Networking Logic - Intro

tuga

. . .
Wammer
Aug 17, 2007
14,341
7,001
173
Oxen's ford, UK
AKA
Ricardo
HiFi Trade?
  1. No
If you look at the history of analogue, it’s a steady progress of greater refinement, additional precision and increased complexity.
If you look at digital, it’s going in the same direction. If you believe that digital can’t be improved beyond a bit-perfect file, that’s fine. But that belief will also hold back discoveries of just how damned good digital and specifically streamed digital can become by removing noise, timing errors and vibration from the entire chain.

I agree that the file is only bit perfect because it is data, and once it is 'read' the digital stream becomes 'frail'.
This was taken from "There's no Such Thing as Digital: a Conversation With Charles Hansen, Gordon Rankin, and Steve Silberman" by Michael Lavorgna, AudioStream June 24, 2013:

Michael Lavorgna: It's common for people to envision and represent a digital signal as a series of 1s and 0s. As such, there's really no room for error, at least according to this binary theory. Is a digital signal simply a series of 1s and 0s?

Charlie Hansen: Unfortunately not. The "1"s and "0"s are just abstractions that are easy to think about. But in the real world, something real needs to represent those two abstract states. In modern digital electronics, we have almost universally chosen a voltage above a specific level (that varies from one "family" of electronic parts to another) to represent a "1" and a voltage below a different specific level (that again can vary) to represent a "0".

In the real world, those two voltages are not the same, so there is a "grey" zone between the "black" of the "0" and the "white" of the "1". Also, it takes time for the signal to change levels, and the time required to do so can depend on dozens (or even thousands) of other external factors.

All of the problems with digital are analog problems.



I disagree with your vinyl player analogy only because the moajority of people were/are addressing real problems whilst in digital audio a few people are addressing imagined problems (mostly for lack of understanding).
Driving a Lamborghini to the supermarket will make as much difference to the time it takes as using a fancy clock in a network switch will improve the data transfer.
 

tuga

. . .
Wammer
Aug 17, 2007
14,341
7,001
173
Oxen's ford, UK
AKA
Ricardo
HiFi Trade?
  1. No
I’m not the only person doing this sort of refinement. Taiko, Innuos and several other companies are making extensive measurements and conducting listening trials to figure out exactly what’s going on and come up with better sound quality. Their research is proprietary but if you talk the Emile Bok or Nuno Vitorino they will share some of their findings.
Which we should believe at face value just because they are 'experts', ot their products are expensive so they must work?
Personally, if the claims are 'bold' and defy logic and current knowledge I wan't to see hard evidence not just marketing blabber.
 
  • Like
Reactions: Strider and Tony_J

JustinTime

Wammer Plus
Wammer Plus
Jan 18, 2023
699
1,397
148
HiFi Trade?
  1. No
If you look at digital, it’s going in the same direction. If you believe that digital can’t be improved beyond a bit-perfect file, that’s fine. But that belief will also hold back discoveries of just how damned good digital and specifically streamed digital can become by removing noise, timing errors and vibration from the entire chain.
I agree, digital can be improved in many ways. Including and after the streamer, it can be improved in terms of (let's call RFI "analogue") noise and it can be improved in terms of (digital) jitter. Up to the streamer it can only be improved in terms of analogue (RFI) noise; the digital data can't be improved. To steal a phrase from your thread title, this is simply "networking logic".
 
  • Like
Reactions: tuga

Cable Monkey

Moderator
Staff member
May 16, 2006
9,069
2,089
158
Birmingham, UK
AKA
Henry
HiFi Trade?
  1. No
Driving a Lamborghini to the supermarket will make as much difference to the time it takes as using a fancy clock in a network switch will improve the data transfer.
The purpose of improving those clocks is not and never has been for the reasons of improving data transfer. And people don't drive Lamborghini's to save time. They do so because they can afford to. So all things considered, not a great analogy.
 

tuga

. . .
Wammer
Aug 17, 2007
14,341
7,001
173
Oxen's ford, UK
AKA
Ricardo
HiFi Trade?
  1. No
The purpose of improving those clocks is not and never has been for the reasons of improving data transfer. And people don't drive Lamborghini's to save time. They do so because they can afford to. So all things considered, not a great analogy.
People buy a £1k switch because they can.
Is there evidence that a fancy clock in a switch will a) improve data transfer and/or b) reduce noise in the network?
 

