News $2,500 Ethernet Switch Effectively Isolates Audiophiles From Cash

[khaakon]

Anyone else notice that these guys are also selling a $3000 1 meter ethernet cable and a $10000 power cable? The power cable actually lists "custom quantum tunneling" as a feature. That is some rich snake oil!

You can get 1(one!) power cable for 45.000 USD - 10x the price of a decent complete stereo system. That's for a 2,5m. - if you take the measly 1,25m, you get it for only 35k USD...
https://audiovisionsf.com/collections/nordost/products/nordost-odin-gold-power-cord
"The symbiotic nature of the electrical and mechanical characteristics demonstrated in this revolutionary power cord produces an overwhelming performance upgrade."

 

[calol]

I’m a networking engineer that works in a field that’s highly sensitive to radio interference and involves high voltages.

The TL/DR version:

This is product is designed for people with lots of money and no sense.


The longer version:

Let’s start with the plausible part, cheap switches use cheap power supplies. A cheap switch mode power supply can produce a lot of noise, that noise can be emitted for the device either via the AC (Power) line, which can affect devices which use the AC line, or by electromagnetic force which will affect devices nearby. An example of this would be take the input line to your amp (unplugging it at the remote end) and putting in near a cheap phone charger, you should hear a ‘humming’ noise. Any decent switch will have measures in place to filter the noise emitted both on to the AC line and RF environment. Any decent amp would have isolation/filtering on the AC line and environment. So really you would need to pair a bad switch with a bad amp to have a noticeable effect.

Copper Ethernet is based on transmitting and detecting a differentiated voltage signals across (multiple) twisted pairs. A switches brain, ASIC, doesn’t work with Base-T signals, instead there is usually a PHY (and isolation between the switch and the ethernet cable) which reduces noise, amplifies the signal, and converts it to something like SGMII for the ASIC which works in binary. It results in a very reliable way to transmit data, Bit Error Rates (BER) are typically in the range of 1E^8 – 1E^9 (Fibre is better, WiFi worse) even in noisy industrial environments. When an error does occur it can be detected, due to Ethernet having an error detection (CRC) per frame. If a switch detects an error the expect behaviour is to drop the frame.

For most streaming audio services, it goes across a WAN (Wide Area Network) this is typically where 95% of issues would come from rather than the LAN (Local Area Network.) Most last mile technologies such as Cable (Coax) DSL, FWA etc (basically any last mile technology other than fibre) have much high chance of introducing problems resulting dropped packets.

A dropped frame/packet would be bad for audio, right? Not necessarily, streaming services use a transport layer protocol like TCP, if you lose a packet TCP ensures that its retransmitted. That means that what is transmitted by the server is what is received without even a bit of error in the payload. (Audio isn't really that sensitive to a flipped bit (well as long as it’s not the MSB) but encryption, and almost everything is encrypted these days, is highly sensitive to a flipped bit. A flipped bit after decryption would result in a completely corrupted pay load. Even for real time audio in UDP things like congestion and jitter are way more important on quality.

So in summary, the network protocol stack (never mind any measure put in place by the application stack) has so many measures in place to ensure errors don’t happen, when they do they can be detected, and when they are detected the can be recovered to ensure what is sent is what is received.

Another thing of note, their website talks about the quality of materials used, features, reviews and subjective opinions, what it lacks is an empirical testing or evidence to show the benefits. It’s easy to test, there are lots of 3rd party labs that can test it, but they are not even claiming support for basic standards.

 

[Deleted member 14196]

I going to start with, I am no sound expert, not even by a long shot. However, music from streaming services is encoded digitally, transmitted digitally and then decoded into analog signals once it reaches its final destination device (phone, tablet, AV unit, TV, etc.). So how exactly does this do anything for that digital signal in transport? Seems like it would have little effect me based on what I know, but maybe someone can enlighten me.

it doesn't. this is pure SNAKE OIL

it's like buying gold plated printer cables because it make printing faster and betterer... hahahahah

and i prove it by noting the prices of the cables!! HAHAHAHA, only retards buy this stuff

 

[JamesAK]

Synergy Research has made an ethernet switch for audiophiles that smooths out electric signals in the switch, to improve audio quality.

$2,500 Ethernet Switch Effectively Isolates Audiophiles From Cash : Read more

An audiophile and his money are soon parted.

 

[TheOtherOne]

Hey you, yes YOU there! Wanna hear a joke? .... "Audiophiles" 🤣

 

[hannibal]

Audiophiles need $1000+ USB cable until it sounds good enough
So this was just waiting to happen!

