[bluecar]

@bluecar: I can only repeat what I said earlier: However, everything you say does not explain why there should be any noise/jitter/whatever left after it has been written and read from disk again.

I also have to strongly agree with @bfreedma since I am also working in the same field. Also there is another forum with science in the same where practical tests have been made to prove that ethernet really does not make a difference.

Fair enough.

I’m sorry, but your understanding of Ethernet and Wi-Fi are fundamentally incorrect and seem based on vendor marketing material rather than the 802.11 standard.

I’ve read the standards docs many times and understand them well (professional requirement). I’ve even contributed to a few. They make it abundantly clear that what you describe, for example some rider on the carrier wave in the manner and impact you describe isn’t technically possible. You’re also ignoring all of the buffering and storage between the Ethernet end point and analog or digital output.

I‘m amazed that vendors have literally created a market for solutions to problems that don’t exist in any reasonable home setting.

...Is this where you tell me you're a engineer?

@bluecar: I can only repeat what I said earlier: However, everything you say does not explain why there should be any noise/jitter/whatever left after it has been written and read from disk again.

I also have to strongly agree with @bfreedma since I am also working in the same field. Also there is another forum with science in the same where practical tests have been made to prove that ethernet really does not make a difference.

Is that the field where you believe that writing a file to a disk or buffering it in memory means that all RF noise from the various connections in the device and its circuits is suddenly rejected? Which field is that

On the final point in the "science" forum - all they really proved is that their testing regime is inconclusive. If someone can prove to me that no RF noise flows through a set of connected ethernet devices, or at least, lets say, below an average of 10uV RMS and within the range 50hz to 5gHz, then I'll nominate you for the noble prize myself D

 

[bfreedma]

Fair enough.


...Is this where you tell me you're a engineer?


Is that the field where you believe that writing a file to a disk or buffering it in memory means that all RF noise from the various connections in the device and its circuits is suddenly rejected? Which field is that

On the final point in the "science" forum - all they really proved is that their testing regime is inconclusive. If someone can prove to me that no RF noise flows through a set of connected ethernet devices, or at least, lets say, below an average of 10uV RMS and within the range 50hz to 5gHz, then I'll nominate you for the noble prize myself D

Please explain how, if it even did make a difference at all, RF noise can be preserved in disk storage or ram buffer, then passed on through the processing chain. Be specific.

If you were right, we would see huge amount of error correction in the original reads on all Ethernet data transfers. A simple look at a resource monitor will show this is demonstrably not a problem. Issues of that magnitude would make the current digital universe virtually impossible due to excessive error correction traffic eventually creating more latency than could be sustained.

Bluntly, this is getting absurd

 

[bluecar]

Please explain how, if it even did make a difference at all, RF noise can be preserved in disk storage or ram buffer, then passed on through the processing chain. Be specific.

If you were right, we would see huge amount of error correction in the original reads on all Ethernet data transfers. A simple look at a resource monitor will show this is demonstrably not a problem. Issues of that magnitude would make the current digital universe virtually impossible due to excessive error correction traffic eventually creating more latency than could be sustained.

Bluntly, this is getting absurd

I agree with you, and I appreciate the points you make

To be fair, I never said that RF noise can be preserved (or stripped away) by writing to storage or buffering - that was TiJo - and that *is* an absurd statement, logically and physically.

Re: error correction, latency etc, you are 100% right. Ethernet devices reject noise in the frequencies that affect successful tx/rx of packets modulated into the carriers. However, outside of those frequencies, nothing much is done (why bother? it can't affect data transmission).

Step down from the logical layer of these circuits and into the electronics.....

Every cable (ethernet, power, signal, whatever) forms part of the electrical circuit it's connected to. It could be a bellwire, and it would be capable of carrying a current between points of potential. That's physics.

Ethernet devices (adapters/switches, *not* cables) have some circuitry to attenuate and reject some noise, but not all noise - just enough to assure network function by keeping high SNR in the frequencies that matter.

