Network Improvements and their Impact on Sound Quality

[nenon]

Which makes me wonder how many people have tried 2 nics in a music server - 1 for access to the internet and 1 direct - connect to renderer versus connecting each component to a switch.

I wish there was an easy and simple answer to all these questions. I have done many of these tests myself and have some valuable information I can share, but it's never easy, because the answer is never Yes or No. Or good, better, worse.
Every time I see a question and want to jump on it to provide what I think would be valuable feedback, I realize I have to spend an hour to write an extensive answer that describes the different aspects of the problem. The main reason is that there are too many factors to consider with every change you make.
So without trying to answer this particular question, because I don't have an hour at the moment, I would give you a couple of pointers to think about. I'll do that in the form of questions that we should be asking ourselves and trying to answer when introducing a change like that.
- Will the two NICs be taking power from the motherboard, ot will they be powered by external power supply?
- If powered by the motherboard, wouldn't that pollute the power quality for the other components also taking power from the motherboard? Certainly as the NIC consumes power it generates some noise... how does that impact your overall sound quality?
- If powered separately, is there a way to isolate the ground? Or will the ground be common? If the ground is common, how do you prevent the noise that comes into one of the NIC to go to the other NIC or the rest of your system?
- Also if powered separately, what kind of power supply are you using? And how does that interact with the rest of your system?
- Could DC cables act like antennas? If yes, how adding another DC cable to power the additional NIC affects your system?
- Do Ethernet cables act like antennas? If they do, what happens with your system when you connect two Ethernet cables to your computer instead of one?
- Are you using the same NIC model or different? If different how does the second driver impact your computer?
- How is your OS impacted by having to deal with two NICs and how does that impact the sound quality?
- What is the overall impact on your computer sound quality by changing it from one NIC to two?
- Are you preventing some noise coming from your network by using two NICs? Or are you injecting twice more? Or somewhere in between?

All these questions are just touching the surface of what is happening at a micro level. I can easily dive deeper and provide 10 times more. And that would be all from a technical aspect not even taking into account personal preferences, musical taste, and your local environment. This is why some of those generic questions are difficult to answer. There is too much to consider, and it's very difficult to predict what the overall result would be.

On the other hand, there are things that are easy to recommend. For example, most home networks have a modem, router, firewall, WiFi access point, and a switch. ISPs often give you one device that can perform all 5 functions. You can think of this device as a surround receiver in HiFi. Those receivers are all-in-one devices too. They have a power amplifier, line stage, DAC, phono stage in some cases, etc. Chances are that replacing a cheap surround receiver with a good quality mono blocks, line stage, DAC, phono stage, etc would give you good results.
Well, the same way replacing the all-in-one ISP device with a good modem, a good router (i.e. Ubiquiti EdgeRouter), a good switch, etc. would give you better results for streaming. That's what makes this an easy recommendation. And when you do that, another easy recommendation would be to replace the cheap wall wart on each of those devices with a good power supply. And when you do that, another easy recommendation would be to replace the cheap $1 ethernet cables with better ones. But pay attention to what we are doing here - we are not really changing the design much. We are sticking to the design we have and upgrading the components. This eliminates the almost unlimited number of variables. As soon as you make a design change, we are back at the hundreds of questions we need to answer and not comparing apples to apples.

Just to be clear, I am not picking up on this particular question. It's a valid question. But I meant to post why some questions are difficult to answer.

 

[sbo6]

I wish there was an easy and simple answer to all these questions. I have done many of these tests myself and have some valuable information I can share, but it's never easy, because the answer is never Yes or No. Or good, better, worse.
Every time I see a question and want to jump on it to provide what I think would be valuable feedback, I realize I have to spend an hour to write an extensive answer that describes the different aspects of the problem. The main reason is that there are too many factors to consider with every change you make.
So without trying to answer this particular question, because I don't have an hour at the moment, I would give you a couple of pointers to think about. I'll do that in the form of questions that we should be asking ourselves and trying to answer when introducing a change like that.
- Will the two NICs be taking power from the motherboard, ot will they be powered by external power supply?
- If powered by the motherboard, wouldn't that pollute the power quality for the other components also taking power from the motherboard? Certainly as the NIC consumes power it generates some noise... how does that impact your overall sound quality?
- If powered separately, is there a way to isolate the ground? Or will the ground be common? If the ground is common, how do you prevent the noise that comes into one of the NIC to go to the other NIC or the rest of your system?
- Also if powered separately, what kind of power supply are you using? And how does that interact with the rest of your system?
- Could DC cables act like antennas? If yes, how adding another DC cable to power the additional NIC affects your system?
- Do Ethernet cables act like antennas? If they do, what happens with your system when you connect two Ethernet cables to your computer instead of one?
- Are you using the same NIC model or different? If different how does the second driver impact your computer?
- How is your OS impacted by having to deal with two NICs and how does that impact the sound quality?
- What is the overall impact on your computer sound quality by changing it from one NIC to two?
- Are you preventing some noise coming from your network by using two NICs? Or are you injecting twice more? Or somewhere in between?