 

[doolittle]

After reading this, I feel like saved a ton of $$ by getting one of these instead... That warm, fuzzy feeling 🤓

https://www.apc.com/shop/us/en/products/Line-R-1200VA-Automatic-Voltage-Regulator/P-LE1200

 

[BakedAlaska]

They are nothing but scam artists.
"appealing to only the most diehard of audiophiles. "
should read "appealing to only the most diehard of mugs. "
but seriously, these guys should be in jail, not conning rather gullible audiophiles.

 

[digitalgriffin]

Even streaming audio is BUFFERED to account for uneven packet delivery cadence which CANNOT be controlled. Snake oil to prey on those who have more money than brains like the product below.

Product Review (hometheaterhifi.com)

 

[digitalgriffin]

Well, some people have reported that there is in fact a difference. The question is only "where" this difference comes from. Of course you are correct and this does not come from affecting the digital data as is. Some ideas:

a.) Inside the switch for 2500 USD you could have a bunch of DSP logic, to recognize audio, unpack the network packets, modify the audio and repack it back. Maybe there is some typical fault in the streaming audio, which can be fixed algorithmically. For example:

• Increasing sampling frequency by 2x by using bandlimited (perfect) interpolation. Many audio receivers do 96kHz or more up-sampling before playback.
• Use a harmonizer to introduce higher harmonics from existing audio in to the new higher frequency range. (above16kHz)
• There could be any number of algorithms that can take the mp3 and aac faults and try to mitigate the problems and improve something.

Why to put that in to a switch rather than a special driver? Maybe its not possible to do on all devices?

b.) There is also this thing called "power supply". You can affect the performance of any device on the power grid, by plugging in one more device. Each new device and even every power plant which is on or off the grid will have an influence on the power supply and the frequency spectrum of the 50/60Hz signal. You can tell exactly which power plants are online and which devices in your houshold are working or not, just by looking at the power grid voltage signal. It is possible that some of the power supply issues (1 or 2 bits) could leak in to the D/A converter (the soundcard) of the computer, which is playing back the audio. The source of that can also be the ethernet connection.

This does not mean that the 2500 USD switch does any of this. As much as the internet is overwhelmed with false information, you could have hustlers selling anything.

Atmapuri

NO on both counts. 99% of all ethernet controllers use opto-isolated interfaces to prevent floating ground from damaging equipment. The hardware layer which is responsible for receiving the electrical signals and generating them is always isolated from the computers main equipment.

Cheap line protection uses line diodes which are horrid for signal integrity. If the packet is damaged it's a NAK response (typical) for TCP/IP and ignored for UDP. Higher quality use opto-isolated repeaters. But that requires a two separate isolated power planes and voltage boosters which drives up cost. That's why you see cheap low reverse voltage diodes (Schottyky diodes) on most isolators like UPS and surge strips.

The TCP header consists of 11 fields, of which only 10 are required. The eleventh field is optional and called “options”. The details of TCP header fields are

• Source port (16 bits) – identifies sending port
• Destination port (16 bits) – identifies receiving port
• Sequence number (32 bits) – Ensure correct sequencing of the arriving data
• Acknowledgment number (32 bits) – Next expected TCP octet.
• Reserved (4 bits) – Reserved for future use and set to zero
• Flags (8 bits) (or Control bits) – contains 8 1-bit flags
• Window (16 bits) – Number of bytes that receiver is currently willing to receive
• Checksum (16 bits) – Used for error-checking of the header and data
• Urgent pointer (16 bits) – Indicates the end of urgent data
• Options (Variable 0-320 bits, divisible by 32) – It’s length is determined by the data offset field. Options 0 and 1 are a single byte (8 bits) in length. The remaining options indicate the total length of the option (expressed in bytes) in the second byte.

 

[hotaru.hino]

Well, some people have reported that there is in fact a difference. The question is only "where" this difference comes from.

Unless they've measured the output of the signal before it hits a driver, any "difference" is likely subjective and influenced by the bias of "I bought expensive equipment, ergo, the audio quality is better"

 

[InvalidError]

NO on both counts. 99% of all ethernet controllers use opto-isolated interfaces to prevent floating ground from damaging equipment.

Apart from fiber, I have never seen Ethernet hardware using optical isolation.

Most copper Ethernet implementations use classic signal TRANSFORMER (magnetic) isolation while dodgy implementations use capacitive isolation instead.

 

[InvalidError]

Unless they've measured the output of the signal before it hits a driver, any "difference" is likely subjective and influenced by the bias of "I bought expensive equipment, ergo, the audio quality is better"

When you get into arguments like the "color", "fullness", "warmth", etc. of sound which are all caused by a particular audio setup's characteristic non-linearities (distortion) profile, you are indeed deep into personal preference territory. There is no winning an argument about customizing your own audio preferences.

