@sns and others, like I said previously, John and UpTone are not wrong when they say there can be leaks via Ethernet that would interfere with internals of a streamer or other component. This used to a huge issue in the 90s for computers, and it has continued with cheap components that have found themselves into non computer related devices like audio equipment. But most networking cards today don’t have these issues, even the really cheap ones. And I am hoping that streamers and other components that costs thousands of dollars like my NAD M50.2 don’t use really crappy components, but it wouldn’t surprise me if they did, after all, they are audio engineers, not networking experts.
But, and more troubling for John and UpTone, you can measure this very easily, and UpTone Etherregen does not eliminate any of this at all.
|
So, here we go again, another thread about bunch of people applying pseudoscience audiophile terms to Ethernet. Nope. Ethernet does not work that way. Ethernet, along with the TCP/IP stack are several layers of error detection and in some layers error correction. If an error is detected the frame, or packet, or datagram is discarded, or in the case of error correction, discarded and the retransmit requested.
So why fiber at all? Copper Ethernet cabling has limited range, 100ft is a safe bet, but depends on speed and cable standard.
but fiber for super short runs like 5ft is crazy waste of money.
|
And, final point, Ethernet is a layer 2 protocol. The type of cable does not change this, as the cable is layer 1, regardless if it is fiber or copper, or barbwire (yes, you can run Ethernet over barbwire).
|
@yyzsantabarbara yeah, no. Read up on what you are typing before you type is my suggestion. If you program this “stuff”, you really should know more than what your post implies, which is 0.
look at the cable itself.
@sns you have never heard of it because it is not your area of expertise, or even foundational knowledge… largest manufacturer of networking equipment in the world and its landing page for routers :
|
@yyzsantabarbara I certainly did not, here was my answer:
"@yyzsantabarbara yeah, no. Read up on what you are typing before you type is my suggestion. If you program this “stuff”, you really should know more than what your post implies, which is 0."
You don't know what you are talking about. There is no broadcast protocols in streaming.
|
@jumia just because you have 2 ethernet ports, doesn't change the protocols involved. It's still bidirectional, and it's still error detection and correction involved.
Go with what is proven, and not what someone else is trying to tell you will work better.
|
@yyzsantabarbara lol, can’t you humor us and explain how a compressed, and in most cases digitally signed or encrypted signal is broadcasted and lost packets are not retransmitted. Can’t wait for the laughs!
@sns and you, you think that your optical cable and FMC somehow can modify the payload, which is where the music resides during transmission. Again, this fmc and optical cable somehow figured out how to break encryption, flac decoding and pcm in real-time and then modifies it to sound better. I have a few bridges for you, just send me all the money you and your children will ever make up front.
I listen plenty, don’t worry about me.
|
|
@jumia the summary of this thread is, either you believe the impossible, or you teach yourself about the realities of streaming and networking, and spend your money elsewhere.
|
@sns absolutely. What it would take to actually make a difference is that you modify the payload. I put every thing you said against the trillion of bank transactions every minute happening globally. Because if FMC actually had a impact and could change the payload, people would be stealing money with the technology.
you don’t understand the technology, that is very obvious, and that is why you can fool yourself in believing it. It is not the same as a speaker cable, interconnect, or even a power cable. It is infinite more complex system with built in safe guards against what you are trying to achieve. The entire protocol set is built to withstand your efforts, and you think you can do anything to impact it. It is laughable.
Like I said, if your premise is correct, trillions of bank transactions are at risk. You really think so?
or do you think a few hundred, maybe even thousands, audiophiles have confirmation bias?
|
@yyzsantabarbara man up now. You don’t know what you are talking about, it is as simple as that. But don’t take my word for it, ROON says so. At least admit it.
|
|
@yyzsantabarbara oh wow. You are special! Can’t even admit you have no idea what your are talking about after your vendor of choice throws it in your face.
UDP is not broadcast. It is a stateless L3 protocol. And no, it is not just Roon. You still don’t get it, if you lose a packer when you send a compressed stream, you can’t uncompressed that part of the stream at all. In addition, DRM requires at least digitally signed music, so, the entire song must be received in its entirety for it to work for downloads, and larger chucks of the stream that is digitally signed. The only thing you have proven is that your really have no clue, which I don’t blame you for, it is infinite more complex than a speaker cable, except that you keep saying you know something about it, which you absolutely do not.
from Wikipedia “ The Internet Protocol Version 4 (IPv4), which is the primary networking protocol in use today on the Internet and all networks connected to it, supports broadcast, but the broadcast domain is the broadcasting host's subnet, which is typically small; there is no way to do an Internet-wide broadcast. ”
but hey, you “program this stuff”.
|
@jumia No clocking benefits, if noise is entering your streamer/DAC through Ethernet, you have some serious other issues.
so the take away is that you should not waste your money. Like I have pointed out and proven, the likes @sns and @yyzsantabarbara have no clue about how networking works.
|
@jumia so, I would skip the “audiophile” switches, and go with a enterprise switch. You can find cheap Cisco, Juniper, or other enterprise grade switches on eBay.
