I have earlier posted about Audio trends and snake oil. What annoyed then and still today in marketing and media that too often today then talking on hi-fi, science is replaced by bizarre belief structures and marketing fluff. It seems that there too many snake oil rubbish products marketed in the name of hifi. It is irritating to watch the stupid people in the world be fooled with things like exotic materials, directional cables, thousand dollar IEC power cables, and all that rubbish. “If you pay ridiculous money for these cable you will hear a difference, but don’t expect your friends to”
I can agree that in analogue interconnect cables there are few cases where better cables can result in cleaner sound. And there are very many cases where there is no objectively noticeable difference.
In digital interconnect cables story is different than on those analogue interconnect cables. Generally in digital interconnect cables the communication either works, does not work or sometimes work unreliably. The digital cable either gets the bits to the other end or not, it does not magically alter the sound that goes through the cable.
But this digital interconnect cables characteristics has not stopped hifi marketers to make very expensive cable products that are marketed with unbelievable claims. Ethernet has come to audio world, so there are hifi audiophile Ethernet cables that do not make sense to me. With Ethernet the data either gets through the cable without any changes to it, or it does not get through at all. Ethernet has checksum on every data packets to detect for any errors (which are rare) in the transmission and all the packets that have data changed in any any way are discarded. So Ethernet cable can not not magically slightly alter the digital sound that goes through the cable.
Here are links to two articles on such expensive audiophile Ethernet cables:
Is streaming cable more or less expensive than $1000 audiophile snake-oil ethernet cable ?
https://audiobacon.net/2019/11/02/the-jcat-signature-lan-a-1000-ethernet-cable/
Gallery: We tear apart a $340 audiophile Ethernet cable and look inside
https://arstechnica.com/gadgets/2015/07/gallery-we-tear-apart-a-340-audiophile-ethernet-cable-and-look-inside/
Apart from the absurdly high price tag and more mechanically robust connectors, I see nothing special on those cables for carrying the data signal. I suppose those does look quite attractive to some users and this makes them willing to pay the high price.
Despite the fact that some people desperately want there to be audiophile Ethernet cables, there simply is no such thing. The IEEE standards do not include a superset of specifications that make a regular Cat-7 cable into an “audiophile” Cat-7 cable. If you still believe those “audiophile Ethernet” cables sounding better, please inform yourself how an Ethernet and Ethernet cable works.
133 Comments
Tomi Engdahl says:
There’s also this by Rod Elliott, if you want something slightly easier to digest: https://sound-au.com/articles/capacitors.htm
Tomi Engdahl says:
https://www.facebook.com/groups/DIYAudio/permalink/5489072734491861/
All technically correct. Must assume modern capacitors are made differently than they were 40+ years ago. Time for a reevaluation?
So true actually. Most caps these days are used for power supply bypassing and filtering. The quality of ones that are in the signal path are far better than ones 40 years ago. Also the materials used today are less diverse and packaging has changed as well. Still all valid info in the article though, and worth the read!
Modern capacitors work the same today as they did in 1980. This paper covers mature and proven technology that will never need a reevaluation. The paper is difficult to read and hard to understand for an audio amateur, DIY Audio homebrewer, and laymen. Perhaps it could be transcribed and re written without all the hi math in a way so that more people could understand, benefit, and apply it.
.
Skip the math and technicals unless you are designing a circuit. His discussions are quite clear without them and apply now as then to the currently available cap offerings.
The most important takes: Most of the discussions about which cap is better for what are foobah based. Electrolytics for power supplies, foil rather than filp caps for speakers (far more current capacity), poly caps are not high frequency limited,
Ok. Very good Larry.
I might add mylar polyester caps for small signal op amp and transistor audio circuits. And, never ever use the old “ceramic disk” caps in or near any audio circuit.
Tomi Engdahl says:
Jakosuotimet, aktiivi vs. passiivi
https://foorumi.hifiharrastajat.org/index.php?threads/jakosuotimet-aktiivi-vs-passiivi.135109/
Tomi Engdahl says:
Not directly related to the cable, but in your case I’d check the output resistance of the phono stage, to make sure it can drive such long cables without sigbificant high frequency loss…
many Le Pacific Phono Preamp circuit plans I found with Google seem to have quite high output impedance, around 2 kohm typically. That high output impedance can easily cause attenuation of high frequencies due cable capacitance and cable picks electrical noise more easily than some other designs. A preamplifier that is designed to properly handle longer cables (low output impedance, preferably balanced) is much less sensitive to cable related issues.
http://diyaudioprojects.com/Solid/JFET-Phono-Preamplifier-Kit/diy-jfet-riaa-phono-preamp.html
http://www.klaus-boening.de/le%20pacific.htm
http://www.next.gr/circuits/Le-Pacific-JFET-Phono-Preamp-for-MM-and-MC-Cartridges-l24644.html
Tomi Engdahl says:
http://www.zinamp.co.uk/articles/TTCabling.html
Tomi Engdahl says:
They even tried to get Linus to be an audiophile.
If you dont watch it, this network switch improves audio quality.
It’s time for some hard truth
https://m.youtube.com/watch?v=NMFQ3YvR3Eo
Tomi Engdahl says:
Audiofiles are the new flat earthers.
Tomi Engdahl says:
$800.00 for a network switch that costs $30.00. They don’t even rebadge them. They claim this ETHERNET NETWORKING SWITCH immediately improves audio quality.
https://aqvox.de/index_en.html
Tomi Engdahl says:
Audiophile logic – the more it costs, the better it sounds.
Tomi Engdahl says:
It’s time for some hard truth
https://www.youtube.com/watch?v=NMFQ3YvR3Eo
Every industry has its fair share of people and companies aiming to perfect the craft… but that aim has lead to a number of shady snake oil solutions that don’t actually improve anything. Today, we look at one of those.
CHAPTERS
—————————————————
0:00 Intro
1:41 What is this thing?
2:57 Their claims
6:14 Testing their claims
11:09 Teardown
15:04 Why it doesn’t work
17:39 Why people are still buying it
Tomi Engdahl says:
https://www.facebook.com/groups/avdisasters/permalink/5508445685950564/
So this guy is doing audiophile network switch tests, and banned me for saying either 0101011 get from a to b or they don’t
https://www.facebook.com/MMHiFi
Tomi Engdahl says:
An electrical study of single-ended analog interconnect cables
https://www.iosrjournals.org/iosr-jece/papers/Vol.%2016%20Issue%206/Ser-1/E1606014053.pdf
Abstract:The influence of interconnecting cables on an audio system’s performance is a controversial issue.
This is partly because commonly measured parameters such as resistance, reactance, frequency response, and
common distortions do not show meaningful differences. Thepresent electrical studyof line-level single-ended
(unbalanced) analog interconnects, provides a more comprehensive picture surpassing the common
specifications. It was found that uncommon time-domain effectssuch as reflection sequences and non-ideal
capacitive behavior, along with noise, can better distinguish the electrical performance of interconnects of
different grades.
Key Words:High-end audio; High fidelity; Hifi; Hi-fi; Cables; Interconnects; Psychoacoustics
Tomi Engdahl says:
Ediscreation Firebird LPS & Silent Switch OCXO Extreme Review.
https://www.monoandstereo.com/2023/02/ediscreation-firebird-lps-silent-switch.html?m=1
In my latest review, I dive deeper into the world of linear power supplies and Ethernet switches and uncover what’s special and unique about Ediscreation Firebird LPS and Silent Switch OCXO Extreme.
Ediscreation was founded in 2014 by Edison Wong, an electrical engineer who devotes all his attention to making audio products.
Ediscreation audio components are 100% proudly manufactured in Hong Kong
Ediscreation considers the LPS (linear power supply) to be the most important part of any audio system, regardless of what type of components it is paired with. Be it a phono preamplifier, DAC, NAS, switch hub, router, external hard drive, etc.
A well-designed LPS should deliver a significant improvement in sound quality. On the contrary, most power supplies that are usually provided are switching mode (SMPS) or cheap iron core-based power supplies.
