Do router and ethernet cables affect sound quality?

Please get someone else to swap the cables without showing or telling you which one is connected ( also not showing you which one is not connected :cool: )
Just call them cable 1 and cable 2
I definitely would be interested in your observations!

Blind tests are not very reliable. It may happen that there are so little and minute differences in sound that are not obvious in short blind tests. I fail most MP3 and lossless tests online but still am sure wav sounds better than MP3. How is this explained.
 
Blind tests are not very reliable. It may happen that there are so little and minute differences in sound that are not obvious in short blind tests. I fail most MP3 and lossless tests online but still am sure wav sounds better than MP3. How is this explained.
Your basically saying that I believe what I believe despite evidence going the other way. If you can't tell the difference in blind test, then there's no difference. Period.

If you closed your eyes while eating, would you not be able to tell the difference between roti and paratha?
 
<snip>
a) Data related issue ( called as jitter, timing issue etc) --> This is easy to understand for most
b) Current / voltages related issue in the signal (also called as noise / current noise / minor fluctuations / not the perfect analog signal)
--> This point #b is i was not aware in the past, and it took some time for me to grasp its understanding.
--> This point #b is where the issue happens even if point #a is perfect.
All of this was solved in networking at speeds of 100Mbps 25 years ago.
Networking does not rely only on physical line quality. There are protocol level redundancy and error checks built into data transmission.

We often have the view that "digital is digital" and so passive network elements like Ethernet cables should all work the same (except for bandwidth differences inherent in the various generations of Ethernet cables). Or active network elements like routers and switches when used in audio specific networks should not produce perceptible sonic differences, etc., but the strange fact is they all produce different sonics, sometimes slight and subtle, but often very obvious.

<snip>
In networking "digital is digital". There is no chance of a few bits flipping or losing a few bits here and there.
If there is loss, it is a whole bunch of packets. If data integrity is paramount, protocols like TCP have re-transmission mechanisms.
A data file is "packetized" so that it does not have to be sent in one go or in sequence. Algorithms like Segmentation And Reassembly allow for out of order transmission, re-transmission only to be reassembled correctly at destination.

When a digital signal is sent over the network, over hundreds or thousands of miles hopping through multiple routers, the serialized approach does not work. TCP/IP which is the predominant networking concept has been in vogue since the 70s (it started as a DARPA project). Packetization and upper/application layer checks make sure that data integrity is maintained, this is absolute. Otherwise, imagine conducting a bank transaction or using GPS or relying on networked medical devices.

When a digital signal is sent from player to DAC, it is not packetized, it is serialized and in sequence. Data can be corrupted or misread by the DAC, though it is unlikely.

I believe if many of our FMs understand that psychoacoustics play a big role in our subjective assessements, most of the difference of opinion and perspectives …..including towards measurements will disappear.

Objective….and hence measurements becomes important when we want to compare and have an idea of transparency not biased by subjective bias.
Psychoacoustics is real when it involves the senses, i.e hearing. That is why this study is about how we hear and why we hear it in a one way or another.
Networking is not related to this aspect at all. It is just about transferring data from point A to point B with absolute certainty.

Cheers,
Raghu
 
Pretty sure there will be a few people who believe it does. Just a few days back there was a discussion about an "audiophile" SSD and the discussion got pretty heated at times. I eventually stopped following the thread after 60-70 replies.

I'm not very technically knowledgeable, but my understanding is that till the data gets converted to analogue, the chain shouldn't affect sound quality at all. This is because as long as the data is digital, it's just 1s and 0s. There are checks and balances at every stage to ensure that these 1s and 0s are not lost during transport. So regardless of what you have in your chain, the end result is same. For video, the result can get affected because of speed, but in audio, mostly that's not an issue, most modern transport layers (except Bluetooth) are way too fast to bottleneck audio streams.

The audio quality can start getting affected from the DAC stage onwards since it's getting converted to analogue here and hereon it's voltage variance and not 1s and 0s. So there is no way for the receiving device to counter any loss in fidelity.

Of course, there are extraneous factors which have nothing to do with the data being digital. For instance, if you feel that CDs make a spinning sound while vinyl doesn't, that's got nothing to do with CD being digital and just the mechanism of how the stack operates.
It is not only about how faithfuly the data can be transfered from one point to another. There is another important factor is involved. The amount of noise it adds up with the data. Generally with wifi modules the noise is higher than a wired connection. That makes the sound little un-clean. But if a streamer is designed to work wirelessly with a built-in wireless module then that shuld not be a problem as the module may be designed to induce low noise.

