Do interconnect cables make a difference? A null test result

[RayDunzl]

Like this (posted here, along with other measurements):

What I meant was, "How does a log sweep recorded through the different cables look when compared in DeltaWave?"

What do they do to a "simple" signal, compared to what's seen with the complex signals above...

 

[solderdude]

Sorry Paul, I did not get that the -42dB null was the original file vs reproduced + recorded file.
Makes perfect sense.
The cable differences themselves are much smaller and in the range of the measurement limit of the ADC-DAC loop (as can be expected) but there are small differences none the less that I would consider as inaudibly small as they are only level differences that are so small they would not impact the actual amplitudes between input and output.

What is the source resistance, cable resistance and input load resistance and capacitance and the differences between the cables ?
That would be a lot harder to determine with the same accuracy.

To put this in perspective, as these appear to be level differences.
As the null is -50dB (there about) in the middle frequencies and -80dB at higher frequencies we can calculate that at -50dB the level difference is 0.3%
When this is converted to dB's again this is 0.026dB level difference for the middle frequencies and 0.0008dB for higher frequencies ?

This would not worry me in the least. moving your head a fraction of a mm would result in far greater deviations.

 

[Blumlein 88]

Here is one of the cable comparisons I've run. March Dac feeding my Zen Tour ADC so clocks aren't locked. One cable is a 10 ft RGB cable which looks to be RG6 coax. The other is 1 meter of Audioquest Diamond Hyperlitz. I'm showing the difference in frequency response between the two. Notice the scale. You do get some uneven results past 20 khz because there so little signal to determine response with up there. So you get almost .005 db variance right at the top and that's mostly the influence of noise upon the very low levels in the music at those frequencies.

 

[Blumlein 88]

For comparison here is a Zen Tour loopback. Both files using 6ft of inexpensive TRS cable. Run a half hour apart. Different musical track in this case. Ribbon microphones so less energy above about 12 khz. That lets the frequency response as determined by the software be bothered a bit more by noise at the upper frequencies.

 

[solderdude]

When I did a null with a long cable I could either get a null in the upper treble or in the lower frequencies but always ended up with some difference heard (either only upper treble or rolled-off upper treble. I could not get 'total silence'.
Back in those days I recorded the nulls as well but tape hiss would get in the way when nulling cables or circuits.

The reason for the above is the delay time combined with the wavelength. The phase alters relatively more at igher frequencies/shorter wavelengths and with nulling phase differences show as amplitude differences so the plot makes sense to me.

0.05dB roll-off is well below levels that would worry me. This is nowhere near 'bass levels/quality improved'/'smoother treble' reports etc. which all require substantial differences (>0.5dB) to become anywhere near audible.
It is quite predicatable that differences between cables exist.
The question is HOW audible it is and if goldeneared people can detect it and if so then what the relation would be with the perceived differences they report.

But it is a good thing that at least it is measuarble.
Now burn-in and cryo treated or not should be provable as well ... unless differences are well below the measurable limits.
There is a reason why none of these effects are ever backed up with (independent) repeatable research showing the benefits.
Explanations are always anecdotal and 'theoretical'.

Thanks Paul (great software b.t.w.) and Dennis for the measurements.

 

[Blumlein 88]

And just for reference here is a Zen Tour Loopback vs the original file. The scale is different, but a little droop around 18 hz and just a little droop in the treble.

 

[MediumRare]

Would a simple test be for @amirm to test the same DAC with two different interconnects to his AP and show the differences, just like he shows the differences between RCA and balanced output?

 

[pkane]

Would a simple test be for @amirm to test the same DAC with two different interconnects to his AP and show the differences, just like he shows the differences between RCA and balanced output?

I don't see why not.

 

[GXAlan]

This is a great thread worth bumping back up.

What is interesting is the -40 dB range between the raw digital file and DAC/ADC combo. I am assuming that’s the PK Metric, which accounts for what you can actually hear.

We cannot hear the digital file, but this would suggest an more accurate DAC and ADC should get you to -60 or -70 dB.

In that case, what would happen if you compared the sound of two different DACs?

 

[pkane]

This is a great thread worth bumping back up.

What is interesting is the -40 dB range between the raw digital file and DAC/ADC combo. I am assuming that’s the PK Metric, which accounts for what you can actually hear.

We cannot hear the digital file, but this would suggest an more accurate DAC and ADC should get you to -60 or -70 dB.

In that case, what would happen if you compared the sound of two different DACs?

Right, the AD/DA filters in the loopback path increase the amplitude of the difference, but mostly in the inaudible regions. PKMetric didn't exist at the time of that test, but was one of the reasons I thought it was worthwhile to create it.

Some of the better devices, such as RME ADI-2 Pro FS and E1DA ADC will also reduce the error significantly compared to the Apogee interface I used for that test. DeltaWave can also correct for the filter frequency response differences, which reduces the error even further.

 

[GXAlan]

Ah. It would be interesting to do PKMetric of digital file against a recording….

I did that for my SACD tests versus the DSD file and recall it being well into the audible threshold. Just have been so busy this Winter to do more hobbyist testing!