Due to cost constraints, the noise and ripple of these power supplies are also very high, which directly affects the sound quality and background noise. It also affects the stability of the device.
Ediscreation LPS is designed to improve the above-mentioned issues.
The Silent OCXO Extreme Switch was conceptualized with one clear goal in mind: to improve the quality of music streaming for any streamer, music server, NAS, DAC, digital front-end, etc.
The Ediscreation team has completely revamped the structure of the network switch and revised it for audiophile use, assuring extremely low noise floor and high precision when connecting network signals.
Silent OCXO Extreme Switch implements a linear power supply with discrete circuits for the main power supply and an ultra-low noise LDO for the OCXO clock supply.
Under the hood is what Ediscreation claims is the world’s first network ground isolation switch that can greatly improve router performance and lower ground noise and also has a built-in OCXO system clock for ultra-low jitter 0.25-0.35ps with ±10 ppb phase frequency stability.
Silent OCXO Extreme Switch also comes with audio-grade LAN jacks for the best signal connection that can host any large-sized LAN cable.
The extreme version adds Furutech IEC inlets, FI -06 NCF, Audio Note Cap, Mundorf Cap, red copper base, UPOCC interconnect cable, and Silver Bybee.
Operational
The hubris associated with switching power supplies is obvious, and for that reason alone my reference listening room has a separate power line (even from the substation) to avoid anything connected to the power grid altering the electricity.
In my travels and visiting various high-end audio systems and with my personal experiences, not only the chargers for phones, and routers, but also LED lighting of all kinds cause and introduces massive pollution to the power grid, not to mention the devices that are connected to the high-end audio system.
The other issue is Ethernet. In the past, when analog sources were prevalent or digital sources were either connected to physical media or streaming was on the rise, we did not have (and didn’t know yet) to deal with the complexity associated with today’s digital audio streaming.
A few years ago, the topic of audiophile Ethernet switches would have surely triggered a heated reaction and questioned sanity :). Fast forward a few years and it no longer sounds like fringe science.
I have often pointed out in my reviews that while analog front ends are complex entities and habitats, with the advent of digital, especially streaming and CAS, the digital realm has become even more complex, where even the smallest anomalies can affect digital playback.
Along with the wave of data handling and processing, audiophile switches became a sort of compulsory product offering. The same goes for audiophile-grade linear power supplies that offer a subjective alternative to cheapo SMPS enclosed with the products, sometimes even with hefty price tags.
The Conclusion
The Ediscreation Firebird LPS and Silent OCXO Extreme Switch were conceived for ardent audio connoisseurs and aficionados. These devices are designed to create a sonic order and bring out the details that can remain hidden.
Edison Wong has approached this particular market segment with great dedication and hands-on experience. As you can see on Ediscreation’s website, the company has a variety of products in its portfolio, ranging from cables to power amplifiers.
Most generic routers and external power supplies for various audio components are mediocre at best. The enclosed universal switching power supplies are more often than not purchased in bulk by manufacturers. The same goes for off-the-shelf switches and their impact on audio streaming.
Many audiophiles and music lovers have already discovered that audio streaming depends on many factors, such as how fast and stable the internet is the question is what kind of “penalty” you get for less than optimal handling of everything involved with streaming, where, to complicate things even further, each streaming service handles encoding differently.
Both the Ediscreation Firebird LPS and the Silent OCXO Extreme are proof of concept
Tomi Engdahl says:
$500 Audiophile SATA SSD Cable With Superstar Crystals Listed
By Mark Tyson published about 3 hours ago
Claimed to be the “best computer Hi-Fi cable ever built.”
https://www.tomshardware.com/news/dollar500-audiophile-sata-ssd-cable-with-superstar-crystals-listed?utm_medium=social&utm_source=facebook.com&utm_content=tomsguide&utm_campaign=socialflow
A SATA 3.0 cable, the likes of which we have never seen before, has been listed on eBay for the princely sum of $500. For your money, the listing claims(opens in new tab) you will receive the “best computer Hi-Fi cable ever built.” We have seen some highly dubious audiophile / computer hardware crossovers previously, so thought this Audiophile Rocks Superstar Crystal Formula SATA SSD cable was worthy of a closer look.
Tomi Engdahl says:
If you like this audiophile Superstar Crystal Formula SATA SSD cable, you might also find the audiophile SSD, and the similarly targeted Ethernet Switch UEF appealing. Please be warned that all these products include a substantial audiophile tax.
Tomi Engdahl says:
AudioQuest fancy ethernet cable teardown pictures
https://www.stereonet.com/forums/topic/86462-audioquest-fancy-ethernet-cable-teardown-pictures/
Gallery: We tear apart a $340 audiophile Ethernet cable and look inside
We bought two, and since we can’t return them, we decided to destroy one and share!
https://arstechnica.com/gadgets/2015/07/gallery-we-tear-apart-a-340-audiophile-ethernet-cable-and-look-inside/
So, with an X-acto knife and spudgers and vice grips in hand, I separated the hefty expensive cable into its components, layer by layer. Potentially fantastical claims about audio clarity aside, the cable itself is of reasonably high quality, with braided and foil shielding around the entire cable coupled with foil shielding around the individual twisted pair bundles. The plugs are high-quality Telegärtner MFP8s, which cost about 9 EUR each depending on where you get them. There’s every indication that the cables conform to the listed Category 7 specifications and, if you were so inclined, you could almost certainly use them for 10-gigabit Ethernet over 100-meter runs and possibly even for short runs of 40-gigabit Ethernet (provided you can find the switching gear for 40GbE with 8P8C connectors). Of course, you can also use other shielded Cat7-equivalent Ethernet cables that cost one-tenth the Vodkas’ price for the same purpose, so the fact that they’re high quality cables doesn’t really justify the price.
When we finally stripped away everything and got down to the actual twisted pair of copper wires, we were pleased to see that they were indeed coated in silver, as the manufacturer’s page describes.
The terminations are rated to CAT 6a. Period.
I would be willing to bet that the NEXT (near end cross talk) will be WORSE than injection molded CAT6a cables due to the pair lengths I saw being out of geometry.
Tomi Engdahl says:
Ars prepares to put “audiophile” Ethernet cables to the test in Las Vegas
We’ve partnered with James Randi’s foundation to prove or debunk magic cables.
https://arstechnica.com/gadgets/2015/07/ars-prepares-to-put-audiophile-ethernet-cables-to-the-test-in-las-vegas/
Tomi Engdahl says:
http://www.image99.net/blog/files/category-diy-cables.html
Tomi Engdahl says:
I love how audiophiles can never quantify how a particular cable is better. It’s always vague nonsense like “sounds more natural” or whatever
Tomi Engdahl says:
he’s not entirely wrong but 2k is excessive. the biggest factor is cable length. because usb audio is isochronous, a longer cable can result in greater jitter and less signal stability. it’s why scientific measurement devices and usb packet analyzers come with short cables. source – design audio thingies for money.
Tomi Engdahl says:
He is entirely wrong. Jitter doesn’t change the signal. Either the signal gets there or it doesn’t. It isn’t impactful until you are doing high speed data transfers. Now if your audio thingies are not designed to handle the jitter that might occur over a bad cable, that’s another story.
Tomi Engdahl says:
it is impactful if you are trying to lower latency or use the full bandwidth of the device. dropouts can occur which is certainly going to affect the audio. isochronous endpoints have no error detection and since thats what audio devices are, you want a good (short) cable. thanks for your input.
dropouts, yes. Audio quality, no. As I said, either the signal gets there or it doesn’t
Jitter is a product of two digital devices passing audio to one another but not synched to the same clock source. “44.1k” for example is not actually identical between devices, which is why one needs a master clock source all other devices in the chain passing digital audio signals need to be slaved to.
Digital latency is defined by buffer sizes which are effectively the size of the buckets data is put into before it’s sent to the CPU to process. Smaller buffer sizes send the data more frequently because there’s less time between filling the bucket and sending it. This results in less latency, but taxes the CPU more which CAN lead to dropped 1’s and 0’s (audible pops and clicks).