JustinTime

Wammer Plus
Wammer Plus
Jan 18, 2023
699
1,397
148
HiFi Trade?
  1. No
People buy a £1k switch because they can.
Is there evidence that a fancy clock in a switch will a) improve data transfer and/or b) reduce noise in the network?
There may not be what would constitute evidence for you… I think @Cable Monkey ’s point is that a “fancy” ethernet clock may indeed positively impact sound quality if its fanciness includes being a low noise component. I believe we three agree that any fanciness in relation to its accuracy is irrelevant to sound quality.
 
Last edited:

bencat

Amplifier Destroyer
Wammer Plus
Feb 6, 2010
10,309
8,115
208
Liverpool
AKA
Andrew
HiFi Trade?
  1. No
But clocks do not need to be fancy or that expensive to make a difference that can be heard . Problem for many is that current digital systems do not make it easy to insert a better external or word clock in the chain . So many have never heard what a change can do and then state that it can make no difference . I am luck I have the inputs needed to change this element and hear the difference . I could spend lots more to get the best possible external clock but I suspect the difference will be less than when the clock I have was added . I could well be wrong and the difference is huge but I do not have the money or inclination to try out one .
 

tuga

. . .
Wammer
Aug 17, 2007
14,341
7,001
173
Oxen's ford, UK
AKA
Ricardo
HiFi Trade?
  1. No
But clocks do not need to be fancy or that expensive to make a difference that can be heard . Problem for many is that current digital systems do not make it easy to insert a better external or word clock in the chain . So many have never heard what a change can do and then state that it can make no difference . I am luck I have the inputs needed to change this element and hear the difference . I could spend lots more to get the best possible external clock but I suspect the difference will be less than when the clock I have was added . I could well be wrong and the difference is huge but I do not have the money or inclination to try out one .
I am not questioning the importance of clocks in digital audio, only in computer networking.
Although the right place for reclocking is by the D/A chip
 
  • Like
Reactions: Nestor Turton

JustinTime

Wammer Plus
Wammer Plus
Jan 18, 2023
699
1,397
148
HiFi Trade?
  1. No
But clocks do not need to be fancy or that expensive to make a difference that can be heard . Problem for many is that current digital systems do not make it easy to insert a better external or word clock in the chain . So many have never heard what a change can do and then state that it can make no difference . I am luck I have the inputs needed to change this element and hear the difference . I could spend lots more to get the best possible external clock but I suspect the difference will be less than when the clock I have was added . I could well be wrong and the difference is huge but I do not have the money or inclination to try out one .
You’re talking about different clocks. This is about ethernet clocks.
 

Cable Monkey

Moderator
Staff member
May 16, 2006
9,069
2,089
158
Birmingham, UK
AKA
Henry
HiFi Trade?
  1. No
People buy a £1k switch because they can.
Is there evidence that a fancy clock in a switch will a) improve data transfer and/or b) reduce noise in the network?
There is evidence that a higher quality clock reduces noise. These are in common use, usually in higher order equipment and it is this that has been trickled down to these HiFi switches.
There is no evidence data transfer is improved because it isn't. And as stated repeatedly before, no one is suggesting that data transfer is what is being improved in that process. But I note it doesn't stop the question being repeated.
 

JustinTime

Wammer Plus
Wammer Plus
Jan 18, 2023
699
1,397
148
HiFi Trade?
  1. No
Am a sucker for a happy ending.....:)
Will stick my head above the parapet.
Have been trying some cable between my Dac & Pre the last couple of weeks.The improvement to my ears is very impressive. I only stream these days, but fortunately am open minded & am pleased to be able to eat my own words. Reg cable = I sit somewhere in the middle.
NVA TIS cable apparently. Impressive lift in s/q. I don't really want to remove it from my system. Icing on the cake.
Ooooer missus. 👀
Interesting report - love the open-mindedness - but in the wrong thread I’m afraid! This is all about ethernet stuff.
 