“Audiophile” switches that I have seen actually have no real networking specs, just terms that doesn’t apply to networking at all.
|
@sns here is the thing, I am guaranteed the foremost expert on networking on this site. I don’t know a lot about everything, but networking is the one thing I do know, love and have made my life and career off.
Therefore, like I said, confirmation bias is a real thing, and what you are proposing, where you could modify, in your favor, the payload, in real time of encrypted and digitally signed protocols, is a joke. Trillions of bank transactions every single day happen securely just because you cannot do what you think you hear. Money talks, and billions, if not hundreds of billions of dollars move every day on the basis that what you say you hear, cannot under any circumstance happen, ever.
those are the facts.
|
@yyzsantabarbara here you go again… you just don’t know what you are talking about. Broadcast is not at a higher level than TCP or UDP (L4), broadcasts happen at L3, a lower level, and is used among other things for ARP.
and for sure TCP can and will reduce throughput, but not by dropping packets like you say, rather it reduces what is called the window size. The window size is basically how much data can be transfer before the ACK is sent. TCP never intentionally drops packets, and what you are quoting doesn’t even say that.
what I have been saying, and continue to say, is that your hearing is up against trillions of secure and undisrupted banking transactions daily. The entire system is built to prevent what you say you can hear. You may think that is mocking, but I find it insanely arrogant to not at all understand networking, like yourself, and say you can hear a difference, which would invalidate everything the Internet able to provide today.
|
@ghasley it is very possible for noise interfere with Ethernet, for sure. The point is, Ethernet is built to withstand it, and that’s why I said a safe bet for a Ethernet run without special conduits, etc. is 100 ft for residential applications. See below the max distance which in some cases is dependent on the speed, however, as a general statement, no benefit to deliberately slow down.
going back interference, the protocol Ethernet have error checking, each frame transmitted includes a checksum to make sure nothing corrupted (changed) during transmission, see below. And this is just one layer. Every layer have these checksums, and the actual application layers deal with compression and digitally signed content (DRM). So if you in anyway, modify the payload, all of these layers will fail all their checks, and it will be discarded and retransmitted. In reality, as soon as one layer detects an issue it is discarded and never goes further.
now, can noise travel via the Ethernet cable into the device itself and cause issues, sure, but that would mean a very poorly network card implementation in the device.
https://www.tripplite.com/products/ethernet-cable-types?utm_term=&utm_campaign=a.Cables+-+Panels,+Jacks+%26+Hardware&utm_source=adwords&utm_medium=ppc&hsa_src=g&hsa_ad=522769106538&hsa_tgt=dsa-1274206709011&hsa_mt=&hsa_ver=3&hsa_acc=4932208510&hsa_kw=&hsa_grp=121359314526&hsa_cam=696012424&hsa_net=adwords&gclid=Cj0KCQjwxveXBhDDARIsAI0Q0x1eE6NPdYOyoq_munQUKOqbWS7kYPopMyKZdsy4k7xHq-qWBnGZPWEaAvijEALw_wcB
https://en.wikipedia.org/wiki/Frame_check_sequence
|
@ddafoe i did. Show me another person here that knows more and I will readily admit so, however, out of the thousands of posts on this topic, my conclusion is firm.
|
@ghasley there you are wrong. Optimal audio quality over a IP network is fully understood by people like me. Do you again believe the banking is comprised daily? Or do you think super hi res simulcast concerts don’t know what they are doing? Look at Ravenna for instance.
The only potential issue is that noise from a short run Ethernet screws up a horribly implemented networking card, and if you remove that noise you could notice a difference. That premise means your super high end streamers use crap components, and that something like Etherregen removes noise. Well, noise is measurable and it doesn’t remove any noise at all. While I am sure that lots of manufacturers use bad components, but even todays bad components, are very good at handling bad noise environments.
|
@theaudiomaniac you specially called me out. My posts are accurate, networking is as close to absolutes as you can come.