USB cable quality is binary. As Ian said. The data goes or it doesn’t. It’s just sending 1’s and 0’s. CHECKSUMs and timestamping are used to assure data integrity, but those are functions of software not the cable.
Tomi Engdahl says:
As vacuum tubes use glass tube, hence their mechanical strength is less. Transistors are mechanically stronger than vacuum tubes. Vacuum tubes are comparatively less efficient. Transistors have very high efficiency compared to vacuum tubes.
https://www.tutorialspoint.com/difference-between-vacuum-tube-and-transistor
Tomi Engdahl says:
Twisted pair Ethernet is transformer isolated on both ends of cable. You can get ground loop issues only if you use shielded cable. If shielded cable has ground loop issues, then replacing the cable with unshielded will solve the Ethernet related ground loop problem immediately.
But that does not stop some companies trying to sell “magic boxes”
iFi Audio LAN iSilencer netværks adapter
https://www.av-connection.dk/?PGr=16341
Tomi Engdahl says:
The data carrying wires are transformer isolated in both unshielded and shielded cable. The cable shield is normally connected straight through from connector shield to another connector shield (which are typically connected to equipment metal case). Shielded Ethernet wiring can create ground loop issues (that generally don’t affect data but but can sometimes cause other issues).
Tomi Engdahl says:
LAN iSilencer Audiophile Ethernet Dongle Claims to ‘Quiet Your Network’
By Mark Tyson published 1 day ago
At ‘just’ $100, it seems like a bargain compared to some audiophile Ethernet cables we’ve seen in recent months.
https://www.tomshardware.com/news/ifi-audio-lan-isilencer-audiophile-ethernet-dongle
Tomi Engdahl says:
“iFi-Audio says that audio signals passing through won’t suffer due to the interference of ground potential differences and AC power transients”
https://www.tomshardware.com/news/ifi-audio-lan-isilencer-audiophile-ethernet-dongle
iFi-Audio shares some tech specs for the LAN iSilencer, which is listed for sale in Japan for the equivalent of $109 including consumption tax (10%). The device supports 1000BASE-T /100BASE-T /10BASE-T transmission speeds, has a shielded case, weighs 17 grams, and measures 73.5 x 19 x 16mm. In the examples showing this dongle plugged into various devices, it looks like it would be prone to accidental damage, as it’s a rigid device that sticks out nearly three inches from your Ethernet jack.
What we can say for sure, is that iFi-Audio’s claims seem to be significantly less wild than other recent audiophile computing gear we have reported upon. For example, the $1,250 AudioQuest Diamond RJ/E Ethernet cable
iFi-Audio’s claims seem to be significantly less wild than other recent audiophile computing gear we have reported upon. For example, the $1,250 AudioQuest Diamond RJ/E Ethernet cable mixed “Solid Perfect-Surface Silver Conductors” and carbon-loaded synthetics to shield and protect your wiring from RFI. Even more outlandish are the claims about the $500 SATA SSD cable with Superstar Crystals, which would apparently work its magic even if you didn’t plug it into an SSD. In recent months we have also reported on an audiophile SSD, and the similarly targeted Ethernet Switch UEF.
Priced at around $100 before tax in Japan, it seems to be a bargain if you believe the spiel. If you don’t, you can save your $100 and spend it on better quality headphones, speakers, or other audio gear.
Tomi Engdahl says:
isn’t the design of the (balanced) twisted pair network cabling, and the physical network interfaces on both sides of it, to remove the interference already?
. Yes from the beginning of the design of twisted pair Ethernet it was like this and still is.
Only application where there is a valid technical reason to add isolation transformer in the middle of cable are some medical networking applications, where the extra transformer provides safety isolation (guarantees higher isolation voltage than built in transformers around 1.5 kV)
Tomi Engdahl says:
If lan noise is a problem, why dont you use om3 Optical fiber?
Tomi Engdahl says:
https://www.dxengineering.com/parts/dxe-iso-plus-2
Tomi Engdahl says:
https://www.tomshardware.com/news/ifi-audio-lan-isilencer-audiophile-ethernet-dongle
Tomi Engdahl says:
It’s not easy to approach a high-end audio setup in the 21st century. Advancements in digital audio technology and extraordinary claims by snake oil salesmen have made high-end audio a complex field.
Tomi Engdahl says:
A network isolator is an electrically insulating separation device for Ethernet-based networks. It disconnects all the electrically conductive connections between the connected network peripherals and the connected device, and typically provides protection against DC and AC voltages.
Keep in mind that twisted pair Ethernet has already transformer isolation in the network interface and in the Ethernet switch.
those components are widely called “ethernet transformer” by component manufacturers and people who wrote Ethernet standards. They used to be 1:1.4 many years ago and nowadays typically 1:1
https://www.pulseelectronics.com/network-signal-transformers/ethernet-transformers/
Tomi Engdahl says:
Galvanic isolation has helped me in getting rid of noise problems in very many analog (and some digital) audio+video connections.
With Ethernet connections that use RJ45 connectors, there is a good galvanic isolation built into every standard compliant Ethernet port (both in device Ethernet port and in Ethernet switch). When you use unshielded Ethernet cables for interconnections, everything is isolated and no problems and no need for extra isolation in normal applications.
When shielded Ethernet cables are used for interconnection of devices with shielded RJ45 connectors in them, the cable shield interconnects the device typically grounded cases, leading to potential ground loop issues.
Tomi Engdahl says:
This LAN dongle claims to eliminate network audio interference
The jury is still out on whether it does anything
https://www.techspot.com/news/98010-lan-dongle-claims-eliminate-network-audio-interference.html
WTF?! Audiophiles go to great lengths to ensure they receive the best possible sound quality, but is it worth spending around $100 to remove audio interference in Ethernet connections? One company seems to think so, offering a dongle that claims to clean up streaming audio.
Audio equipment company ifi has released a small device that attaches to Ethernet ports to remove signal interference from sound streamed through the connection supposedly. The unique product, called the iSilencer, appears to target audiophiles who are already spending a lot of money to get pristine sound from internet sources like music subscription services.
https://ifi-audio.com/products/lan-isilencer/
Network LAN filter
The LAN iSilencer removes the electrical noise that distorts the audio signal from your network.
It reduces the noise floor and interference.
You can connect it:
To your router
To switches
Directly to your Ethernet wall socket
The LAN iSilencer has a zero-jitter memory buffer and galvanically-isolated inputs to eradicate digital distortion.
The LAN iSilencer boasts premium galvanic isolation technology, where electrical circuits are separated to eliminate stray currents. Audio signals can pass between galvanically isolated circuits to block differences in ground potential or currents induced by AC power.
Commenters on forums either scoff at the device or express mixed impressions of it. Some users appear to hear a slight difference in sound quality, while others claim a significant improvement.
Amazon reviews for another ifi device – a USB audio interference eliminator that launched in 2020 for $60 — carry a similar sentiment. It performs the same function as the Ethernet dongle after plugging into a computer’s USB ports, and reviews indicate a noticeable but perhaps not transformative change.
Tomi Engdahl says:
The jury is still out on whether it does anything useful. The LAN iSilencer boasts premium galvanic isolation technology, but looking at what is inside it I do not see any premium technology or claimed magical “zero-jitter memory buffer”
https://ifi-audio.com/products/lan-isilencer/
Tomi Engdahl says:
Coincidentally ‘performance exceeds cost by a good margin’ is my go-to pickup line.
Tomi Engdahl says:
Well if they look like snakes….so….snake oil?
Tomi Engdahl says:
“Our customers are happy in their self-delusions!”
Eye wateringly expensive Audiophile Bling, nothing more.
It really is more. It’s lying in the course of commerce, and that should be regulated.
Lying is not only legal in the US, it seems to be encouraged.
“Astronomical” and “jaw dropping” are not words that should ever be associated with describing the sound after changing a cable.
Maybe for the prices!