A better-built ethernet cable, with fancy shielding, terminations and conductors will carry data no better than a $2 patch cable of the same length. BUT it will collect and retransmit slightly less non-signal noise than a cheap-o patch cable.

Likewise, you can buy a fancy switch, with Mundorf capacitors, an LPSU and a hand-crafted OCXO clock - will it improve your networking? No. Will it generate/propagate less RF noise. Yes.

less noise in the circuit = more accurate conversion of data from files, to time-series streams. As soon as you slice a file up and lock it into a set time period, any errors cause that slice to be ignored by the DAC. hence, why noise affects SQ. In isosync, the data will be right (compared to the source file), but the *timing* will be wrong - this is where the noise causes the errors.

So the answer to the OPs question is NO, for networking, and YES for SQ.

I'm tired of this too - can we talk about Mains Cables and SQ, just for a rest

 

[bfreedma]

I agree with you, and I appreciate the points you make

To be fair, I never said that RF noise can be preserved (or stripped away) by writing to storage or buffering - that was TiJo - and that *is* an absurd statement, logically and physically.

Re: error correction, latency etc, you are 100% right. Ethernet devices reject noise in the frequencies that affect successful tx/rx of packets modulated into the carriers. However, outside of those frequencies, nothing much is done (why bother? it can't affect data transmission).

Step down from the logical layer of these circuits and into the electronics.....

Every cable (ethernet, power, signal, whatever) forms part of the electrical circuit it's connected to. It could be a bellwire, and it would be capable of carrying a current between points of potential. That's physics.

Ethernet devices (adapters/switches, *not* cables) have some circuitry to attenuate and reject some noise, but not all noise - just enough to assure network function by keeping high SNR in the frequencies that matter.

A better-built ethernet cable, with fancy shielding, terminations and conductors will carry data no better than a $2 patch cable of the same length. BUT it will collect and retransmit slightly less non-signal noise than a cheap-o patch cable.

Likewise, you can buy a fancy switch, with Mundorf capacitors, an LPSU and a hand-crafted OCXO clock - will it improve your networking? No. Will it generate/propagate less RF noise. Yes.

less noise in the circuit = more accurate conversion of data from files, to time-series streams. As soon as you slice a file up and lock it into a set time period, any errors cause that slice to be ignored by the DAC. hence, why noise affects SQ. In isosync, the data will be right (compared to the source file), but the *timing* will be wrong - this is where the noise causes the errors.

So the answer to the OPs question is NO, for networking, and YES for SQ.

I'm tired of this too - can we talk about Mains Cables and SQ, just for a rest

Again sorry to be blunt, but you have no idea what you’re talking about. You’re literally making up problems that don’t exist in order to spend money on a “solution”.

If you can produce objective measurements from an independent source showing any of the above phenomena remotely close to audibility, I’d be happy to continue to discuss.

As if data packets containing music data are somehow different than packets containing any other kind of data that somehow retains external noise post storage and buffering as they are processed... Audiophile Bro-Science

 

[endless402]

i think you are having trouble separating noise and bits... the noise isn't in the 0 and 1's, it's the devices that are adding noise which passes through the devices and eventually into your output since they are all connected

 

[TiJo]

To be fair, I never said that RF noise can be preserved (or stripped away) by writing to storage or buffering - that was TiJo - and that *is* an absurd statement, logically and physically.

Okay, but if you agree with me here. How does anything matter that comes before the read from disk?
Essentially you agree, that RF noise can be removed by writing to disk? So why care about RF noise and cables then?

(And let me say: I even disagree that there is accumulating RF noise over the network)

 

[TiJo]

i think you are having trouble separating noise and bits... the noise isn't in the 0 and 1's, it's the devices that are adding noise which passes through the devices and eventually into your output since they are all connected

If this was the case sooner or later bit flips would occur. However, in practice this does not happen.

 

[endless402]

Again, noise has nothing to do with the bits...

