Overview and conclusion
Contents
What a journey this was! It literally took months to gather the cables, set up the test methods and listen to the cables/interlinks. Then it was time to process the data and find the relationships. A tough, but interesting puzzle!
Was it worth the investment? Yes… definitely, we think. We feel strongly that we have at least already been able to get a little whiff of the measurability of cables. No: absolute quality is not measurable. As is “musicality. But we have been able to gain insight into other elements. And now we want more…!
Every cable just works
Let’s conclude the article by saying that every cable just works. We didn’t hear a really bad cable. Even the stock interlink still did certain things just fine, though its character proved a bit unpredictable and the overall feel of it is a bit restless.
Other than that standard €6 cable, we didn’t have any junk. Technically most cables were just well put together and you get value for money. So should you spend a fortune on a decent cable? No. Definitely not. There are very good cables for sale for €40. That is something we experienced during this test.
So is there no room for those esoteric, extremely expensive cables? We wouldn’t say that. There is a place for these products. After all, there are people who can and are willing to dig deep into their pockets to squeeze those last few percent out of the system. Just as there are people who want to buy high-quality tires for their car, like to cook with high-quality kitchen appliances or ride a top-of-the-line bicycle. And that’s just fine. No reason to condemn that.
All data in tables
We have put all the data in some graphs for your convenience. This makes it easier to see the relationships. These graphs show all relevant properties in a few overviews. We have also made a summary table of all the data, including the propagation variables.
What AI has to say
We have fed the data to an AI. This is what an AI has to say about the data:
Key Correlations:
Delay and Time Pulse to Reflection: There’s a strong positive correlation (0.94), suggesting that cables with higher delay times also tend to have longer times from pulse to reflection. This could indicate that longer delays in signal transmission correlate with increased times for signal reflections to become noticeable.
Delay and Propagation Variance: A moderate positive correlation (0.53) indicates that cables with higher delay times may also exhibit higher propagation variances. This suggests a relationship between the timing of signal transmission and the consistency of that timing across different signals or conditions.
Capacitance and Inductance: There’s a strong positive correlation (0.63), which is expected as these two properties are often related in how they impact the behavior of an electrical circuit or cable.
Capacitance and Peak Pulse Voltage: A strong negative correlation (-0.62) suggests that cables with higher capacitance tend to have lower peak pulse voltages. This could be due to the capacitance’s effect on damping the signal strength.
Inductance and Peak Pulse Voltage: A strong negative correlation (-0.72) further supports the notion that certain electrical characteristics (in this case, inductance) can significantly impact the peak voltages that cables can handle or transmit.
Conclusions:
The correlations observed suggest intricate relationships between the physical and electrical properties of the cables and their performance in signal transmission and handling.
Cables with specific electrical characteristics (like higher delay and propagation variance) might be more suitable for certain applications where timing consistency and signal delay are critical factors.
The combined analysis underscores the importance of considering multiple parameters when selecting cables for specific uses, as various factors (capacitance, inductance, delay, etc.) can influence performance in complex ways.
Summary video
Listening tests / Livestreams (purely for transparency)
You are absolutely crazy if you want to look at these livestreams, since they last for over five hours each. But for transparency reasons, we live streamed these listening sessions. But should you want to know how we did the listening tests, here’s where you get an insight into the process.
Conclusion
What a journey this was! It took months to collect, set up and listen to everything. But it's been worth it, we think.Let's conclude the article by saying that every cable just works. We didn't hear a really bad cable. Even the stock interlink still did certain things just fine, though its character proved a bit unpredictable and the overall feel of it is a bit restless.
Other than that standard €6 cable, we didn't have any junk. Technically most cables were just well put together and you get value for money. So should you spend a fortune on a decent cable? No. Definitely not. There are very good cables for sale for €40. That is something we experienced during this test.
So is there no room for those esoteric, extremely expensive cables? We wouldn't say that. There is a place for these products. After all, there are people who can and are willing to dig deep into their pockets to squeeze those last few percent out of the system. Just as there are people who want to buy high-quality tires for their car, like to cook with high-quality kitchen appliances or ride a top-of-the-line bicycle. And that's just fine. No reason to condemn that.
Thank you very much for this fantastic test. Do you think the results can be applied to XLR cables?
Yes
Good morning,
This was a fascinating read and most of it was understandable.
This inspired me to make up an interconnect cable using Sommer Tricone. Switching between this and a QED QNECT2 was night and day, my ears like the difference and this has told me that swapping interconnects is worth considering.
I am jumping onto your speaker cable test next.
Thanks!
I have a question. I think it’s a language thing.
In your measurement of “impedance”, is that actually the resistance value or equivalent of the looped conductors or is it the characteristic impedance of the cable? It doesn’t seem like the latter at all, which might be very interesting to understand.