Tomi Engdahl says:
it is true that there is noise floor in all signal pathways. In Ethernet communications, if data signal level gets too close to the noise level, there will bit errors in data. If there are bit errors, the computer receiving the data will discard the complete data packet. With typical audio streaming using TCP/IP, there are two separate checksums (Ethernet CRC and IP ) that verify that packet was received correctly. If both checksums do not match packet is discarded. The streaming system typically sends the same packet again if it was not received correctly.
Tomi Engdahl says:
A network isolator is an electrically insulating separation device for Ethernet-based networks. It disconnects all the electrically conductive connections between the connected network peripherals and the connected device, and typically provides protection against DC and AC voltages.
Keep in mind that twisted pair Ethernet has already transformer isolation in the network interface and in the Ethernet switch.
Tomi Engdahl says:
I had a guy tells once he needed a cat 6A patch lead to his broadband router because the N rated wifi at 3m distance didn’t give true clear audio.
I said does the bit rate drop then? “No but it sounds different.” It’s literally not possible for buffered TCP audio to sound different over any media without but rate change.
Tomi Engdahl says:
Networking being marketed to audiophiles now?
Yeah, over a decade now when streaming and NAS became normal. Pretty interesting products. Or “interesting”, as sarcasm is easily lost in writing.
Tomi Engdahl says:
The pics that Tomi linked show that it’s pure snake oil.
Two TVS diodes, 12 ceramic capacitors, one Pulse Electronics Ethernet transformer. No “zero jitter memory buffers” of any kind. It’s totally not needed because every Ethernet device is already going to have a transformer like this onboard.
Tomi Engdahl says:
same type of signals isolation transformer module as found in pretty much every modern Ethernet port made to devices and switches. Does not look nothing special, made by company named PULSE, looks like from some of their commonly used series. Typical components around it. Can’t be sure if this device cuts shielded cable shield or does it go straight through from connector to another.
Tomi Engdahl says:
https://ifi-audio.com/products/lan-isilencer/
Tomi Engdahl says:
https://www.emosystems.com/network-isolators/faq/
1. What is a network isolator?
A network isolator is an electrically insulating separation device for Ethernet-based networks. It disconnects all the electrically conductive connections between the connected network peripherals and the connected device, and typically provides protection against DC and AC voltages of 4 kV and greater.
At the same time, it makes possible an almost loss-free transmission of high frequency AC signals with frequencies greater than 1 MHz, which are used in the Ethernet protocol for signal transmission.
EMOSAFE Network Isolators are suited to use in the medical field for the standard-compliant isolation of Ethernet-based signal interfaces (Signal Input/ Output Parts, or SIP/SOP) in accordance with the requirements of IEC 60601-1.
2. How can dangerous voltages be caused in the network?
For copper-based network cabling, it is possible that the signal lines or the cable shield unintentionally come into contact with other electrical parts through installation or design errors, aging or humidity. It is therefore assumed that the local supply voltage is present on the network cabling.
For this reason, network isolators are usually designed to be permanently exposed to a voltage of 250 V AC.
However, high voltage spikes and transients can appear on supply networks, and are caused, for example, by switching operations. Since these overvoltages can be much greater than the mains voltage, network isolators possess high voltage withstand strengths.
Furthermore, network isolators prevent the flowing of current between connected devices, which would otherwise result when electric potential differences exist between the connection points.
to the top
3. Where are network isolators typically used?
Network isolators are commonly used in copper-based network connections in private, public, and commercial sectors.
Typical applications include:
Medical electrical devices whose operation is permitted only when existing signal interfaces present a standards-compliant means of separation. Hence the frequently used term “medical network isolator”.
Sensitive measuring and monitoring devices in electrical testing laboratories, which are connected through an Ethernet network to a control center, and need to be
protected from transients and potential differences.
Computer systems which are electrically connected over long distances via
Ethernet cabling, where potential equalising currents need to be prevented.
Audio applications, where the transmission of low frequency alternating currents (AC hum) over the network connection is to be reduced.
Applications where valuable or especially vulnerable devices need to be protected from overvoltage and line noise from peripheral network equipment
4. When should network isolators not be used?
A network isolator is designed for data transmission over the frequency range from 0.3 MHz to 100 MHz.
Lower frequencies are strongly attenuated. For this reason it is not usually possible to transmit the signals of nurse call systems, telephone systems, or analog audio and video signals through a network isolator. Higher frequencies of up to 500 MHz can be transmitted through a network isolator; however the signal quality in this frequency range will, in many cases, not meet the application requirements.
The supply of electrical power to Power over Ethernet (PoE) devices is not possible via cable paths fitted with network isolators. See question 23.
to the top
5. I am a manufacturer of medical devices which have network interfaces.
Do I need network isolators?
If the housing of the medical device is located in the patient’s environment and the housing contains accessible, electrically conductive parts which are able to come into contact with the patient via the person performing the treatment, the manufacturer of the medical device must ensure that no discharge currents which exceed the IEC 60601-1 specifications for touch currents are able to flow through the patient’s body.
The source for such currents may be any electrically conductive connections from the device to its environment, which includes the network connection. (Please also see question 11).
As part of the prescribed type examination of the ME device, the network (signal lines and shield) interface are connected to a 250 V AC source. A measuring device with a test probe is then placed in contact with all accessible parts of the housing.
If, under normal conditions (NC), a current exceeding 100 µA is discharged through the test probe from any part of the device, it will not receive approval. Please note that approval may still be refused if the device passes the type examination, but the measures taken do not meet the requirements of IEC 60601-1. (Please also see question 25).
Network isolators create a galvanic separation point directly at the network interface, in accordance with the design requirements of IEC 60601-1, hence cutting off the current sources responsible for the touch currents.
The measures taken by the manufacturer must also fulfil single fault safety requirements;
i.e. they must also offer sufficient protection even when a second safety measure for patient protection which the same purpose fails, e.g. the protective earth connection.
This is tested separately. The touch currents under single fault conditions (SFC) may not exceed 500 µA.
It must also be taken into consideration that the interface and non-isolated parts of the connector may also be accessible. This problem can be solved with the use of suitable covers or by designing the layout such that a network isolator is located in the supply line outside the patient’s environment. (Please also see question 17).
Manufacturers may also obtain approval for the network interface of their medical device solely for connection with other medical devices, and identify the interface correspondingly.
In this case, the test described at the beginning is no longer necessary, and the network interface no longer needs to be galvanically separated.
However, this is only a viable solution when the interface is not intended for connecting to a clinic network, otherwise such a connection could be made by medical personnel within the scope of reasonably foreseeable misuse.
On the other hand, as a manufacturer, you do not necessarily need to install or provide a network isolator. It is sufficient when the accompanying documents prescribe their use in cases where your device is to be connected to a non-medical device or a clinic network.
By doing so, you transfer the responsibility for ensuring that the device has a safe electrical connection to the operator.
As long as your device is installed in an irremovable manner and possesses a permanently installed protective earth connection which satisfies the requirements laid out in IEC 60601-1, you do not require any network isolators.
to the top
6. I operate medical electrical devices.
Do I need network isolators?
Operators of medical electrical (ME) devices, according to the German Medical Device Directive (MPBetreibV), are responsible for ensuring that the standards for proper installation, operation and applications are adhered to.
13. The technical documentation for the EMOSAFE Network Isolators often
refers to “MOPP”.
What does this mean?
MOPP is an acronym of Means Of Patient Protection.
This term is defined in IEC 606011 as a “means of protection for reducing the risk due to electric shock to the patient”.
Here several protection systems usually work side by side or with one another, confronting various different potential hazards:
A high dielectric strength will protect the electrical equipment from frequent impulse voltages, which can be caused by switching operations.
A suitable insulation construction will reduce the probability of inadequate or failing insulation due to manufacturing defects or product aging.
Compliance with creepage and clearance distances will provide protection from electrical shock in humid or dusty environments.
Reducing leakage currents through protection devices to levels below the device-dependent maximum leakage currents prescribed in the standard will ensure that the patient is protected from permanently acting leakage currents.