Even the type of hard drive sounds different.
A song playing from my external hard drive plugged into my streamer is different than one playing from the internal ssd. I’m sure the 0 and 1s are the same but one is powered by a high quality isolated power supply and one is through a wallwort

 

[bluecar]

Again sorry to be blunt, but you have no idea what you’re talking about. You’re literally making up problems that don’t exist in order to spend money on a “solution”.

If you can produce objective measurements from an independent source showing any of the above phenomena remotely close to audibility, I’d be happy to continue to discuss.

As if data packets containing music data are somehow different than packets containing any other kind of data that somehow retains external noise post storage and buffering as they are processed... Audiophile Bro-Science

**Read this bit carefully**....Like you, Im sorry to be blunt, but you didn't read what I wrote. - I never said, nor ever would, that packet data can retain external noise, post-processing - that's a preposterous notion - and that came from another poster, not me - "Bro-science" as you rightly say.

noise affects D/D conversion in the time domain, That's not my opinion, it's a fundamental of DSP. I don't feel the need to share papers or proofs for a concept that's there to read in any yr1 uni textbook for any electrical/electronics engineering course.

Equally, the idea that non-signal noise cannot propagate through an ethernet network defies the physical realities of electronics - it can be attenuated, but not eliminated.

The OP asked "switches/cables - is it worth it?"

My view is that there is a sound set of reasons why better switches and cables could result in better SQ, and as I've said several times, it has nothing to do with the receipt or transmission of packet data, but the role of the switches and cables as noise sources and carriers in the overall system. Ultimately of course, SQ is subjective, and tough to test - if you want to get into ABX or DBT then the Sound Science board will fill that need for you. I've seen good results using modestly priced, well-made cables and making some teaks to pretty basic cisco switch - cost about $30 to do. Maybe high end, multi-hundred $$cables work better - doubt it, but open to the idea if I can test them myself.

 

[bluecar]

Okay, but if you agree with me here. How does anything matter that comes before the read from disk?
Essentially you agree, that RF noise can be removed by writing to disk? So why care about RF noise and cables then?

(And let me say: I even disagree that there is accumulating RF noise over the network)

No...I don't agree on the 'removal of RF noise by writing to disk - how ridiculous.

And as for RF noise propagating through a network, you may disagree with that - I will respect your view, in the same way as I would respect your view that the earth is flat.

 

[bfreedma]

Okay, but if you agree with me here. How does anything matter that comes before the read from disk?
Essentially you agree, that RF noise can be removed by writing to disk? So why care about RF noise and cables then?

(And let me say: I even disagree that there is accumulating RF noise over the network)

I don‘t care about RF noise and cables. You may be confusing my posts with someone else.

 

[cpurdy]

Please dont try internet censur. You have bias that it cant sound different. You want to think what you want to think.

*censor

My bias is towards truth. There are others, on this thread, whose bias is towards ripping gullible people off.

It would be great to hear a bit more about your system and listening tests? i assume you've done the due diligence, tried these products or similar and you are speaking from a position of experience. Otherwise you risk appearing little highhanded dismissing someone else's experience just base on what you think.

I don't have a system. I have too many systems. Because I have the same disease that everyone else here has

My main office has a pure sine wave AC->DC->AC filter feeding an Ares II into a WA5-LE with a half a dozen different cans, but mostly using the Ether C's balanced. Where I am now, I'm using a portable FLAC DAC with balanced output into some cheaper 1060c planars (bass heavy fun) or custom Laylas (blocking out the rest of the world).

I love beautiful cables and crazy nice gear, just like the next person. And I generally don't mind spending too much to get some desired result, aesthetic or otherwise. And it absolutely makes things sound better, because I want it to. My typical A/B test is A="I don't have what I want, it sounds flat or awful" and B="I have what I want, it sounds awesome".

