest. You perhaps ought to also tell Naim, Arcam and Chord not to use the QC/measurement systems he supplied them as they may also be based on similarly flawed assumptions.
Perhaps you also should apply for his job. You obviously know far more about all of this than he does!
Be sure to post his response here. I shall be waiting to read it with interest. You perhaps ought to also tell Naim, Arcam and Chord not to use the QC/measurement systems he supplied them as they may also be based on similarly flawed assumptions.
Perhaps you also should apply for his job. You obviously know far more about all of this than he does!
Last edited by a moderator:
Murray, just to start,
But how you get from the value of this piece of writing to the value of his 'QC/measurement system' is obscure. I assume he is catering for his target markets in both cases.
As I said I'm more interested in the results of the blind listening test.
Paul
The skin effect is demonstrably insignificant, see the reference above, and 'non-linear inter-strand conduction' is undefined but it sounds good in a typical cable bollocks sort of way.
But how you get from the value of this piece of writing to the value of his 'QC/measurement system' is obscure. I assume he is catering for his target markets in both cases.
As I said I'm more interested in the results of the blind listening test.
Paul
- Joined
- Mar 23, 2006
- Messages
- 418
- Reaction score
- 0
His email address is there Paul. Take it up with him. It shouldn't take long to show him where he's gone wrong. I'll be waiting to read his reply.
Cheers,
Murray
Cheers,
Murray