I am not surprised at people laughing at my answer, OR giving explanation about computer screen/pdf files etc.

I too was in the same boat few months ago, and I also did not belive that router / ethernet cable can affect the final sound. So at that time I had decided to experiment myself, and I was able to find some difference, thats it. If you can hear the difference fine, if you cannot then also its fine. Experiment and try and find it out yourself.

Having said that, i will try to make more clear, as it got clear to me gradually over the time.

We are talking about 2 things. First point is easy to understand and digest for many. It is the second point which most people do not focus on and miss the point.

a) Data related issue ( called as jitter, timing issue etc) --> This is easy to understand for most

b) Current / voltages related issue in the signal (also called as noise / current noise / minor fluctuations / not the perfect analog signal)

--> This point #b is i was not aware in the past, and it took some time for me to grasp its understanding.
--> This point #b is where the issue happens even if point #a is perfect.
You are correct. Anything you put in the signal chain will have the effect on it.

Yes wifi routers/lan cables everything will have their effect on sound quality.

For an example if a router transmits data with a lot of added emf noise then the audio streamer will also receives that. It will have it's noise filters but still how clean the output will be depends on it's capability and amount of noise added from router. The same rule also applies to wired connection.

So better routers and better streamers will perform better together. Btw some streamers perform better on wired connection and some on both type connections and even some on wireless connection. It all depends on the device's particular ciruit design.
 
It is not only about how faithfuly the data can be transfered from one point to another. There is another important factor is involved. The amount of noise it adds up with the data. Generally with wifi modules the noise is higher than a wired connection. That makes the sound little un-clean. But if a streamer is designed to work wirelessly with a built-in wireless module then that shuld not be a problem as the module may be designed to induce low noise.
You are correct. Anything you put in the signal chain will have the effect on it.

Yes wifi routers/lan cables everything will have their effect on sound quality.

For an example with a routet transmits data with a lot of added emf noise then the audio streamer will also receives that. It will have it's noise filters but still how clean the output will be depends on it's capability and amount of noise added from router. The same rule also applies to wired xonnection.

So better routers and better streamers will perform better together.
Wow!!
We badly need a "bang your head against the wall" emoji.

Guys,
Don't theorize just because you have a keyboard, display and network connection.
The same network connection you used to type posts on this forum. It came through and got posted, right?

TCP/IP is reliable down to the last bit. If there is data corruption anywhere in the network segment, packets are dropped or culled.
They don't end up with noise added, just because you believe so.

If you still want to buy a fancy router and fancy ethernet cable, please go ahead.
It's your money, who are we network engineers to deny you the pleasure of pristine, unmolested-by-noise audio.

Cheers,
Raghu
 
Last edited:
Wow!!
We badly need a "bang your head against the wall" emoji.

Guys,
Don't theorize just because you have a keyboard, display and network connection.
The same network connection you used to type posts on this forum. It came through and got posted, right?
TCP/IP is reliable down to the last bit. If there is data corruption anywhere in the network segment, packets are dropped or culled.
They don't end up with noise added, just because you believe so.

If you still want to buy a fancy router and fancy ethernet cable, please go ahead. It's your money, who are we network engineers to deny you the pleasure of pristine audio.

Cheers,
Raghu
I do not think you got my opinion. I am not saying anyone to buy new router/lan cable. I pointed out why people think their wired connection sounds better than wireless connection.

Simply put on budget devices the wired circuit is better designed than it's wireless module.
 
Blind tests are not very reliable. It may happen that there are so little and minute differences in sound that are not obvious in short blind tests. I fail most MP3 and lossless tests online but still am sure wav sounds better than MP3. How is this explained.
Well, if done properly you can ensure that all the usual subjective biases are removed.
What remains is whatever is reported.
The results may or may not be pleasing to us.
But that’s not what the test it’s for.
The results are what they are.
It’s up to us to then try and figure out why.

Consider your example: there maybe a number of physiological or technological reasons why you could not distinguish between a mp3 and a WAV file of the same recording played on the same system, same room, one right after the other.

However our well informed brain tells us that there is more information in a WAV file than in a mp3 file and this should be apparent when we listen.