Misinformation? I am not reporting the news, I am reporting my reality and my experience.
Regardless of the technical theory, there are many people hearing differences between network cables and in product group tests. You obviously believe they are delusional. I certainly am not.
So when you buy any new product, do you have someone to assist in a blind listening test? If yes then lucky you!

Sure, we already know that people will hear the same exact sound in different ways based on their expectations. I don't doubt that people hear different things if the test is not blind A/B. I don't see anything wrong with that. As I've said before, the important thing is enjoyment. But misleading people on purpose to take their money is wrong, so I don't mind speaking up when I see that happening.

I wish I could do real blind A/B tests, but I can not. (I suppose I could buy the equipment to do so, but I just don't want to.) I like doing my A/B tests by myself, and letting my various biases screw up the results to my own liking. There's nothing wrong with bias, and there's nothing wrong with buying fancy toys that one does not really need. But there is something wrong about being dishonest about something in order to swindle people out of money, so I tend to say something when I see that.

If my file can be transmitted bit-perfectly via cheap cables and routers, why does a music file (e.g. an mp3) need better hardware? Since it is bit-perfect, the only bottleneck should be the connection to the DAC right?

A fitting question. Trust me, though, it will sound better if the cables look better. I have personally tested this over and over and it's clearly true. Even if scientifically there is absolutely no difference in the sound.

It's not a gross effect, but it definitely causes incorrect blocks in the USB stream, that are ignored by the DAC - tiny errors in the DAC inputs cause tiny errors in transients at the DAC output, giving effects like digital 'grain', or 'darkness' or 'harshness', depending on how the stream was affected in the Streamer, and (bluntly) the quality of the DAC.

Yeah, no. It is possible with USB 1.0, maybe, if the bitstream is at the limit of the USB throughput and the clock is real-time, but since USB 2.0 (or maybe SS, I can't remember), this is impossible. Scientifically impossible.

So, to recap, yes, it was actually possible 25 years ago (1996) that USB could lose data and a DAC would have to fudge the analog output accordingly. Modern USB basically has no way (other than, in theory, emulating the deprecated protocols for old times' sake) to have this happen.

Cabling upstream of the player carries noise from every other network component to the ethernet receiver at the player - a bit-perfect copy of whatever music file you send it is delivered to the player, but the received noise affects how well that player converts it to time-series streams. Ethernet is not a digital connection (no such thing really exists) - in each ethernet cable, there are 2 pairs of conductors that carry data.

Yeah, no. It's not that this is theoretically impossible; it is theoretically possible. However, no one builds equipment in a manner that would allow it to happen. No one.

WiFi, by definition, is galvanically isolated. With no cable connection...

^ this!

regardless of use, an ethernet connection is a bunch of wires, wrapped into a cable that connects two network cards. Even if there is no data *at all* being sent, ethernet creates several circuits between the two devices. In addition to carrying signals, the ethernet cables will carry any noise in the signal between devices - the cables also act as aerials and collect noise via EMI - unless that noise is detrimental to the ethernet standard, it is largely unmanaged.

Correct. This is all covered by the IEEE 802 specs. Not a problem at all. Absolutely zero influence on the data. And any of this so-called electrical noise is 100% filtered out of the data. It is possible that a network card could experience such a substantial amount of noise that it would cause a ripple on e.g. the 5v rail, but that would probably be enough "noise" to fry everything in your PC and the entire room. Furthermore, the designs have gotten far more reliable over time, such that flaws that could theoretically be measured 30 years ago (back when we were using 10-base-2 or 300kb token ring) simply don't exist anymore. I can remember when a vacuum cleaner could mess up a network backup, for example.

again, forget the digital aspect of this - that noise transmits freely into your player.

Wrong. There is no logical or scientific basis for this claim.

it all comes down to the quality and shielding related to the power supply.

when i'm listening to my Cayin N6ii, when the wifi is downloading from tidal, i can hear it buzzing in my iem. If i turn off the wifi and play from internal storage, there isn't any buzzing. this is likely because the wifi chip is requiring more power and as a result, is injecting noise into the amps.