Keep up the great work!
Impedance is complex resistance.
https://en.m.wikipedia.org/wiki/Impedance
That’s actually not my question.
There’s the loop resistance/impedance of the conductors and there’s also the characteristic impedance of the transmission line.
For example, there’s 50 Ohm coaxial cable. That’s the characteristic impedance. Zip cord has a characteristic impedance of around 90 Ohms. That’s all a combination of the inductance, capacitance, resistance of the wire, and so on.
https://en.wikipedia.org/wiki/Characteristic_impedance
This value rises considerably as you go below 100 KHz.
http://k9yc.com/TransLines-LowFreq.pdf
It is just the resistance. Not the characteristic impedance.
Great! Thanks!
I’m looking forward to your speaker cable tests.
Hopefully we will have it live at the end of april.
I did wonder. In a youtube video you talk about the HP pulse generator being 500MHz generating a 2ns pulse. At 500Mhz the wavelength is about 60cm, which is shorter than te length of the cable. So you enter transmission line theory territory, and the characteristic impedance comes into play.
In the video you state that channel 1 is terminated at 50 Ohm and channel 2 at 1 Mohm (which could be considered open ended).
And in between channel 1 and 2 is the cable., with its characteristic impedance for transmission line situations, which 500MHz is I think. So every cable that does not have a characteristic impedance of 50 Ohms will cause reflections. It does not matter if the cable is good or not, it is just caused by the transmission line impedance mismatch.
The characteristic impedance of the TPR is given as 110 Ohms, which causes the reflections. I cannot find this value for the SQM, but if that would be 50 Ohm, this would explain its non-reflection-behaviour in the test.
But in audio frequency territory, this characteristic impedance is of no importance and it does not say anything about the quality of the cable.
Just a few thoughts that kept me busy 🙂 Love to hear your view on this.
Hi F.Vcp.
You are right. That’s why we didn’t do all tests on those frequencies, but did a series of tests to see what is important and what isn’t important. The ‘reflection-test’ was purely for propagation speed and -variance. Not to analyze impulse behaviour or anything. For obvious reasons, we needed a frequency that gave a shorter wavelength, for we needed to isolate the puse.
Jaap, first very well done. Probably the only way to do this better would have been if you had a reel of the cable 1000M long to determine better what the cable attributes are at 1M.
But I was talking to a good friend physics about this. Just like component differences, I think the sonic difference has to do with the designers. I always say this at a show, throw 10 designers in a room with the same parts and you get 10 different products.
In the cable realm you have metal, dielectric, insulation, wind, layout, shielding and directionality…. all these designers have their own take on what is important and that is why cables sound different.
Plus look at the variables you have on either end!
Thanks again, really great stuff!
Gordon
Hey Gordon,
I think you are right. And very nice analogy!
Hey all, there was/is an interesting exchange on this test and its results over on the Dutch side. Jaap has given me permission to use the translation I produced (Apple Translate) and repost it here for all to see. What I love about it is that this is exactly the kind of helpful result a researcher wants to see. Good research typically should generate more questions than answers, especially in its early stages. That’s normal and how the process and what it examines becomes more refined.
============================
April 6 2024 — Posted by Ad Braam (translation from Dutch) with Jaap’s responses following each question.
Jaap, upon further study I still have a number of questions (sometimes for confirmation) to be able to understand the research well. First of all, it is clear to me that the intention is to investigate whether a relationship can be found between the measurements and the listening tests and not so much to determine what is a good interlink (in my opinion very dependent on the rest of the audio system).
1. The ‘explanation of measurements’ says ‘Ideal there is no difference’. Is it meant: Ideal is there no difference when the voltage is increased from 0.3 to 1.5 volts?
A: There is ideally no difference in propagation when the voltage changes.
2. Does the voltage in an analog interlink vary between 0.3 and 1.5 volts?
A: A line output works at 2 volts at RCA. But of course the tensions vary with the music. I chose these 3 voltages to gain insight into small and large differences in amplitude.
3. Is the propagation time the difference in time it takes the pulse to go through the cable?
A: No. The propagation time is the time it takes pulse to go through the cable. So sometimes there is a difference in that when the amplitude changes.
4. How is it that the pulse that goes through the cable (green) is higher than the reference pulse (yellow)?
A: This has to do with the closure impedance, among other things. On the yellow that is 50 Ohm. On the green that’s 1 MOhm. That to imitate an exit / entrance.
5. As for noise measurements. Is it possible to explain the ‘spectral view’ in more detail. What do the green and brown lines mean? Because of this I don’t understand the statements between vd Hul and Grimm.
A: Spectral view is not noise measurement. This one displays decay. Often: extinction after a signal. Explaining this goes too far for a comment section.