If the test results don’t match out expectations we humans commonly tend to dismiss the test. But the test only can achieve what it was designed to do. Nothing more or nothing less.
 
Psychoacoustics is real when it involves the senses, i.e hearing. That is why this study is about how we hear and why we hear it in a one way or another.
Very true. But this flaw in hearing affects almost the majority and hence has not been classified as a mental disorder. Just replace hearing with vision, the moment I see things that are not actually there, then I will either be classified as an 'enlightened one' or the 'blessed one' and with some luck I may even found a major religion. But in today's age I will be regarded as someone requiring treatment. It took someone like Sigmund Freued to convince medical fraternity to treat such symptoms. Imagine if Amir had founded Video Science Review before Sigmund Freued and others were born? We would have been blasting Amir and VSR and claim science cannot explain everything that we see.
 
I can see this thread is converging towards and will soon meet up with another thread that is several pages long titled “objective vs subjective (AKA beating a dead horse) :)
 
I do not think you got my opinion. I am not saying anyone to buy new router/lan cable. I pointed out why people think their wired connection sounds better than wireless connection.

Simply put on budget devices the wired circuit is better designed than it's wireless module.
There is no guesswork here.
Wired is always full duplex, ideally. Wireless/WiFi is always half duplex because air is a shared medium.
When using wired, you are guaranteed some amount of bandwidth for audio application.
In WiFi, this is not always true. WiFi uses channel hopping and also has something called "backoff". Various algos are used here.
Whatever be the case, TCP/IP gives you data integrity. If anything, due to a poor connection, there may be packet loss.
In audio, if packet loss occurs and is not recovered in time, there is a buffer under run. It means no music for a split second or an abrupt jump.

VoIP calls have these nulls. Where you say "Hello, helloooo ..." and look at your screen with annoyance. A possible cause is packet loss.
Late packets are not very useful in a live call.

Streaming services may also use adaptive streaming. It can gauge the network speed up to the app and decide to stream at lower bit rate.
Hence you may (or may not) hear a difference.

As a guy who has worked a quarter of a century in networking, designed and tested numerous HW chips (for applications all the way from modems to full blown data centers), this is all I can say.
"Have fun!! May the pixies rule"
Cheers,
Raghu
 
Last edited:
There is no guesswork here.
Wired is always full duplex, ideally. Wireless/WiFi is always half duplex because air is a shared medium.
When using wired, you are guaranteed some amount of bandwidth for audio application.
In WiFi, this is not always true. WiFi uses channel hopping and also has something called "backoff". Various algos are used here.
Whatever be the case, TCP/IP gives you data integrity. If anything, due to a poor connection, there may be packet loss.
In audio, if packet loss occurs and is not recovered in time, there is a buffer under run. It means no music for a split second or an abrupt jump.

VoIP calls have these nulls. Where you say "Hello, helloooo ..." and look at your screen with annoyance. A possible cause is packet loss.
Late packets are not very useful in a live call.

Streaming services may also use adaptive streaming. It can gauge the network speed up to the app and decide to stream at lower bit rate.
Hence you may (or may not) hear a difference.

As a guy who has worked a quarter of a century in networking, designed and tested numerous HW chips (for networking applications all the way from modems to full blown data centers), this is all I can say.
"Have fun!! May the pixies rule"
Cheers,
Raghu
Again!! You are eyeing the software part only. Hardware also is important. Circuit design and implementation is vital in any hi-fi category. That is why a raspberry based streamer is cheap and a streamer from PS Audio is so costly, essentially both doing the same thing.

Don't you believe hardware also contribute to sound quality besides software?
 
Again!! You are eyeing the software part only. Hardware also is important. Circuit design and implementation is vital in any hi-fi category. That is why a raspberry based streamer is cheap and a streamer from PS Audio is so costly, essentially both doing the same thing.

Don't you believe hardware also contribute to sound quality besides software?
I would suggest you to read my posts all over again. Networking is not software alone. It is complicated software running on complex/dedicated hardware to give the user absolute data integrity when required.
The network part/segment of any streamer, cheap or expensive is completely reliable when the connection is "good".
Streamer, router, ethernet cable does not control data integrity. Protocols guarantee data integrity.
When connection is bad, packets may get dropped in the network or arrive late (dropped by the app/player because it is useless).