I can hear my phone when it is sending/receiving data. For example, in the middle of the night, the iOS does some various scheduled daemon activities, and it wakes me up with the little bits of buzzing inside of its radio and what not. Almost no one besides me can hear it, but it's clearly audible to me, and annoying. (My daughter, also a musician, can hear it as well.) Wireless modems are noisy. Both in terms of their electromagnetic abuses and in terms of the resonance noise that they can create. (I'm not a physicist; I'm not good at explaining what's going on here, because I don't fully understand it.)

An IEM is basically a big antenna with a super efficient means of turning electromagnetic signals into sound. It is super believable to me that your IEMs are picking up interference from your WIFI, and it is even conceivable (albeit unlikely) that your DAC is picking up interference from your WIFI. Your DAC power supply can also, in theory, be affected by substantial electromagnetic noise, although power supplies are supposed to filter that out, and DACs are supposed to filter out noise from their power supplies. So, I would guess that the noise that you hear is coming from the IEMs themselves, caused by the WIFI signals. Just like some people can listen to the radio via silver fillings in their mouths: https://mythbusters.fandom.com/wiki/Tooth_Fillings_Radio_Myth

noise affects D/D conversion in the time domain, That's not my opinion, it's a fundamental of DSP. I don't feel the need to share papers or proofs for a concept that's there to read in any yr1 uni textbook for any electrical/electronics engineering course.

You are using words that I know the meanings of, but you are putting them together in ways that I do not recognize. Also, when someone says "I don't feel the need to share papers or proofs", that's usually a dead giveaway that they are spouting nonsense. If you are not spouting nonsense, please take the time to translate what you were thinking into a sentence that an electrical engineer could recognize. Thank you.

 

[bfreedma]

*censor

My bias is towards truth. There are others, on this thread, whose bias is towards ripping gullible people off.



I don't have a system. I have too many systems. Because I have the same disease that everyone else here has

My main office has a pure sine wave AC->DC->AC filter feeding an Ares II into a WA5-LE with a half a dozen different cans, but mostly using the Ether C's balanced. Where I am now, I'm using a portable FLAC DAC with balanced output into some cheaper 1060c planars (bass heavy fun) or custom Laylas (blocking out the rest of the world).

I love beautiful cables and crazy nice gear, just like the next person. And I generally don't mind spending too much to get some desired result, aesthetic or otherwise. And it absolutely makes things sound better, because I want it to. My typical A/B test is A="I don't have what I want, it sounds flat or awful" and B="I have what I want, it sounds awesome".



Sure, we already know that people will hear the same exact sound in different ways based on their expectations. I don't doubt that people hear different things if the test is not blind A/B. I don't see anything wrong with that. As I've said before, the important thing is enjoyment. But misleading people on purpose to take their money is wrong, so I don't mind speaking up when I see that happening.

I wish I could do real blind A/B tests, but I can not. (I suppose I could buy the equipment to do so, but I just don't want to.) I like doing my A/B tests by myself, and letting my various biases screw up the results to my own liking. There's nothing wrong with bias, and there's nothing wrong with buying fancy toys that one does not really need. But there is something wrong about being dishonest about something in order to swindle people out of money, so I tend to say something when I see that.



A fitting question. Trust me, though, it will sound better if the cables look better. I have personally tested this over and over and it's clearly true. Even if scientifically there is absolutely no difference in the sound.



Yeah, no. It is possible with USB 1.0, maybe, if the bitstream is at the limit of the USB throughput and the clock is real-time, but since USB 2.0 (or maybe SS, I can't remember), this is impossible. Scientifically impossible.

So, to recap, yes, it was actually possible 25 years ago (1996) that USB could lose data and a DAC would have to fudge the analog output accordingly. Modern USB basically has no way (other than, in theory, emulating the deprecated protocols for old times' sake) to have this happen.