6. What is the explanation that at lower frequencies the induction and capacity swings like this?
A: I don’t know exactly either. It has to do with wavelengths, among other things, but I also have to dive into this.
7. It is confusing that the vertical scale in the messenger plot is not the same when comparing Ricable and Mogami. I didn’t understand the conclusion at first.
A: You can’t do anything about that. The software scales itself and that is also necessary to keep it readable.
8. Propagation variance table very interesting. Do I understand correctly that ‘gets faster’ means that with increasing tension the propagation time becomes smaller?
A: Yes
9. Apart from the conclusions listed under the table, can you also conclude that a ‘good’ cable should have a low variance?
A: Yes. Less variation is better. Ideally it is 0.
There are a lot of questions, but hopefully I’m not the only one who has these kinds of questions.
END OF REPOST
it’s pure nonsense, what about the cheap connectors and simple wires inside a unit and i’d like to see these kind of things done by a university, not by a annoying guy which is close to the source……and all those terms:the highs are too promonent,but vague……for example, how do you know it’s not by bad wiring instead of the interlink. As an engineer micro-electronics i can say that you’re an indiot if you buy overexpensive interlinks, you’d better buy a more expensive amp
Thank you for the feedback.
Dear Richard. How do you explain the difference that is evident from the measurements of the cables?
Well, I’m a retired university researcher of 30+ years and I don’t have a problem with what these guys are doing. Is it the rigorous and pure scientific method backed up by statistical analysis? No. But I can tell you that this kind of casual but thoughtful and carefully-managed research, performed within their admitted and openly public constraints, has led to more complex and sophisticated follow-up by others quite often. After all, it is the repeated casual observations (yes, including the subjective) of others with sincere and deep interest in a subject that stimulate the start of funded research in a wide variety of fields. I would suggest to others who have the background to be helpful with ideas, money, and/or time instead of throwing bricks. We’re adults, not children. Carry on gentlemen, you’re doing fine.
Thanks Alan!
👍 I couldn’t think of a better response
I wonder if it would be possible to feed all of this tabular measurements, graphs, and subjective observations into an AI model and ask what characteristics are most correlated since it is so complex. Awesome effort guys. I can’t think of anyone else having attempted this publically. Mahalo!
Thank you! Maybe… Hmmm… Interesting thought.
Agreed
Is it fair to consider that the composite values shown in the All Data table, when coupled with the subjective observations, could hint at a kind of value proposition when MSRP is taken into account? You know…..loosely. Have to factor in the rationality of those doing the listening of course. 😉
I have looked at that as well, but haven’t found a logical link…
Regarding the rationality of the listeners or that the lowest composite value+observations:cost? Both equally…challenging I suspect. But I think there is some guidance that could be drawn or inferred from the latter.
I really tried to find a link between what we hear and what the data in ‘All Data’ says. What I did notice is that too high a capacitance or inductance is definately not good. Cables that show a decent distribution between the two mostly show a good sound balance.
The thing is: the data in that table doesn’t show the frequency response, nor the spectrum behaviour. And those two really gave insight in the performance of the cable. Along with the temporal behaviour (propagation variance).
There are other variables affecting quality. So its only part of the picture then. Not enough to guide a purchase decision. And, in the end, that’s also going to come down to buyer motivations, which is another can of worms. So I see what you’re saying. Still, this data provides a place from which you or others can build in the measurable areas. Valuable nonetheless. Well done.
Thank you. But the freq response and spectrum measurements are in this article. The thing is: they are not a simple value I can put in a table.
Good work 😉
I wish you would have done the tower and evergreen rca from Audioquest 😉
Hence my question : if someone is on a budget and need a coaxial cable , I usually advice an Audioquest forest as they are great for the money.
Is there an rca cable that you recommand also around 50€ ? Audioquest tower and evergreen would be the equivalent of the forest coaxial but they’re completely different type of cable 😉
Thanks !
(The rca are for a cd player and a rega io )
Ready the article. There are cables around €50 and less in there.
First, BIG shoutout to you guys for having the cojones to take this on. Just one look at the data sheet….HUGE undertaking. I have to think that this is a significant contribution to audiophiles everywhere who have been wondering. Perhaps it will stimulate follow-up on the part of others even. That’s what research is all about. Secondly, am I missing something? The data summary doesn’t show any values, Impedance-Inductance-Capacitance, for the Transparent Plus, yet it seems to show in its individual graphs. I’m sure you didn’t miss something so obvious, so can you explain to me? Thanks.
Hey Alan!
Thanks! I also hope that others will pick this up and make some progress. It’s needed :-).
The Transparent is sort of a weird cable. The values of impedance, capacitance and inductance are way higher due to the filter-network. If I would put them in the graphs, the scale would be messed up. That’s why I left it out.
Ah. Got it. Makes sense. Thx.