Your comment is more or less exactly what I said…
As for wrappers around UDP, most applications do build some sort of reliability wrapper around it. In the case of Roon, they specially call out the need to support DRM, which requires reliable transmission.
Other applications like gaming have indicators of packet loss, which is a reliability wrapper, if you didn’t, the game would either crash or just hang if you lost connection somewhere.
VoIP also have call quality indicators, another form of reliability wrapper. So, it is more common than not to use UDP and a reliability wrapper.
|
@theaudiomaniac so, you called me out. What is wrong in any of my posts?
to your posts, I would add while UDP does not provide reliability at the transport layer, it leaves that to the higher level protocols. Which is huge headache, and probably why Roon ultimately switched to TCP.
|
@ghasley of course I could be nicer, but is that how you react when someone says you don’t know what you are talking about and then just makes stuff up?
but you are correct. I can of course be nicer, and you are also right that I should be nicer.
|
@theaudiomaniac you are saying you don’t have any more knowledge on the topic and therefore no one else does either. Good argument. like I have stated before, I know more about the topic that anyone on this forum, including yourself, until proven that it is not so.
one thing you certainly did not was to dispute any absolutes, in fact you, yourself said that the forum is anti Audioscience…
What you don’t call a reliability wrapper is up to you, UDP specifically states it leaves reliability to the higher level protocols. VoIP measures it for quality, gaming and gamers are obsessed with their “ping” there are controls. Neither requires that every packet is received, but have reliability safeguards to ensure a minimum level of performance.
And you are wrong when you say it is a wrapper around UDP, it is on top of UDP. Also wrong when you say the purpose of a reliability wrapper is to manipulate the data, it is literally the opposite, to ensure data arrives intact. How much data? well that is up to each individual application accepted performance level.
and finally, it is just flat out a lie that Roon did not support Tidal/Quboz before they switched to TCP. They most certainly did, and it is easily verifiable too. Which means they had to have a reliability wrapper on top of UDP to ensure DRM, proving you wrong again.
in summary @theaudiomaniac , yes, you do have limited knowledge, and can’t even bother to do research. We can agree there.
|
@sns John is not a networking expert or anything networking. He is, as he says himself, a silicon chip designer.
”John Swenson is widely known as a talented engineer, with decades of experience in silicon chip design (he was a senior project lead at a major integrated circuit firm for 30 years).”
|
@yyzsantabarbara EtherRegen doesn't support QOS, so there is that. Need managed switches for that, and EtherRegen is not managed. And while you can configure a lot, most QOS configuration do not limit or prioritize anything until you hit the defined congestion threshold. However, EtherRegen, if you have a congested network, by design, can cause issues, since you have "A" ports that are gigabit, and the "B" port that is supposed to go your streamer/dac, which is 100Mbps. If you are pushing over 100Mbps on the "A" side, and for some reason push it to the "B" side, you will introduce packet drops. It is even possible that the EtherRegen will flood the "B" side with everything going on on the "A" side. UpTone does not give any type of specifications for this, or any other specifications at all, only "ultra low jitter" and "ultra low noise" without and numbers.
What we do know about the UpTone is that they use cheap components according to their own pictures of the circuit board.
Did you ever think about this in a objective manner?
1) Uptone provides 0 specifications
2) There is no measurable improvement to their claims
3) There is no theoretical improvement to their claims
4) They use cheap components
5) Some people hear a different, but others do not.
Maybe there is a confirmation bias involved, maybe?'
You are the one who brought in the money, I get it that you feel stupid when you thought I had a $500 marantz receiver from BestBuy with a Klipsch $300 bookshelf pair, but if you want to argue that Classe Audio for Pre/Power is not good quality, you are in a league of your own, and will really feel stupid.
|
@yyzsantabarbara here you go again, John Swenson is NOT an audio streaming expert. That is a lie. 100% lie. Again, from UpTone’s website “John Swenson is widely known as a talented engineer, with decades of experience in silicon chip design” .
Stop lying. Please. John Swenson is a silicon chip designer, which is not even remotely close to audio streaming expert.
and your post never make any sense, if you lose a packet, how the hell can you regurgitate that lost packet, and the associated sounds in that packet?
However, the correct use of regurgitate is @yyzsantabarbara regurgitates things he reads on the Internet, meaning you just repeat statements out of context without understanding and comprehension of a topic you have claimed to program….