Protective earth (PE) connections need to be in place to divert hazardous currents to earth before they reach the patient.
15. What does leakage current mean, and what are its permissible limits?
Galvanic isolation devices are never 100% perfect isolators.
This means that whenever a voltage is present on one side of the device, a very small amount of current will still flow (or “leak” through), and this is referred to as leakage current.
It is assumed that the largest voltage that a network isolator will be permanently exposed to is the rated root mean square (RMS) AC voltage of the supply network, with an extra 10% safety margin.
Because most network isolators are designed for a rated voltage of 250 V AC RMS, this results in a test voltage of 275 V AC RMS.
The permitted leakage current for the protected network interface is derived from the requirements of IEC 60601-1.
This allowable leakage current for a particular device is dependent upon several factors, and can only be determined by the manufacturer.
Only in exceptional cases are network interfaces within the medical field required to have leakage currents of 50 µA or less.
EMOSAFE Network Isolators typically have leakage currents significantly lower than this limit.
16. Must network isolators be subjected to regular safety checks?
Safety checks are not required for network isolators in medical applications, because network isolators in themselves provide no diagnostic or therapeutic purpose, and are therefore not classified as medical devices. However, together with connected medical devices, network isolators form medical electrical systems, which may to be subject to control.
Therefore, safety checks on network isolators within such systems are meaningful when the risk of a patient forming part of an electrical circuit with the Ethernet connection is to be reliably prevented.
Recommended safety checks:
visual inspection for external damage
inspection for the ingress of substances, especially liquids
function test
insulation test
availability and completeness of the accompanying documentation
17. How far away from the equipment to be protected can I install the
network isolator?
In principle, a network isolator should be located as close as possible to the equipment to be protected. For this reason, many ME device manufacturers construct galvanically isolated network interfaces directly into their equipment. However, there is a risk that the unprotected incoming cable enters the patient environment, and that the patient, the attending physician, or a member of staff can come into electrical contact upon cable connection or disconnection.
The various products from our EMOSAFE Network Isolator family allow this risk to be countered as appropriate.
The most common method is to
utilise an EMOSAFE Network Isolator housed in an external enclosure, for example the EN-30, which is then located outside of the patient environment.
This design offers the advantage that devices without network isolators can easily be retrofitted with the necessary protection.
EMOSAFE Network Isolators for device installation are intended for device manufacturers.
They have optional contact touch guards, and accidental removal protection.
21. How long are cable routes equipped with network isolators permitted to
be?
A widespread assumption is that network isolators reduce the maximum length of cabling.
In fact they do so, but in practice only very rarely.
Explanation: the maximum cable length for Ethernet connections is firstly limited by the signal propagation velocity required for a specific data transfer rate, secondly by the minimum allowable signal strength received at the far end of the cable, providing a
satisfactory signal-to-noise ratio.
A network isolator has no influence over the signal propagation velocity.
The insertion loss of EMOSAFE Network Isolators is approximately equivalent to that for
a cable of length between three and ten metres. This effect only begins to limit the maximum cable length when the insertion loss of the cable itself is borderline. In practice this occurs very rarely, as commercially available cable and connector components usually have much lower insertion losses than that required by the standards.
When in doubt, a cable certification device can help by providing a measurement of the insertion loss for the affected cable route. If the values are borderline, consider
alternatives to shortening the cable, such as using a better quality cable, or fitting better
quality terminal connections.
to the top
22. Is it possible to arrange multiple network isolators in a row?
In especially critical environments such as operating rooms, it may be appropriate to have both the ME device and the wall outlet fitted with separation devices, in order to reliably rule out any risk of unwanted voltages and currents reaching the patient.
In other cases, a duplication occurs coincidently when already-isolated ME devices are connected to isolated wall outlets. Such duplication is generally uncritical, because neither the signal strength nor the signal quality is significantly reduced by a network isolator.
However, for cable routes already operating on the edge of their specifications, or for devices with older network interfaces, this can lead to losses, which is why we strongly recommend testing such a situation before the patient is subjected to such an ME system.
23. I want to operate a Power over Ethernet (PoE) device on a cable route
equipped with a network isolator.
Is that possible?
A network isolator can be used in a PoE network without the network isolator being damaged or impaired in its effect.
However, because of its intrinsic electrical separation, PoE devices cannot be powered via a network isolator.
Should PoE devices nevertheless need to be operated, either the network isolator must be removed from the cable routing, or the device must be supplied with power by a separate power supply, being a medical power supply if necessary.
24. Network isolators interrupt the cable shield.
What effect does this have?
The total shielding or outer shield of network cables serves primarily to prevent the cable interacting electromagnetically with neighbouring cables and equipment.
Such interacting (or interference) affects the signal quality and the reliability of signal transmission. This can be especially important for installations where many cables are in a common cable duct or conduit.
The extra wire pair shielding supports the effect of the total shielding, and additionally it reduces the electromagnetic interactions between adjacent pairs inside a cable.
The interruption of the cable shield over a short distance is generally negligible, as the protective effect is maintained over the entire screened portion.
However, it can be entirely appropriate to continue cable shielding right up to the network isolator.
Twisted pair configurations in network cables provide effective protection against
electromagnetic interactions. In the United States of America, for example, it is rare to
find shielded Ethernet cables, as unshielded cables compete better in the cost-sensitive market.
Cable shields are almost indispensable, however, when used in 10-gigabit Ethernet applications (i.e. 10GBASE-T), due to the extremely high demands imposed on crosstalk performance.
Furthermore, a shield may be required when the network wiring is used for other forms of data transmission, for example analog audio and video signals.
The transmission of such signals is not supported by network isolators, which makes their shielding also irrelevant.
For some applications, electrically connected screening will be required to form a connection between the potential voltage differences of spatially distributed network components.
For this reason, some members of our family of network isolators provide the means of connecting shield connections to ground.
The EN-50 series offers plug connections for this purpose, allowing incoming cable shields to be electrically connected to the chassis ground of the devices.
In the context of EMC testing, it may be necessary to capacitively couple the shield connections across network isolators which are built into devices and equipment.
This is possible with capacitors which have especially high dielectric withstand strengths. Such a connection can assist in controlling the electro-magnetic emissions.
The EN-50 series of EMOSAFE Network Isolators offers such capacitively coupled connections.
Sometimes it is desired that a device, which is earthed via an Ethernet network cable, has a potential voltage matching that of its surroundings.
To ensure this, the incoming and outgoing shield connections of some of our network isolators are equipped with special high-resistance, high-voltage-rated resistors. These resistors allow equipotential bonding through the shield conductors, but do not permit significant current flow.
to the top
25. In my Ethernet network, I only have unshielded cables.
Do I still need a network isolator?
In the transmission and reception units of Ethernet network components, the signal conductors are electrically isolated with the help of transformers.
Therefore, it is often assumed that an additional separation of the cable screen (e.g. the use of unshielded cables) would be sufficient to comply with the isolation requirements specified by the standards.
However, the transformers commonly used in Ethernet network components only provide a functional isolation. The IEC 60601-1 requirements on the technical implementation of electrical isolation devices cannot be met by these transformers.
Specifically, these requirements are those of dielectric withstand strength, creepage and clearance distances, and the insulation construction.
Enamelled wire, for example, which is commonly used in the construction of transformers, is explicitly stated in the standards as inadequate – even if the design attains the required dielectric withstand rating.
26. How does EMO Systems guarantee the quality of their EMOSAFE
Network Isolators?
EMO Systems is certified to DIN EN ISO 9001 and DIN EN ISO 13485.
Tomi Engdahl says:
https://www.meilhaus.de/en/med-mi-1005.htm
Network Isolator MED-MI-1005
The MED-MI-1005 network isolator offers a flexible, easy option for immediately isolating industrial or medical equipment. In less than 5 seconds, equipment, users and patients are protected from dangerous leakage currents flowing via the data network. The MED-MI-1500 doesn’t require a separate power supply and behaves completely transparently on the data network. Therefore, no installation of drivers or additional software is required. The MED-MI-1005 was tested under stress conditions for up to 10 kVDC (10 s) isolation and 5 kVAC (60 s).