All these questions are just touching the surface of what is happening at a micro level. I can easily dive deeper and provide 10 times more. And that would be all from a technical aspect not even taking into account personal preferences, musical taste, and your local environment. This is why some of those generic questions are difficult to answer. There is too much to consider, and it's very difficult to predict what the overall result would be.

On the other hand, there are things that are easy to recommend. For example, most home networks have a modem, router, firewall, WiFi access point, and a switch. ISPs often give you one device that can perform all 5 functions. You can think of this device as a surround receiver in HiFi. Those receivers are all-in-one devices too. They have a power amplifier, line stage, DAC, phono stage in some cases, etc. Chances are that replacing a cheap surround receiver with a good quality mono blocks, line stage, DAC, phono stage, etc would give you good results.
Well, the same way replacing the all-in-one ISP device with a good modem, a good router (i.e. Ubiquiti EdgeRouter), a good switch, etc. would give you better results for streaming. That's what makes this an easy recommendation. And when you do that, another easy recommendation would be to replace the cheap wall wart on each of those devices with a good power supply. And when you do that, another easy recommendation would be to replace the cheap $1 ethernet cables with better ones. But pay attention to what we are doing here - we are not really changing the design much. We are sticking to the design we have and upgrading the components. This eliminates the almost unlimited number of variables. As soon as you make a design change, we are back at the hundreds of questions we need to answer and not comparing apples to apples.

Just to be clear, I am not picking up on this particular question. It's a valid question. But I meant to post why some questions are difficult to answer.

Agreed there are many variables (your questions essentially apply to switch scenarios also, power, cables, proximity, etc.). My question (if I was not clear) was more specific to why it seems most people do not implement a direct connect via ethernet omitting the switch in between and whether this was due to either a) not having experimented without a switch or b) after experimenting it proved to be sonically worse with their current setup. Details forthcoming could answer the variables articulated in your questions.

 

[Rhapsody]

Agreed there are many variables (your questions essentially apply to switch scenarios also, power, cables, proximity, etc.). My question (if I was not clear) was more specific to why it seems most people do not implement a direct connect via ethernet omitting the switch in between and whether this was due to either a) not having experimented without a switch or b) after experimenting it proved to be sonically worse with their current setup. Details forthcoming could answer the variables articulated in your questions.

Sb06,

I'm not sure if I understand your question, it seems too easy to answer. I'm pretty sure that everyone that has inserted a switch into their chain, first was connecting directly from their router to their server. They listened and understood that sound.

Then they inserted a switch and noticeably experienced a not so subtle uptick in sonic performance. Insert a good LPS and notice even more of an uptick in performance.

Again, I may not be understanding what you are saying. Please let me know if I don't understand your question.

Bob

 

[nenon]

Agreed there are many variables (your questions essentially apply to switch scenarios also, power, cables, proximity, etc.). My question (if I was not clear) was more specific to why it seems most people do not implement a direct connect via ethernet omitting the switch in between and whether this was due to either a) not having experimented without a switch or b) after experimenting it proved to be sonically worse with their current setup. Details forthcoming could answer the variables articulated in your questions.

In the context of a server/streamer solution it does make sense and gives very good results in some cases. I think you mentioned utraRendu - yes, when I was using one, setting a bridge on my JCAT Net Femto gave good results. But I gave up those two box solutions a long time ago...
But absolutely - we have done a lot of experiments. Check out the novel thread on AS - there is tons of info and feedback about our bridging experiments.