 

[digitalgriffin]

Apart from fiber, I have never seen Ethernet hardware using optical isolation.

Most copper Ethernet implementations use classic signal TRANSFORMER (magnetic) isolation while dodgy implementations use capacitive isolation instead.

I always saw them use two separate power planes The signal transformer being one. The opto isolated section is built into the chips themselves (or so I thought). Heck all you really have to do is measure the voltage drop across the diode. You don't even need to fully isolate it. 1 = voltage drop across diode. 0 = no voltage drop.

 

[Deleted member 1353997]

This is some "highest quality pixels that anybody has seen" level of nonsense...

 

[hotaru.hino]

Also, while we're at it, I want to bring up this gem again: http://www.hisdigital.com/us/product2-388.shtml . I believe its retail price was $100.

 

[xyvyx]

It's comforting to see that geeks of a different ilk come to the same conclusions about this sort of BS. It's like anytime there's a complex, technical system where the average consumer isn't aware of how/why something works, it creates this market for snake-oil products to take advantage of them.

Anyway, if you want to see actual measurements of a device like this, it's been done:

UpTone Audio EtherREGEN Switch Review

This is a review and detailed measurement of UpTone Audio's Ethernet "audiophile" switch. It is on kind loan from a member. The EtherRegen costs US $640. The EtherRegen as the name indicates, "regenerates" and Ethernet signal. This is a functionality that is provided with any Ethernet switch...

www.audiosciencereview.com

 

[gabrielex]

The audiophile market is one of the richest in cows to milk.
They'd believe everything you throw at them and spend any amount of money no matter what.
While there can be some setups that actually do sound better than others, most of the times their price is totally unjustified, while in other cases, such as this and many other ones it's just pure scamming and taking advantage of the naiveness of some people.
Also if you don't really want interference (and I don't think it really matters that much on a digital file, while you should instead be careful about not having interference in your outputs), there's optical fiber, it is still expensive but no way near that, and in that case interference is not even a thing since it's, well...optic.

 

[InvalidError]

I always saw them use two separate power planes The signal transformer being one. The opto isolated section is built into the chips themselves (or so I thought). Heck all you really have to do is measure the voltage drop across the diode. You don't even need to fully isolate it. 1 = voltage drop across diode. 0 = no voltage drop.

There is no opto-isolation in PHY chips. Pairs terminate directly into the PHY chip's analog front end which contains differential amplifiers to cancel out the pairs' common-mode noise, a variable gain amplifier to normalize the signal, a low-pass filter to limit noise bandwidth and an equalization network to compensate for the bulk of expected cable-related issues. On 1GBase-T which uses all four pairs in full duplex, the AFE's output then goes into a DSP which handles echo cancellation, crosstalk mitigation, further equalization and decoding of the PAM-5 modulation.

You cannot detect and decode this with a diode.

 

[Jim90]

I like this title, it has long-life potential....let's see....

The 3080Ti Effectively Isolates Fools From Cash
The 3090Ti Effectively Isolates Fools From Cash

 

[digitalgriffin]

There is no opto-isolation in PHY chips. Pairs terminate directly into the PHY chip's analog front end which contains differential amplifiers to cancel out the pairs' common-mode noise, a variable gain amplifier to normalize the signal, a low-pass filter to limit noise bandwidth and an equalization network to compensate for the bulk of expected cable-related issues. On 1GBase-T which uses all four pairs in full duplex, the AFE's output then goes into a DSP which handles echo cancellation, crosstalk mitigation, further equalization and decoding of the PAM-5 modulation.

You cannot detect and decode this with a diode.

Well then I stand corrected.

 

[LuxZg]

Well, if you wrap a cable going from your output to speaker/headphones around normal switch vs thus switch, MAYBE you get a difference in sound ))

 

[jasonkaler]

I going to start with, I am no sound expert, not even by a long shot. However, music from streaming services is encoded digitally, transmitted digitally and then decoded into analog signals once it reaches its final destination device (phone, tablet, AV unit, TV, etc.). So how exactly does this do anything for that digital signal in transport? Seems like it would have little effect me based on what I know, but maybe someone can enlighten me.

Dude, paying more for something makes the music sound better.
The same as paying more for a meal makes it taste better.
Any apple fan will tell you this.

 

[blppt]

Never mind that most 'super high spending so-called audiophiles' that I know would swear off anything but vinyl anyways.

Or they'd only listen to 24/96 SACDs with their equipment air-gapped from any internet source. Definitely would not listen to super-compressed streaming audio.

 

[SKD46]

Snake oil??