Cable Pathways Between Audio Components Can Affect Perceived Sound Quality

[March Audio]

On 6/11/2021 at 7:24 AM, PeterSt said:

 

100%.

But all I know from design is that differential is mainly a pain in the *ss implying more THD than less with SE.

Nicely controversial, right ?

100 pages to discuss ...

 

PS: Strange claim: If you know what can be achieved with balanced, you can achieve the same with SE just by pushing (yourself) harder. For example, common mode noise can also be diminished by eliminating noise in the first place.

So soo many ways to Rome ...

Not controversial, just simply wrong.

[March Audio]

To be pedantic (sorry), as mentioned, wires will generate (Johnson) noise, but this is dependant upon temperature and resistance.  In a cable with a resistance of fractions of an ohm this is totally negligible.  Triboelectric noise can be generated by the movement of conductor insulators from vibration or bending.  This generates an electric charge.  A non issue in line level low impedance interconnects, but high impedance sources can be an issue.  Such as electric guitar pickups/cables.  Flick them and you may hear a noise.  Its usually combatted by adding an additional partially conductive layer around the central conductor insulator.

 

This can be a real PITA when doing things like using charge output accelerometers for vibration measurement.  Special, very tightly wound cable and minimising cable vibration helps.  I have very direct experience of this from jet engine testing.

 

However, as I mentioned, in normal low impedance line level interconnects both are a non issue.

[March Audio]

I have just had a quick squiz through the paper so I may well have missed pertinent things, so dont shoot me, but I raise the same concerns as others have already stated.  

 

1. Comparing different types of cable i.e. RCA and XLR.  The cables behave fundamentally differently with respect to noise rejection.  The electronics driving each, balanced V single ended perform differently WRT noise and distortion.  RCA will be more susceptible to noise currents which can be flowing in the low signal conductor (shield).

 

2.  The noise measurement was performed incorrectly.  It was driven with a 100 ohm source impedance (quite representative) however the far end was only terminated by the oscilloscope which will have a very high input impedance of typically 1Mohm.  This is not representative of real world amplifiers which may typical range from 10kohms to 100kohms.

Secondly scopes are generally useless for measuring noise.  The one used has an 8 bit resolution.  Its own noise floor is way to high.

 

[March Audio]

38 minutes ago, jventer said:

Let's make it simple: he was trying to see if people can reliably hear differences between 2 sets of interconnects. His paper say they can.

 

The only other possible variable in the system is/was what the dac does with outputs via balanced and single ended. The author assumed / measured certain things: 

"This DAC’s total distortion (all products) at the present output level is specified at -120 dB of full scale. 10 This DAC’s absolute voltages at the left and right RCA outputs and the XLR inverted and non-inverted outputs (pins 3 and 2 relative to pin 1) are equal within a standard deviation of 0.13% (i.e., 0.011 dB). The six voltages were measured at the pins while playing a -12 dB (25.12% of full scale) 1 kHz sinusoidal tone from a 16 bit, 44.1 kHz wave file.

 

I am not experienced with this dac or qualified enough to know what other measurements should /could have been done to measure that the outputs were the same?

 

Assuming the outputs were the same, the author proved that people can hear differences in interconnects.

 

If we go off on tangents about the quaility of the interconnects, their cost, their lengths etc., which may all influence the why people heard a difference, it still does not refute the fact that the majority of the people in this experience reliably heard a difference.

 

(The paper do raise many further questions re larger sample size, measuring of noise etc., but it does point in the direction that further tests are necessary to fully understand how our ears and brains work.) 

 

I hear what you say but the problem with the paper is that it is testing a system of components which will behave fundamentally different in the two configurations.  

 

As such it isn't testing just the cables, it's testing the whole system. 

[March Audio]

On 6/10/2021 at 9:55 AM, The Computer Audiophile said:

I still have an open mind about it but more questions the more I read. 
 

I found it very interesting he used Spectral amplification. Spectral “requires” using MIT cables and has ultra wide bandwidth. Great stuff, but perhaps not idea for such a test?

 

Would you say he really tested for a difference between a single ended route and balanced route, rather than a difference in cables per say?

Well a radio receiver has the potential to suffer problems other amps may not.  I was surprised to see they even admit to this in their own literature:

 "To prevent oscillations in the Spectral system, conventional "wire" cables have been replaced with precision tuned cable "networks" developed by MIT. The modular terminations of the MIT cables tune them to function like ideal low-pass filters"

 

So I agree it was a dubious choice of amplifier for the test.  That's not a criticism of spectral, just whether it was appropriate for the test as it has special requirements to keep it stable. Requirements the tested cables would not meet.

 

 

 

I would possibly suggest that he tested for neither because it's difficult to differentiate between the effect of the balanced v single ended electronics compared to the effect of balanced v single ended cable.