 

[sbo6]

In the context of a server/streamer solution it does make sense and gives very good results in some cases. I think you mentioned utraRendu - yes, when I was using one, setting a bridge on my JCAT Net Femto gave good results. But I gave up those two box solutions a long time ago...
But absolutely - we have done a lot of experiments. Check out the novel thread on AS - there is tons of info and feedback about our bridging experiments.

Yep, I read some of it, super long thread and much of it meandering. But thank you will revisit it.

 

[sbo6]

Sb06,

I'm not sure if I understand your question, it seems too easy to answer. I'm pretty sure that everyone that has inserted a switch into their chain, first was connecting directly from their router to their server. They listened and understood that sound.

Then they inserted a switch and noticeably experienced a not so subtle uptick in sonic performance. Insert a good LPS and notice even more of an uptick in performance.

Again, I may not be understanding what you are saying. Please let me know if I don't understand your question.

Bob

Sorry asking about comparing a config with a switch typically connected to a music PC and a renderer versus without the switch - direct connect ethernet form the Music PC to the renderer.

 

[nenon]

Sorry asking about comparing a config with a switch typically connected to a music PC and a renderer versus without the switch - direct connect ethernet form the Music PC to the renderer.

It depends on a lot of factors. I will give you two scenarios, and hopefully that gives you an idea.

Scenario #1
Your server is just an off the shell desktop with a crappy switching mode power supply and a lot of noisy components. And that desktop has two onboard NICs. You also have an etherREGEN switch with a really good LPS.
Chances are that connecting the server to the A-side of the etherREGEN and your streamer to the B-side of the etherREGEN would give you much better result than connecting the streamer directly to the server.

Scenario #2
You have a really high quality server with really clean power (i.e. top linear power supplies). Think Innuos Statement level, or PinkFaun, or Taiko Extreme or something similar. And let's say you are using the built-in switch on the all-in-one cable gateway your ISP has provided to you. Chances are you would never ever even for a minute want to listen to your ultraRendu again. But let's assume for some crazy reason you still want to listen to the ultraRendu... Then chances are that connecting the ultraRendu to your server (and doing what is called network bridging) would give you much better results than connecting the ultraRendu to the switch. The logic is that the signal from your high quality server is cleaner and reclocked with a better clock.

Those are two extremes and you can be anywhere in between. That's why such questions are difficult to answer. Experiment with what you have. If you have a JCAT Net Femto card (seems like you do) but no decent switches, bridging might be a good way to go.

 

[sbo6]

It depends on a lot of factors. I will give you two scenarios, and hopefully that gives you an idea.

Scenario #1
Your server is just an off the shell desktop with a crappy switching mode power supply and a lot of noisy components. And that desktop has two onboard NICs. You also have an etherREGEN switch with a really good LPS.
Chances are that connecting the server to the A-side of the etherREGEN and your streamer to the B-side of the etherREGEN would give you much better result than connecting the streamer directly to the server.

Scenario #2
You have a really high quality server with really clean power (i.e. top linear power supplies). Think Innuos Statement level, or PinkFaun, or Taiko Extreme or something similar. And let's say you are using the built-in switch on the all-in-one cable gateway your ISP has provided to you. Chances are you would never ever even for a minute want to listen to your ultraRendu again. But let's assume for some crazy reason you still want to listen to the ultraRendu... Then chances are that connecting the ultraRendu to your server (and doing what is called network bridging) would give you much better results than connecting the ultraRendu to the switch. The logic is that the signal from your high quality server is cleaner and reclocked with a better clock.

Those are two extremes and you can be anywhere in between. That's why such questions are difficult to answer. Experiment with what you have. If you have a JCAT Net Femto card (seems like you do) but no decent switches, bridging might be a good way to go.

Yes, I'm already bridging with a JCAT Femto to a SOTM Ultra neo. Let me try one last time and apologies for any derailing of the thread.

Scenario: Assume using a non - SOTA server like Innuos Statement, Taiko extreme (which most folks don't own) but a DIY PC with linear PSUs and a Pink Faun or JCAT nic with a LPSU bridged to a SOTM Neo or Sonore Ultrarendu (either with a LPSU) versus the exact same hardware setup non - bridged and instead connected to a <insert audiophile switch here>. Would any switch sound better than direct connect with both devices implementing high quality LPSUs? I'm guessing the answer as folks have said above is - devil is in the details of components and config.