Put it another way:
Network path from Spotify/Tidal to RPi based playback app buffer is the same as Spotify/Tidal to PS Audio (or whatever) playback app buffer.
Audio data/file is not molested in this segment.
Once it enters the player (SW) and then on to the DAC or I2S, circuit noise and other artefacts can creep in.
Trust the network segment to deliver or not deliver. There is no such thing like it'll deliver with bits mangled due to noise.
If you can't trust this, stop using data networks in totality, not just audio.

Cheers,
Raghu
 
Blind tests are not very reliable.
Blind Tests like ABX tests don't have a Statistical Power to draw universal conclusions. In fact, They should not be used when a Expert Opinion for an Variable Acquired ability (Like Audiophile listening) is being utilised , The number of variables is quite huge and your sample size has to be equally huge and universal. In Short, absence of difference in Blind tests doesn't have any significance , statistically.

After ABX testing For Eg,
Finally Your Conclusion will be represented like this; that 'X' group of individuals(with Z level hearing accuracy), used 'Y' test tracks to Compare component 'A & B ', in set up comprising of 'PQRST' Devices, in a Room Acoustic Parameters 'MNO' with test duration of 'C' Minutes, with Component inter change time duration 'D"minutes & Conclude that There is no Acoustically discernible difference between component A&B :)
 
I would suggest you to read my posts all over again. Networking is not software alone. It is complicated software running on complex/dedicated hardware to give the user absolute data integrity when required.
The network part/segment of any streamer, cheap or expensive is completely reliable when the connection is "good".
Streamer, router, ethernet cable does not control data integrity. Protocols guarantee data integrity.
When connection is bad, packets may get dropped in the network or arrive late (dropped by the app/player because it is useless).

Put it another way:
Network path from Spotify/Tidal to RPi based playback app buffer is the same as Spotify/Tidal to PS Audio (or whatever) playback app buffer.
Audio data/file is not molested in this segment.
Once it enters the player (SW) and then on to the DAC or I2S, circuit noise and other artefacts can creep in.
Trust the network segment to deliver or not deliver. There is no such thing like it'll deliver with bits mangled due to noise.
If you can't trust this, stop using data networks in totality, not just audio.

Cheers,
Raghu
I never said data bits will be corupted with the noise. When a wire carries a signal it also transmits electrical noise. Do you think it delivers corupted data? No it does not, you are right. But with the data it also introduces emi noise and the signal that passes the data carries the noise.

Similarly the radio wave that carries data packets in a wireless connection also carries noise from router and other rfi sources.Then the whole signal gets to the wireless module of the streamer and enters it. That radio signal has both data packets and the transmitted noise and it does not have to alter the data itself.it can not.

Then the device's input signal filters do their job and the better device it is, the more it cleans the signal and process it further.

Tell me if I am wrong.
 
Blind Tests like ABX tests don't have a Statistical Power to draw universal conclusions. In fact, They should not be used when a Expert Opinion for an Variable Acquired ability (Like Audiophile listening) is being utilised , The number of variables is quite huge and your sample size has to be equally huge and universal. In Short, absence of difference in Blind tests doesn't have any significance , statistically.

After ABX testing For Eg,
Finally Your Conclusion will be represented like this; that 'X' group of individuals(with Z level hearing accuracy), used 'Y' test tracks to Compare component 'A & B ', in set up comprising of 'PQRST' Devices, in a Room Acoustic Parameters 'MNO' with test duration of 'C' Minutes, with Component inter change time duration 'D"minutes & Conclude that There is no Acoustically discernible difference between component A&B :)
That’s true. they are not for statistical assessments.
But they can be useful for us as individuals to explore and find out for ourselves.
 
I never said data bits will be corupted with the noise. When a wire carries a signal it also transmits electrical noise. Do you think it delivers corupted data? No it does not, you are right. But with the data it also introduces emi noise and the signal that passes the data carries the noise.

Similarly the radio wave that carries data packets in a wireless connection also carries noise from router and other rfi sources.Then the whole signal gets to the wireless module of the streamer and enters it. That radio signal has both data packets and the transmitted noise and it does not have to alter the data itself.it can not.

Then the device's input signal filters do their job and the better device it is, the more it cleans the signal and process it further.

Tell me if I am wrong.
Do you have a basic understanding of electrical circuit theory?
If so, you should be able to follow this.