Easy to handle, no software installation required.
Compact, sleek and universal range of applications.
Protects equipment, users and patients in less than 5 seconds from dangerous leakage currents.
No fixed cables. No need to replace the entire isolator in the event of a faulty network cable.
Lightweight, durable housing.
Can be connected to any system with a network connection
Inexpensive and efficient operation.
No dedicated power supply required.
Tomi Engdahl says:
https://www.audiosciencereview.com/forum/index.php?threads/reducing-noise-in-ethernet-cable.2371/
If I am streaming music from my server over an Ethernet cable to my SoTM SMS Ultra do I have to worry about noise? If so, how do I reduce it?
The only thing I have seen is that if they using a switching wal-wart, replacing that with a linear supply reduces mains noise. But that noise is not audible.
No. Ethernet is not used as a clock source (“it runs asynchronously”). So it can’t cause jitter by itself. And it is transformer coupled so very good isolation.
If you had extreme AC mains coupling on it, you could use one of these isolators:
I am a bit unclear on your target usage scenario. I quickly looked up your network player and it uses UPnP so its likely its over TCP/IP and reclocks from the buffer, so whatever this SoTM does is largely irrelevant as long as you can stream to the network player. You can use a wireless dongle to talk to a server in another room.
I used to own a streamer (a Linn DS which I used for a few months and sold for a nice profit. weird but true) and it made sense to not have an intermediate PC plugged in, but then it stopped making sense once I realized UPnP is not a secure protocol and that most servers have limited functionality. A fanless PC + USB DAC made more sense in the end, especially because this was also cheaper overall. I have shared network drives mapped on the fanless PC and the music is stored on an external HDD attached to my main PC upstairs. Works well, except that Windows does not automatically reconnect the mapped drives. I’ve too lazy to write a startup script that will connect them automatically.
Great post! So much BS in the audio world it is incredible. In order not to fall prey to these con men I would have to have an electrical engineering degree or be a member of Audio Science!
I am in the medical field and I do have something very similar to what you show in your post. I do not have 12 inches between my ethernet cables and my AC mains. It is looks like a pasta salad behind my racks and TV stand.
Review and Measurements of EMO EN-60KDS Ethernet Isolator
https://www.audiosciencereview.com/forum/index.php?threads/review-and-measurements-of-emo-en-60kds-ethernet-isolator.1869/
This is a review measurement of an Ethernet isolator cable (passive) which is sold to medical industry. Ethernet is normally an isolated interface with transformers at both end. Addition of this is likely for high voltage surges and such.
I was loaned one of these to test. As you see, it is a simple, in-line cable: http://www.emosystems.com/product/en-60kds/
The unit retails for $204 in US and about 153 euros.
Searching on CA forum, seems like folks advocate its use to get rid of leakage noise. So I used the Sonore microRendu networked audio adapter which seems to be pretty sensitive to AC mains leakage. To make sure it didn’t contribute any, I used my lap supply to power the microRendu.
The Ethernet connection is from my local lab switch which in turn connects to another larger switch in basement equipment closet. The cable from switch is about 6 feet. The isolator was placed at the end of the cable and terminating into microRendu.
Here are the measurement results:
As we see, there is no change other than run to run variations.
Conclusion:
The use of Ethernet isolators is for life safety and belts and suspenders at that. There is no reason to use them with computer audio or data for that matter.
Use that money to buy pizza and ice cream and you will be happier for it.
Edit: further testing with ethernet cable wrapped around mains cable to force induction of mains, some difference was found. See: http://www.audiosciencereview.com/f…o-en-60kds-ethernet-isolator.1869/#post-48753
ust a few words about functionality of this device.
1. As noted, Ethernet already has isolation transformers. The reason it can still leak is because the transformer is not ideal. It has some capacitance and that allows forward conduction of mains frequencies. The EMO has a higher quality transformer in it with less capacitance. Hence the reason it doesn’t leak as much. A good transformer in the device would have done the same thing.
2. Secondary, it has clamping diodes to clip surges. There is very little space in there so the energy it can shunt will be quite small. But it is there as a secondary measure of sorts.
My testing shows that #1 is true at least in the case of Sonore microRendu (so much for these devices being higher quality than computers in this regard). Its Ethernet input transformer did let some 60 Hz mains harmonics through which the addition of EMO eliminated. But maybe it is the combination of both that allowed that. Hard to say.
Still a good job on debunking some of the claims of high end ethernet cables, audiophile switches etc.
Literally all anybody needs to do to debunk these types of clamis is to take a few minutes to actually read the wikipedia page on TCP. That’s it! It debunks all of it simply by the nature of how TCP works (and you can apply what you’ve learned to more than stupid audiophile myths). I mean, this paragraph alone drives a nail right thru it (my bolding for emphasis):
At the lower levels of the protocol stack, due to network congestion, traffic load balancing, or unpredictable network behaviour, IP packets may be lost, duplicated, or delivered out of order. TCP detects these problems, requests re-transmission of lost data, rearranges out-of-order data and even helps minimize network congestion to reduce the occurrence of the other problems. If the data still remains undelivered, the source is notified of this failure. Once the TCP receiver has reassembled the sequence of octets originally transmitted, it passes them to the receiving application. Thus, TCP abstracts the application’s communication from the underlying networking details.
In other words, TCP fully reassembles all the packets BIT PERFECT (to use an audiophile term) to what came over the “wire” before any audio application ever sees it. If you are using TCP, the layers below it do not matter at all as long as the packets are coming in fast enough.
Oct 19, 2022
#15
anotherhobby said:
Literally all anybody needs to do to debunk these types of clamis is to take a few minutes to actually read the wikipedia page on TCP. That’s it! It debunks all of it simply by the nature of how TCP works (and you can apply what you’ve learned to more than stupid audiophile myths). I mean, this paragraph alone drives a nail right thru it (my bolding for emphasis):
In other words, TCP fully reassembles all the packets BIT PERFECT (to use an audiophile term) to what came over the “wire” before any audio application ever sees it. If you are using TCP, the layers below it do not matter at all as long as the packets are coming in fast enough.
TCP works, but audiophile lore has also evolved and now includes various explanations for why even a bit-perfect playback may not sound good. With no real evidence, measurements, or controlled listening tests, various audiophile manufacturers started producing “cures” for these “issues”. Cures that include things like ethernet clock regeneration for jitter reduction, super-capacitor power supplies for the ethernet switches to reduce PS noise and ground leakage currents, and heroic measures to add galvanic isolation into network switches and network cards. All, supposedly, to reduce the electrical noise that can travel from the source device all the way to the output of a DAC (or to the clock power supply causing additional jitter). This also resulted in expensive audiophile ethernet cables, overpriced outboard clocks, and monstrous ethernet switches that are overbuilt and overpriced for what they do, all based on faith that somehow the expected reduction in network noise will result in better sound quality.
This is the kind of stuff Eric tried to test for and demonstrated that the cure can be much worse than the disease, in some cases increasing noise and jitter by a large margin — just the opposite of what these devices supposedly claim to achieve.
Some streaming protocols run over UDP, rather than TCP, where the responsibility for end-to-end data error correction and ordering is the responsibility of the application. Meanwhile there’s a strong move to QUIC and HTTP3.
I’ve done troubleshooting on large Ethernet/IP networks. It’s a mistake to assume end-to-end errors don’t exist. A router, for example, rewrites Ethernet headers and recalculates the checksum. I’ve see routers with slight memory errors corrupt the data field, then create a correct checksum for the packet.
However, everything I’ve seen so far in audiophile land to “fix” Ethernet is daft.
Some equipment may not have good galvanic isolation, or may be sensitive to noise, but those are flaws which should be fixed, not mitigated.
Yeah, well it’s easy to see where he went wrong.
‘The causal connection “Better Ethernet jitter value” = “Better sound” cannot be proven here.’
The equation he needs to assess is instead “More money spent + more fancy boxes littering the living room” = “Lower feeling of inadequacy when browsing audiophool forums”.