 

[Rhapsody]

I'm guessing the answer as folks have said above is - devil is in the details of components and config.

I guess that your guess is correct

 

[onsionsi]

@Rhapsody i came across to check with you if you figure out any additional enhancement after burning in the M12 Gold second Switch or the difference is subtle.

 

[Rhapsody]

@Rhapsody i came across to check with you if you figure out any additional enhancement after burning in the M12 Gold second Switch or the difference is subtle.

God only knows I listen for 12 hours a day. I'm lost in the forest. My sound is as good as I have ever experienced with systems that I have owned. Since the 50's.....seriously.

It was really good when I inserted the switches. I can't really tell if it has been getting better. My issue is I flip from one set of speakers (10+ sets available) and many more combinations of electronics, coming up with new variations daily.

With the enhancement of the network AND the Extreme, I don't think the dac is as critical to how my systems are performing, but I don't think you could take away EITHER the Extreme or my current network, which is still a work in progress and get the "firing on ALL cylinders" sonics that I am experiencing.

For example I just turned on a system, it's Friday night, going to find something special to listen to. As the system goes on with the electronics needing an hour to warm up I have RadioSwissClassic playing/MP3 quality to do the warm up. It sounds SO GOOD to me that I know for the next 4 or 5 hours I will not switch from listening to the MP3 Radio Station.

I know that I can just hit an icon and get 5X the sonic quality, but it doesn't matter. I'm lost in the music. I do love Radio SwissClassic.

THIS IS WHAT THE NETWORK AND THE EXTREME HAVE DELIVERED FOR ME.

 

[onsionsi]

God only knows I listen for 12 hours a day. I'm lost in the forest. My sound is as good as I have ever experienced with systems that I have owned. Since the 50's.....seriously.

It was really good when I inserted the switches. I can't really tell if it has been getting better. My issue is I flip from one set of speakers (10+ sets available) and many more combinations of electronics, coming up with new variations daily.

With the enhancement of the network AND the Extreme, I don't think the dac is as critical to how my systems are performing, but I don't think you could take away EITHER the Extreme or my current network, which is still a work in progress and get the "firing on ALL cylinders" sonics that I am experiencing.

For example I just turned on a system, it's Friday night, going to find something special to listen to. As the system goes on with the electronics needing an hour to warm up I have RadioSwissClassic playing/MP3 quality to do the warm up. It sounds SO GOOD to me that I know for the next 4 or 5 hours I will not switch from listening to the MP3 Radio Station.

I know that I can just hit an icon and get 5X the sonic quality, but it doesn't matter. I'm lost in the music. I do love Radio SwissClassic.

THIS IS WHAT THE NETWORK AND THE EXTREME HAVE DELIVERED FOR ME.

Really appreciate your feedback and what I know from you the second Switch became a mandatory as same as the first switch.

 

[musicfirst1]

So guys, through all these dizying variations on a theme, I've decided to purchase an Edgerouter and Buffalo switch as the core of my connection to the internet. Do fiber optic cables make a difference on one's network, either in terms of length or brands?
So for me for now its ISP modem>Edgerouter>[copper or fiber optic cable]>Buffalo switch>server

 

[Rhapsody]

Really appreciate your feedback and what I know from you the second Switch became a mandatory as same as the first switch.

All I know is what happened in my set up. Each time I added even the Edge, prior to the two JCAT switches, it was like turning up the gain control. The most noticeable overall gain boost ( to me gain boost is a REALLY good thing) was adding the LPS to the one JCAT and the Edge router.

I would have taken something out of the chain if I did not think it made an instantly noticeable difference. I'm talking 5 seconds and it's like "wow".

YMMV. YMMV. YMMV.

 

[Ultrafast69]

I substantially increased sound quality by relocating the modem on the other side of the wall from the music room, it just required 45’ of coaxial.

I do run separates; modem, router and grounded etherREGEN switch all connected with Audioquest Diamond to Gigafoil v4 to Aurender.

All Keces P8 linear power supplies, Oyaide cabled on DC and Audioquest Dragon on AC. Mechanical isolation on Adona platforms.

ADDPWR gives it a boost.

Results are so good, streaming is very close to files, and I bet most can’t tell the difference.

 

[wil]

God only knows I listen for 12 hours a day. I'm lost in the forest. My sound is as good as I have ever experienced with systems that I have owned. Since the 50's.....seriously.