Regarding the streaming, TCP is incredibly reliable, and some of the topics that come into play are latency, window size, and yes, jitter. QOS has nothing to do with it as you can’t run QOS over the Internet (sure, you can tag your packets, and your internet provider will respectfully give you the finger and remove the tags, unless you pay for this, and then it is still only within that provider’s network and not the next network you need to traverse). It is all TCP.
and finally your comment about the 2k system. That is final sign of defeat, nothing factual backs up your claims, so now the listeners system isn’t resolving enough, spend more money. Well, I have a 45k system that I really enjoy, and I use enterprise grade networking equipment from juniper and extreme, and I certainly couldn’t hear a difference when I tried the Etherregen. Right now I am even running wifi to my NAD M50.2 due to a reno project, and no issues.
|
@sns no, not at all. It is a mature technology far beyond your comprehension. It is used in pro audio and pro video daily. You don’t understand, along with others on this forum, but calling it immature is just a spectacular show of your ignorance.
you don’t understand it, but it is used daily at level far beyond your reach.
I would advise you to open your eyes, read up on the topic, I have given you numerous sources to educate yourself, yet you chose to be ignorant.
|
I encourage anyone to learn, and as you learn, you understanding will evolve. For pro audio, which have been using high quality streaming for 20+ years, there are two standards that most seems to use, AES67 and SMPTE ST 210. Now, I am not an expert at either protocol, but I have designed and built networks to support these protocols across the US and across the Atlantic.
Companies like Ravenna puts both of these into practice, and provides a wealth of information, and they don’t use products like Etherregen, but if you really really want to twist it, they do use optical for SQ, since the distances involved far exceeds copper Ethernet capability, so if you want to twist it you could say that pro audio uses optical to provide high resolution audio and improve SQ over Ethernet copper, since copper would not transmit anything at all.
but for short runs, all regular Ethernet.
|
@jumia no humor, just harsh reality. @sns keeps mentioning that “we” don’t know anything streaming yet and that it is a immature technology. That is a lie.
why call it a lie instead of no true? Because I have given him links to educate himself, especially on the pro audio applications and protocols, but he insists. So his statements are then simply a deliberate lie. Why did I say it is beyond him? Well, obviously, he either can’t understand protocols used, either in consumer audio or pro audio like Ravenna, or he desperately want to hold on to his limited world view, which then, clearly, the truth is far beyond him.
|
@sns there you go, that is my point. You made a statement that is wholly inaccurate. I gave you information to educate yourself, and you continued down the same path.
how you perceive me is up to you, I just have no patience for ignorant people, especially people who chose to be ignorant. While that is not popular in today’s political climate where being ignorant and stupid is cool and everyone has an equal voice on all topics, I don’t subscribe to this, rather, I find it appalling.
I don’t know a lot about a lot of the topics discussed on this forum, but like I have said, I know more than anyone on this forum about this particular topic. Or at least until proven that it is not so. I am not apologizing for this, nor do I have to play nice with people who argue without any knowledge or sources, or regurgitate things from a webpage that they don’t understand, or outright lie.
|
@sns no, that is not accurate. You still can expect incremental improvements of mature technologies, and I assume a lot more metadata will be integrated in the future to augment the listening experience.
Certainly intentions to be malicious in any posts, but if you can’t call out people on BS and at least try to have a fact based discussion, what is the point?
I did miss that you had acquire a microtik, although, I wouldn’t call that enterprise. Better than consumer, for sure, so a step in the right direction. So that begs to question, what is “enterprise” grade? Gartner defines it as follows: “Enterprise-grade describes products that integrate into an infrastructure with a minimum of complexity and offer transparent proxy support.”
so what does that mean? Mainly that the products are not designed to operate a single unit, like an access point or switch for a consumer, rather a part of a total solution, solving more than one particular issue. And with the appropriate support structure for this.
|
If you want more details on enterprise grade, Forbes defines it here:
|
@sns the point why I am recommending enterprise grade equipment is due to consistent and predictable performance. It’s been tested over and over again and the support from the company and reputation of the company is behind it.
but like you said, you don’t need that typically for home use, and your microtik is fine, a bit lower throughput that I would want with 16 ports and only 16gbit throughput, which means it can’t support all ports at full speed, as that requires 32gbit throughput (duplex, each port can transfer 1gbit in each direction). These specs matter if you want to minimize jitter and latency.
Another concern I have with microtik is the cpu vs asic/fpgu. Also doesn’t promote consistent latency and jitter. And then it is shared data and control plane, which means if you are doing configuration changes, it could impact performance.
anyway, there is a lot to this topic, but in most cases a netgear from bestbuy is fine.
|
@sns thats what it is about right? It is a hobby, so have fun with it!
|