And it all gets buffered into RAM and reclocked out anyway. Even for USB and S/PDIF these days. Bits are indeed bits.
If someone in a hifi shop tells you one device has a better sounding digital output than another, just hold on to your wallet and walk away.
Some streaming protocols run over UDP, rather than TCP, where the responsibility for end-to-end data error correction and ordering is the responsibility of the application.
UDP is typically used only in applications where low latency is more important than error-free transfer, such as telephony. Home music streaming invariably uses TCP.
I’m staggered that people are cheating audiophiles. When did this start happening?
Around the same time audiophiles were created, I think.
Some streaming protocols run over UDP, rather than TCP, where the responsibility for end-to-end data error correction and ordering is the responsibility of the application. Meanwhile there’s a strong move to QUIC and HTTP3.
Can you name a single audio streaming service that applies to this discussion that uses UDP that’s not somebody’s open source project on github? I’m going to guess there is a zero percent chance of that. UDP is generally used for very low latency where dropped packets are okay (i.e. gaming). UDP is not a good tech for audio streaming, so I’m pretty sure it’s not a thing. The discussion of UDP in this case is irrelevant.
Around the same time audiophiles were created, I think.
It’s a self-reinforcing loop: audiophiles “hear” a difference, speculate as to what causes it, then some go into business to manufacture devices believing that they discovered something new. The audiophile press starts reviewing them, audiophiles read about them, buy the products and post exuberant reviews.
New audiophile businesses spring up, encouraged by the press and the growing audiophile interest, along with overly positive forum and web “reviewers”. Old, established business see a high profit market, and join the fray. They experiment with tweaks without understanding of physical or psychological principles behind what they (or their customers) hear, and the loop closes with more and more expensive products being created based on nothing but speculation and uncontrolled listening tests.
I’m sure that a few audiophile tweak manufacturers are lying and know better. I bet most are just as deluded as the rest of audiophile community and are simply not interested in learning the truth.
Anyone honestly believing that certain Ethernet cables, audiophile grade switches, Ethernet reclockers, and so on, can make an improvement either doesn’t have a clue how TCP/IP and Ethernet work or started out with faulty kit.
While I have to agree, in the context of audiophile tweaks, it’s really no worse than cable risers, the Shakti Hallograph and numerous other magical implements. They are all a way of improving PRaT and separating a prat from his money.
ome equipment may not have good galvanic isolation, or may be sensitive to noise, but those are flaws which should be fixed, not mitigated.
When we consider that a typical wired Ethernet network in a large home consists of an Ethernet line going from the cable modem to one to three second-level Ethernet switches – usually 1GbE – and the cascade depth of the network is two, and these switches are typically (almost always) using old-fashioned IEEE 802.1 bridging (i.e. Spanning Tree Protocol), this is not high-tech stuff in the modern sense. And even 1GbE is serious overkill for audio (uncompressed 24/192K digital audio requires – what? – about 10Mb/sec per stream).
I was recently having a casual conversation with a custom HT installer, and he told me that most of the new homes he’s outfitting are not being wired for Ethernet, but for WiFi 6 meshes, even homes with a 4K TV in several rooms. He did do one home recently wired for 10GbE, but that was an outlier. Wired Ethernet is apparently so yesterday.
anotherhobby said:
Can you name a single audio streaming service that applies to this discussion that uses UDP that’s not somebody’s open source project on github? I’m going to guess there is a zero percent chance of that. UDP is generally used for very low latency where dropped packets are okay (i.e. gaming). UDP is not a good tech for audio streaming, so I’m pretty sure it’s not a thing. The discussion of UDP in this case is irrelevant.
Some professional studio video and audio protocols are based on UDP. Domestic audio streaming protocols are TCP, but QUIC may change that:- there are reductions in server and multiplexing overhead which the stream providers may benefit from. Also reducing the delay to your home (e.g. for Radio and TV streams) is a goal of providers, so that you don’t see the World Cup goal 20 seconds after your neighbours have been cheering when watching via DTT. https://fiestajetsam.github.io/draf…rements/draft-gruessing-moq-requirements.html
anotherhobby said:
So have I. I’ve been in this industry for over 30 years. That’s a bad/broken router. That’s a very specific problem in a very specific circumstance, and your stream would stop working or rebuffer if that happened, but you would not lose “audio quality.” If TCP did not work the way it did, the Internet literally would not work. TCP guarantees you either get all the stream’s contents in the correct order, or not at all (an error). A broken router is irrelevant in the context of this discussion.
You are right it was a broken router. But the error occurred once every million frames or so! The only way I could find what was going on was by correlating card data and netstat over a bunch of devices, some routed through it and some not – the Ethernet errors were lower than they should have been relative to IP/UDP and IP/TCP. TCP/IP-based systems silently recovered from the junk they were given, but some UDP-based systems were confused. Once I caught a corrupted packet it was clearly rubbish, but the Ethernet checksum was correct!
anotherhobby said:
Agree since there is literally nothing at all wrong and there is nothing to fix.
Agree, and again not relevant to the actual claim, as there is zero validity in the base claim.
I don’t mean to be argumentative, I just don’t think this claim deserves any help whatsoever.
I’m not in any way supporting the claims by these vendors.
@all,
sure, bits are bits, data integrity is not affected as it is guaranteed by design of the protocols used. That is not the point here.
The point here is that at the last stage of the digital stream, at the input of the DAC chip proper, noise and timing issues might affect analog output (which is certainly true for bad DAC designs), also the noise can sneak directly into the analog circuitry. And that’s why less noise and better timing right at the DAC input can make a difference if better cables or whatever do actually achieve that… as I said before, this is not so easy to examine (measure) but it is not impossible and thus cannot be dismissed by philosophical arguments (like “bits are bits”). But, alas, this is not what’s been tested by Eric, rather he tested for potential disturbing mechanisms which may or may not affect the analog output.
You’re forgetting about end station buffering. Every audio streaming app I’ve tried (several, including Apple, Amazon, Spotify, etc.) buffer at least two seconds of audio. All of the effects you’re discussing will be negated by buffering.
ahofer said:
Interesting to me that, once again, adding low level noise seems to be correlated with finding the sound “more spacious”. In this case, unlike vinyl/tape/tubes, it may not be audible.
That statement caught my eye. We as humans have a sense of what sound ‘sounds like’ in a spacious room so by modifying the signal (introducing some type of distortion) to have that same effect, it would be interpreted by the brain as making the sound seem spacious, although likely less accurate. People may prefer this, even calling it an improvement, even though it may be some type of distortion or noise. I recall an older AV receiver I had that had different room effects to make it sound like different rooms: outside, jazz club, church, stadium – I think they were using digital filters to adjust (distort) the sound to give a similar effect.
It is trivially true that even a buffered digital stream will exhibit timing artifacts on output that depend on the stream contents, RFI, phases of the moon, whatever. There is a whole class of cryptographic attacks based on detailed timing artifacts, and those folks at the NSA and GCHQ have put a lot more work into the field than any audio company.
The web browser you are reading this on contains code to deliberately blur the system clock to any code running inside the browser, precisely because timing attacks are so easy that allowing access to a precise clock is a security risk.
Heck, it’s probably possible to look at the jitter spectrum of a signal going into a DAC and do better than chance at predicting if the music is Kenny G or Metallica.
But it just doesn’t matter. The information content of the timing variations is so far below audibility it‘s like listening for a baseball hit in New York from London. What was the residual jitter on the SMSL500, -120dB?
KSTR said:
You still don’t understand and still think in the digital domain only, it seems. Buffering does nothing against common-mode RF noise entering the system, this happens right at the RJ45 input jack and it is an analog phenomenon.
But PHYs have been magnetically coupled and have had 49.9 ohm load resistors on each pin for over a decade. If you have enough noise coming through to impact the end device ground, you probably don’t have a working connection.
