The thing with audiophile networking equipment... where are the proofs?

[jabbr]

3 hours ago, PeterG said:

 

Of course, if someone showed you "proof" that a certain product was better, but you could not hear a difference, I hope you would ignore the "proof"...  I see you are a brand new member.  If you surf around this site you'll find countless debates on the general difference between people who describe themselves as objectivists and people who describe themselves as subjectivists.  Whatever way you think about this stuff, I encourage you to take an empirical approach and listen before you commit to big $--borrow gear from your local dealer before you buy, or buy from someone with a generous return policy.  Good luck!

 

Since we are talking specifically about network equipment here and not audio equipment it is a really strange concept to think that a network should have a "SQ". In my mind the best network has **no sound**. The highest quality commercial networking equipment has been engineered to physical standards which are simply out of the range of the capabilities of any audiophile company. That said we don't need 800Gbe at home nor for audio. My own **listening tests** identified common ground loops as well as common mode noise transmitted through copper ethernet cables as something that *can* affect "SQ'. This is trivially eliminated by using fiberoptic ethernet.

 

There is no need to buy network gear from a local audio dealer!!!

[jabbr]

7 hours ago, PeterG said:

"Need" is a funny word in audiophiledom, isn't it?  There is really no need to spend more than $2-3K on a stereo that one could buy in an hour or so--see @DuckToller's very good thread on this.  But of course, once the bug bites...we NEED better sound.

 

I do not stream because every time I have listened to a streamer, I have preferred local files.  Maybe I need to try more expensive streamers or better networking?

 

To be very clear. and I've spent considerable amount of time looking into this as well as listening, the quality of my equipment, lets say my 100Gbe Mellanox switch, QSFP modules, cables ... essentially my entire network ... is very simply unmatched by any "audiophile" network equipment in existence. Its not even close. I do grant that endpoints and equipment that sits on the audio rack should be engineered to the requirements of audio.

 

Now since you don't stream, you are commenting on an aspect for which you admit that you don't have direct knowledge. I haven't "listened" to every piece of network equipment but I can say with a high level of confidence that the single most glaring **reduction** in SQ from a network is by transmission of common-mode noise as well as ground loops (similar things) via Ethernet cables. Once you go above the very cheap network equipment that is in the common home network, and move to 10Gbe which at this point in time is almost legacy, then things like network jitter is irrelevant. 

 

Regarding "streaming", the SQ from HQPlayer streamed to a low power NAA (which sits in the audio shelf) is nothing short of amazing. The ability of the modern network to definitively isolate the high noise/high powered workstation needed to process HQPlayer at its highest level from the low noise audio area is also nothing short of amazing. I say this from both a technical as well as vast listening perspective.

[jabbr]

On 10/23/2022 at 10:40 AM, AnotherSpin said:

 

This whole love of music thing is very confusing indeed. For example, some people like Wagner and others prefer Die Antwoord. All this is strictly opinion based with total absence of any scientific method and slightest attempt to really prove any point, instead of preaching 😎

 

I will say that the role of the network is not to provide of subjective SQ/color such as Wagner vs Die Antwood, rather the role of the network is to provide a perfectly black background for the music to come through. This is what a faulty or substandard network sounds like: a fuzz or haze -- at the very worst a buzzing (ground loop). A good network will have a perfectly black background: no coloration! 

 

Those things are both audible and measurable (since this is Objecti-Fi)!

[jabbr]

1 hour ago, firedog said:

One: all of that is very expensive and very difficult to actually test properly. That's why it's almost never done for audio.

Two: it probably won't actually make much of a difference. People who "believe in their ears" will continue to do so, even if testing and evidence say they are wrong. Those who tend to accept objective testing will continue to accept it vs. sighted listening. 

Thus the endless arguments about it on audio forums, even when such testing is occasionally done.

 

The thing with pro-grade network equipment ie 10Gbe ethernet and faster, is that **all** of it is required to be tested and not pass network jitter, so you can hear a difference, and know that its been compliance tested in a very stringent fashion.

[jabbr]

1 hour ago, semente said:

 

I disagree. There is a universal definition for better performance: accurate reproduction of the recorded signal.

 

Better-sounding, on the other hand, depends on the listener's preference. It's subjective and thus not universal.