It was really good when I inserted the switches. I can't really tell if it has been getting better. My issue is I flip from one set of speakers (10+ sets available) and many more combinations of electronics, coming up with new variations daily.

With the enhancement of the network AND the Extreme, I don't think the dac is as critical to how my systems are performing, but I don't think you could take away EITHER the Extreme or my current network, which is still a work in progress and get the "firing on ALL cylinders" sonics that I am experiencing.

For example I just turned on a system, it's Friday night, going to find something special to listen to. As the system goes on with the electronics needing an hour to warm up I have RadioSwissClassic playing/MP3 quality to do the warm up. It sounds SO GOOD to me that I know for the next 4 or 5 hours I will not switch from listening to the MP3 Radio Station.

I know that I can just hit an icon and get 5X the sonic quality, but it doesn't matter. I'm lost in the music. I do love Radio SwissClassic.

THIS IS WHAT THE NETWORK AND THE EXTREME HAVE DELIVERED FOR ME.

Interesting. It sounds like your Extreme/Network improvements are raising the tide so that all formats (including MP3) are rising with it.

Have you noticed that poorly recorded/mastered sources that you couldn't tolerate before have lifted to a more tolerable level? I can imagine they might improve a little, but are essentially a lost
cause!

 

[Rhapsody]

Interesting. It sounds like your Extreme/Network improvements are rising the tide so that all formats (including MP3) are rising with it.

Have you noticed that poorly recorded/mastered sources that you couldn't tolerate before have risen to a more tolerable level? I can imagine they might improve a little, but are essentially a lost
cause!

Hi Wil,

I always test out RadioSwissClassic with any updates. At one time it was just a background not really good for listening. Now I get a 3D holographic soundstage with it and it sounds glorious, depending on the track they are playing.

On poor original recordings whether it is with RSC/MP3 OR Qobuz/Tidal I have not found that all of these latest enhancements have made the poor recordings more listenable.

The MP3 format when they play good tracks, which I find is most of the time, really has been enhanced.

 

[hifimckinney]

As a second entry point to my networked system, I use https://www.fullcompass.com/prod/53...dante-enabled-hdmi-audio-de-embedder-embedder. This extracts audio and converts it to Dante format and on the other end I have a Dante to AES converter feeding my DAC. I would not say that this setup is 100% switch/fiber agnostic but Dante brings to much much closer to real stuff. I would encourage folks to try this out and see how other methods of transportation robs music of its soul.

 

[Blackmorec]

I agree completely with Rhapsody’s reply and
would like to give feedback on my own experience.

Let’s start with Radio Swiss Classic. This is MP3 based 128kbps internet radio.....typically something you’d use to warm up a system without serious listening. But not anymore. The music they transmit are careful selections of joyously uplifting pieces by composers I’m often not familiar with....things like oboe, bassoon or flute concertoes that i wouldn’t normally gravitate to. Beautifully performed by master musicians the recordings are generally extremely well Made. Like Rhapsody I started using the station as a warm up and found that it became increasingly engaging, getting to the point it becomes a real act of will to switch off, even though the higher resolution formats do sound better. Frankly, Swiss Radio Classics may be the closest I’ve ever come to the true sound of a concert hall, where instruments are beautifully located in space without being artificially highlighted or spotlighted....no doubt the output of a few, highly skilled recording engineers.
But these are not bad recordings.....they are great recordings transmitted at a low resolution, that when processed properly sound magnificent.

On the other hand, poor recordings are still poor recordings.....for example when all musicians share the same crowded centre stage, all sounds emanating from the exact same spot, which in real life is of course physically impossible....recordings without bloom or ambience, cold, sterile and without warmth and beauty. These recordings are fairly rare, less than 5% probably but no amount of system or network improvement can help such recordings, other than to let you understand why they’re so unpleasant. But where the true gems lie are in really good recordings that’s you may not have fully appreciated that suddenly bloom to generate a sense of amazement and wonder....where the true talent of the artist is revealed in subtle details of timing, rhythmic interplay, note shape and micro-dynamics. When that happens you know you’ve made a truly fundamental upgrade of your system’s performance.

 

[musicfirst1]

Gentlemen

Can you tell us what app you are streaming Radio Swiss Classic with?

Thanks