I work in IP telephony where mixed TCP and UDP is my day to day. It does not take much to upset TCP. Ethernet will fault (and probably the PHY/controller) before ANYTHING even comes close to impacting the endpoint ground via some sort of potential difference which might cause analog “sneaking in.” This is doubly true for PoE devices. They have 0 tolerance for any funny business. Triply true for a 10g link. Want 802.3bt on 10G-baseT? You will know at roughly the 500ms mark if your cable supplier has cut any corners.
But if you are really worried about PHY transmission, there’s nothing stopping you from replacing your copper runs with OM1-OM4. It would still be multitudes better value than any audiophile funny business because OM1 is still rated for 10g up to 33m. And before anybody jumps on the latency bandwagon, I get motion sickness very easily, but manage to use fiber-based DP and USB for 8k VR between rooms.
You can even run fiber inside HVAC plenums without any special rating. It saves a lot of heartache on drywall and paint. (That’s a return plenum, hence why I need to clean beneath the grille and make a custom unit.)
Isochronous transmission is never used in consumer audio devices. Buffering a few seconds of music is always the cheapest solution, even in the cheapest mobile devices.
IMO, Ethernet really does suck at the link level for high performance data center use, which is why most other interconnects like PCIe, CXL, FC, IB, and HPC interconnects work differently, but digital audio is just not a challenging application for even a 1GbE network. TCP is hardly a state of the art protocol either, but this notion that digital audio streaming is somehow a demanding or challenging application for modern Ethernet/IP networks is just nonsense.
Hi all,
I just saw this thread about my work!
Will join the discussion later in more depth as I am currently time limited.
I’ve been a bits are bits guy for a long time, and had granted myself a very long time against the claim that Ethernet can affect sound – until I heard it for myself and then started investigating this.
I am currently investigating in sound changes by LAN cables and here too you can see quite a few differences in how the cables produce or pass on interference, which must be the reason for the sound changes.
Listening tests are made after my measurements by independent people who do not know the measurement results. Initial results confirm the thesis that noise is responsible for the sound changes.
I will get back to you later.
Best regards,
Eric
Some more words for those for whom the subject is new:
The sound difference is not caused by influencing the data signal. This is quite bulletproof over Ethernet as it is a differential signal.
But with the differential signal also common mode noise is transmitted, which gets into the ground of the endpoint (for example Stremer) and is then given further over the digital chain (e.g. USB). With the streamer I use for measuring, all Ethernet noise is transported directly via the USB ground to the DAC. The data are of course bitperfect – there is no doubt about that, so speculations regarding the manipulation of the data is impossible.
The common mode noise are usually quite indifferent to Ethernet, so here with normal “IT” switches only in the context of EMI compatibility is looked.
For more information you can also check this thread:
https://www.open-end-music.com/forum/privatforen/thomas-michael-rudolph-tmr/651298-diskussion-ueber-messungen-an-ethernet-infrastruktur#post688782
https://www.open-end-music.com/forum/privatforen/thomas-michael-rudolph-tmr/651284-messungen-von-ethernet-infrastruktur-switches-nur-lesen/page3#post695734
Paul K bumping this thread had me re-reading this and thinking after seeing your comment again in the present, it has me thinking.
I recall Amir saying that EMI/RFI into unshielded audiophile interconnects can something create a false sense of detail. This provides the scientific basis for how some might hear cable differences.
Sony’s DSEE HX is all about adding extra noise, albeit they have stated but never published the science that the DSEE AI is based upon master recordings from DSD/DXD in the Sony Music catalog, converting them to 44.1 and then feeding it into a machine learning algorithm to try to identify patterns between the 24/44.1 and 24/48 content vs the actual high res recording. But what you are really capturing then is the pattern of ultrasonic noise of the ADCs used in that training set on top of any actual high resolution musical info.
I recently looked at the Topping D90 v. Marantz SA-10 and came to the realization that the increase in ultrasonics in the SA-10 is more truthful to the DSD file. And if there is such a thing as a DSD sound, it’s ultrasonic noise.
Amplifiers like the L-08M (DC to 600 kHz; no worries about cable choice), Spectral (requiring MIT cable to avoid oscillation), and Halcro (250 kHZ bandwidth) and never famed for their warmth. In contrast, Marantz’s super slow roll off filter adds a ton of artifacts in the ultrasonic range in their AVRs and the SA-10 chooses to run well past the 50 kHz of scarlet book.
I tried looking up the SACD standards and found
ANNEX D.3 (Maximum peak signal : MaxPeak/MP)
Peak signal levels above +3.10 dB SA-CD in quasi DC-50kHz bandwidth determined in ANNEX D.3 (MaxPeak/MP) are not allowed.
ANNEX D.4 (Maximum high frequency noise power : HF)
The accumulated RMS signal + noise level of the DSD signal in 40kHz- 100kHz bandwidth determined in ANNEX D.4 is maximally equal to the RMS level of an input sinewave with a peak amplitude of -20 dB SA-CD.
That’s a really high ultrasonic allowance of noise!
Doing DeltaWave comparisons between @pkanes algorithm for DSD decoding and the ultrasonic roll off filter and TASCAM’s (presumably the same as Esoteric’s algortihm) shows that decoding DSD 2.8 generates anywhere from -120 dB to -140 dB PK Metric depending if you are converting to 88, 176, 352, 704 kHz.
Maybe those who think they hear the EtherRegen doing something to the sound are hearing the effects of added ultrasonic noise which then pollutes the audible range like seasoning in a way that is favorable.
If you look at Marantz’s various noise shaping algorithms, they picked the one that allows the most ultrasonic noise as the default for the SA-10 and for the AV10, they also offer the correct filter for measurements but the default slow roll off for the default setting since it sounds better to their team. I have to imaging that a company like Marantz has detailed analytics on what sells compared to the Denon equivalent and their decision away from transparency is intentional and founded upon marketing/sales data somehow.
Noise can certainly make a difference. It can become audible directly by leaking into the analog output stage, or it can create distortions by modulating the D2A clock, creating jitter/phase noise. All of these effects should be measurable at the output stage, and it should be easy enough to test with EtherREGEN+REF10 and without (since these are not critical for operation and can be removed completely). I’ve yet to see a measurement where this improvement is demonstrated including from the manufacturer of EtherREGEN.
For me, any addition of extra noise and distortion to the output of a DAC is an undesirable result of adding nearly $6000 worth of equipment, but I guess others may still want to pay extra for a less transparent, more veiled playback
At least one such test demonstrated that EtherREGEN had no effect at the output of a DAC. At all:
https://www.audiosciencereview.com/forum/index.php?threads/uptone-audio-etherregen-switch-review.10232/
https://www.audiosciencereview.com/forum/index.php?threads/surprise-findings-on-ethernet-cables-cleaners-reclockers.38371/#post-1351513
To be clear, I think audiophiles should all prefer WiFi.
1) No ground loops
2) No worries about cable topology
3) No need for cable lifters
4) No worry about dielectric, skin effect, cable materials, connector corrosion
5) Has it’s own version of Ken Ishiwata, Nelson Pass, Franco Serblin in Hedy Lamarr
I think the whole process of audiophile routers, wired or wireless, is a sad waste of effort.
(Edit: And this is coming from someone who *believes* in sound signature…)
https://www.audiosciencereview.com/forum/index.php?threads/surprise-findings-on-ethernet-cables-cleaners-reclockers.38371/
It’s hardly surprising that in Eric’s tests these so called ethernet “cleaners” actually caused more noise and increased jitter in the digital transmission. The need for an expensive outboard “low phase noise” clock to plug into an expensive Ethernet regenerator/reclocker to improve audio quality is just one idea that especially boggles the mind…
Apparently these guys have never heard of CRC32 and receive-side endpoint buffering. (I did a site-wide search for “CRC”, and nothing came up.) Of course, then there’s IP and TCP checksums, and I’ve never seen a music app that is buffering less than two seconds of digital data. And it is digital data. IMO, some peoples’ thought processes would benefit from an enema.
Tomi Engdahl says:
https://community.roonlabs.com/t/noise-over-ethernet-from-switches-routers-computers-etc/61551/4