 

This is very true! The SS vs tube amp debate is drastically different than when we are discussing network gear. Many expert amp designers understand that various odd vs even harmonics can provide enjoying sonic coloration. Ultimately this is about listening pleasure. 

 

Network equipment do not have a role in tuning SQ. Any sound introduced to the system by a piece of network gear is a fault.

[jabbr]

16 hours ago, Jud said:

 

I do try to go for stuff that measures well, when relevant measurements are available.  There's a frustrating lack of really top-level scientific information correlating much of this to what we can actually perceive. Not surprising - labs with fMRI and the most sensitive equipment test rigs (if such a lab even exists) trying for grant money aren't inclined to say "Let's propose an experiment involving consumer audio equipment, I bet it'll get funded!"  But such measurements along with other information like repair data can at the very least indicate something about quality of design and execution.

 

We are admittedly interested in what can be heard yet networking equipment is far far away from the actual audio circuitry.

Similarly there are only a few ways that networking gear is physically possible to affect the audio circuitry. At some point, and this applies to networking equipment, not amplifiers and certainly not speakers, we can rely on good old fashioned engineering. While there are good old fashioned reasons why different brands of resistors, capacitors, tubes and transistors can indeed alter “SQ” and also are indeed measurably different in their behavior, this is just not the case for networking equipment.

 

 

16 hours ago, Jud said:

 

Where measurements aren't available or where they don't conclusively favor one course or other, I suppose we're into the realm of risk tolerance and disposable income.  How much are you willing to drop on whatever you think sounds good (assuming the claims for the item lie somewhere within the bounds of physical reality)?

 

 

The bounds of physical reality is the salient point. The very and measurably best network switches in terms of vanishingly low jitter and noise are not designed for audio. I will say this for everything including the switch and everything like NAS and servers that sit behind the switch. Everything forward to the switch from endpoint through speakers/headphones is fair game for being the best in terms of audio specific engineering.

[jabbr]

46 minutes ago, Jud said:

I do wonder about very practical things like power supply noise getting into the rest of the system, or system grounding/topology creating noise loops. (I had a noise loop once where the network equipment was involved. An isolation transformer helped but didn't totally resolve the problem. What did at the time was feeding the system from a battery powered laptop locally rather than the network. I eventually made other changes that allowed me to run the system from the network without issue.)

 

Again, as we presumably agree, good measurements in relevant aspects of network performance can generally indicate a level of design and build quality.

 

If the switch is copper, and especially if metal RJ-11 plugs with shielded cables are used, then there is the possibility for the copper cable to transmit both common mode noise and participate in ground loops with the endpoint I.e. inject noise into the audio area.

 

When the switch is fiber then there is **no** electrical coupling of the network cable to the audio system. Cheap wall-warts near the audio system can certainly inject common mode noise via the AC input to the audio area just as can your fridge so AC iso transformers should be used in all cases. I put the networking hardware, and NAS and servers in my basement so they are physically isolated from my audio areas and there is no downside to long fiber cables. So yeah you can power isolate your audio area and I’d use a clean PSU for anything supplied by the audio power circuit.

 

Like you, I’ve also had noise loops associated with network equipment which was my very motivation for using fiber in the first place.

[jabbr]

37 minutes ago, yamamoto2002 said:

When reviewers test those products, I hope they add test vector to measure acoustic noise characteristics. I'd like to know which one is the best performing (most quiet) high-res capable network equipment. Also I'd like to know the most quiet computer monitor and laptop for home high-res recording.

 

My network equipment is located in my basement which has poured concrete walls, and so acoustically quite isolated from my audio areas. 

[jabbr]

1 hour ago, Jud said:

I do wonder about very practical things like power supply noise getting into the rest of the system, or system grounding/topology creating noise loops.

A friend’s house lost a lot of electrical equipment after a lightning storm, possibly via the cable line in the house. The modem was literally fried.

[jabbr]

14 hours ago, Jud said:

 

That part I've got covered.

Yeah I know ;)

 

Video screens are crazy bad at giving off EMI. My first really bad experience came when I was using a fairly good pro-quality digital recording device that I normally USB powered off my MacBook for recording music events. I was using this as double duty DAC for my headphones… and after installed a new HDMI monitor at my workstation … bzzzzzz … and that led me down the isolation rabbit hole