In the network segment using line or RF, if the noise is so high so that SNR is low, data is not received correctly by the application.
So these segments are designed to have an acceptable SNR. No circuit element is cleaning the signal, it just retrieves it for further processing.
In this case, store into a buffer for playback. If it can't retrieve it there will be an application error.

The processor that's running the streaming protocols is also hosting the player which in turn puts it out on I2S or USB or Coax or OPT using relevant circuitry. In practice the noise introduced by a processor, RPi or fancy stuff, is much higher than network side line/RF noise.
Circuits are usually isolated from each other using good design practices such as proper ground plane design, trace shielding, etc.
Once the audio signal, processed by the player and transported (in sequence) over the transport interface, DAC kicks in.
In this segment there can be issues, if SNR is not good or you have a lousy cable or a SBC.

Correlating noise over the network to unacceptable SNR in audio path is .... a huge leap of faith, which many folks take with out understanding.
So they start changing routers, ethernet cables and what not. All I'm saying here is, the network segment is not the culprit here for poor sounding audio. It can be the player SW, the way I2S is laid out on the board, or USB implementation or something else.
Not the network segment, which is what the original post asks.

Do router and ethernet cables affect sound quality?​

The answer still stays true to what I said many posts earlier.
No, it does not.

Cheers,
Raghu
 
Do you have a basic understanding of electrical circuit theory?
If so, you should be able to follow this.

In the network segment using line or RF, if the noise is so high so that SNR is low, data is not received correctly by the application.
So these segments are designed to have an acceptable SNR. No circuit element is cleaning the signal, it just retrieves it for further processing.
In this case, store into a buffer for playback. If it can't retrieve it there will be an application error.

The processor that's running the streaming protocols is also hosting the player which in turn puts it out on I2S or USB or Coax or OPT using relevant circuitry. In practice the noise introduced by a processor, RPi or fancy stuff, is much higher than network side line/RF noise.
Circuits are usually isolated from each other using good design practices such as proper ground plane design, trace shielding, etc.
Once the audio signal, processed by the player and transported (in sequence) over the transport interface, DAC kicks in.
In this segment there can be issues, if SNR is not good or you have a lousy cable or a SBC.

Correlating noise over the network to unacceptable SNR in audio path is .... a huge leap of faith, which many folks take with out understanding.
So they start changing routers, ethernet cables and what not. All I'm saying here is, the network segment is not the culprit here for poor sounding audio. It can be the player SW, the way I2S is laid out on the board, or USB implementation or something else.
Not the network segment, which is what the original post asks.

Do router and ethernet cables affect sound quality?​

The answer still stays true to what I said many posts earlier.
No, it does not.

Cheers,
Raghu
Did I say that noise over network would make the output audio poor? I said better networking components can make subtle improvements in audio. If you disagree that is okay. If you say network has nothing to do with the noise that gets in to a network enabled device, that is okay too.
 
Did I say that noise over network would make the output audio poor? I said better networking components can make subtle improvements in audio. If you disagree that is okay. If you say network has nothing to do with the noise that gets in to a network enabled device, that is okay too.
If noise is the factor to decide "better" or not and to make "subtle" improvements in audio, you just contradicted yourself in back to back sentences.
You tried with your explanations.
I tried with my explanations.
In the end we both wasted some time on the forum.

Enjoy the music!!

Cheers,
Raghu
 
Oh man, going through this thread I can think of only 2 responses.

1. No, network hardware does not affect Sound Quality. period.
2. If it did, I would advise we all get an Audiophile Air Purifier so that it can Purify the air in our rooms so that the signals have a Purified medium to travel through.

MaSh
 
Jitter is only an issue at the DAC. Jitter in the Ethernet stream is not relevant. The reason is there is a buffer that collects the audio samples and that is clocked out at a constant rate. The buffer is in your media player and/or your DAC.

Ethernet is a packetized transmission anyway. The data arrives in high speed bursts at random intervals - extreme jitter! As long as the buffer has data in it, the output stream audio samples will be well within audible jitter standards.

Also as for noise, any Ethernet connection is transformer coupled. So there is no ground loop through an unshielded CATx cable. What is funny is that some audiophiles insist on using shielded Ethernet cables. That shield can actually cause a ground loop - the very thing they are trying to avoid! Shielded Ethernet cables are for industrial factory floor use where there are high levels of RFI. You typically won't find that in a home.
 
Back
Top