Yeah, no. It's not that this is theoretically impossible; it is theoretically possible. However, no one builds equipment in a manner that would allow it to happen. No one.



^ this!



Correct. This is all covered by the IEEE 802 specs. Not a problem at all. Absolutely zero influence on the data. And any of this so-called electrical noise is 100% filtered out of the data. It is possible that a network card could experience such a substantial amount of noise that it would cause a ripple on e.g. the 5v rail, but that would probably be enough "noise" to fry everything in your PC and the entire room. Furthermore, the designs have gotten far more reliable over time, such that flaws that could theoretically be measured 30 years ago (back when we were using 10-base-2 or 300kb token ring) simply don't exist anymore. I can remember when a vacuum cleaner could mess up a network backup, for example.



Wrong. There is no logical or scientific basis for this claim.



I can hear my phone when it is sending/receiving data. For example, in the middle of the night, the iOS does some various scheduled daemon activities, and it wakes me up with the little bits of buzzing inside of its radio and what not. Almost no one besides me can hear it, but it's clearly audible to me, and annoying. (My daughter, also a musician, can hear it as well.) Wireless modems are noisy. Both in terms of their electromagnetic abuses and in terms of the resonance noise that they can create. (I'm not a physicist; I'm not good at explaining what's going on here, because I don't fully understand it.)

An IEM is basically a big antenna with a super efficient means of turning electromagnetic signals into sound. It is super believable to me that your IEMs are picking up interference from your WIFI, and it is even conceivable (albeit unlikely) that your DAC is picking up interference from your WIFI. Your DAC power supply can also, in theory, be affected by substantial electromagnetic noise, although power supplies are supposed to filter that out, and DACs are supposed to filter out noise from their power supplies. So, I would guess that the noise that you hear is coming from the IEMs themselves, caused by the WIFI signals. Just like some people can listen to the radio via silver fillings in their mouths: https://mythbusters.fandom.com/wiki/Tooth_Fillings_Radio_Myth



You are using words that I know the meanings of, but you are putting them together in ways that I do not recognize. Also, when someone says "I don't feel the need to share papers or proofs", that's usually a dead giveaway that they are spouting nonsense. If you are not spouting nonsense, please take the time to translate what you were thinking into a sentence that an electrical engineer could recognize. Thank you.

Nice synopsis of a lot of topics.

I was going to mention that there seemed to be some confusion between older bus based networks and modern packet switched networks but didn’t think anyone else would remember ARCNet and Token Ring. We’re older than most here, I suspect. (or I’m old and you really know legacy network topologies).

 

[endless402]

I should clarify, when I said Wi-Fi, I meant the Wi-Fi card being galvanically isolated from the rest of the components via the power supply...

 

[bluenight]

I’m sorry, but your understanding of Ethernet and Wi-Fi are fundamentally incorrect and seem based on vendor marketing material rather than the 802.11 standard.

I’ve read the standards docs many times and understand them well (professional requirement). I’ve even contributed to a few. They make it abundantly clear that what you describe, for example some rider on the carrier wave in the manner and impact you describe isn’t technically possible. You’re also ignoring all of the buffering and storage between the Ethernet end point and analog or digital output.

Its an easy thing to say,
prove it then.
Like the 802.11 standard is the Bible and answer to everything like how the physics of how electricity works and how noise can travel in it.

I'm tired of this too - can we talk about Mains Cables and SQ, just for a rest

You did great explaning how electrical noise can travel in a chain in an relativly easy way to understand. I am saving it in my notebook incase i forget how it works.

My bias is towards truth.

How do you know your not spreading misinformation about noise in electricity is not an real thing effecting audio in a music system chain with digital devices for us to hear? Thats Not effecting the data in ethernet like @blucar explained.
Prove it.
More peoples listening experiences shared online points more to its real, Yes its good to be aware of placebo and what is a real change or not, this i have learned from this thread. An obvious big change i would be certain about like a constant change in tonality for exe.