  • Haha
Reactions: tuga and garn63

JustinTime

Wammer Plus
Wammer Plus
Jan 18, 2023
699
1,397
148
HiFi Trade?
  1. No
There is evidence that a higher quality clock reduces noise. These are in common use, usually in higher order equipment and it is this that has been trickled down to these HiFi switches.
Intrigued by common use: common use where?

Intrigued by "higher order equipment": examples being?

Thanks
 
  • Like
Reactions: tuga

Blackmorec

Newbie
Wammer
Feb 6, 2019
90
65
23
Intrigued by common use: common use where?

Intrigued by "higher order equipment": examples being?

Thanks
Any application requiring a clean timing source
Any app requiring accurate frequency control
radio transmitters
Cellular base stations
Military comms equipment
Precise frequency measurements
High accuracy test and measurement equipment

The clocks are standard off-the-shelf items offered by a variety of manufacturers so audio manufacturers have plenty of choice when it comes to selecting clocks

I do get the point that a network shouldn’t need super accurate clocks in order to function correctly. Its only when that network stream is converted into an analog signal and used to drive an audio system that we find that certain things that ought to have zero affect on the data have a major affect on the music produced. We know that affect isn’t caused by the data itself, because its bit perfect and any change or deviation would be a move away from accuracy. So it comes down to the fabric of the bit stream i.e how ‘perfectly‘ each bit is formed, how many retransmissions are required, how much error correction is required, how much noise is included, how much noise is generated, how much noise is picked up, how much latency is introduced, the speed and shape of the voltage transitions between 0v and 2v, how much power supply noise is generated, how much jitter and therefore phase noise is included etc. A computer network is built to ignore all that ‘stuff’ and just look at the actual voltage transitions. A DAC is designed to ignore all that stuff, but unfortunately that stuff leaves an imprint on and changes the final sound signature. It has an impact on the creation of the analog voltage from the bit stream.
If networking was indeed as robust in the audio domain as it is in the IT domain, then nothing should change how the resulting music sounds as long as we start with a bit perfect stream. But we know that‘s not the case and that many things can and do change how that music sounds. All I’m saying is that when you remove any and all those variables and stick as closely as possible to the ideal spec for noise, timing, vibration levels etc. the effect on the final sound is nothing short of spectacular.
I took network optimisation to an obsessive level because with each step, I got very clear rewards and those individual rewards compounded into a musical presentation that was way beyond my wildest expectation for streamed audio.
Companies like Taiko and Innuos listen to music, identify shortcomings, postulate where those shortcomings may be coming from, make measurements to confirm or deny their postulations then design circuits and products to address the shortcomings. From that work we know that power supplies, RAM, network traffic, EMI, clocks, storage devices, wiring, vibration control, DC cabling, network cabling, processors, mains supplies, connectors, switches, bridges, routers, modems etc etc. all affect how the final music sounds. Clearly these things don't effect the data itself, because its already bit perfect but they do affect how the bit perfect data sounds when its finally converted to music.
 

Cable Monkey

Moderator
Staff member
May 16, 2006
9,069
2,089
158
Birmingham, UK
AKA
Henry
HiFi Trade?
  1. No
Intrigued by common use: common use where?

Intrigued by "higher order equipment": examples being?

Thanks
We operate at the lowest levels of Ethernet transmission. 10/100/1000 and at a stretch 2.5 and 10gig.
This is carried over ISP networks with backbone switches capable of carrying massive amounts of traffic.
Those switches and their design attributes are what are selectively trickled down to lower order gear.
The aim is to harness lower noise performance but marketing pimps more readily apparent accuracy numbers.
 
  • Like
Reactions: JustinTime

JustinTime

Wammer Plus
Wammer Plus
Jan 18, 2023
699
1,397
148
HiFi Trade?
  1. No
We operate at the lowest levels of Ethernet transmission. 10/100/1000 and at a stretch 2.5 and 10gig.
This is carried over ISP networks with backbone switches capable of carrying massive amounts of traffic.
Those switches and their design attributes are what are selectively trickled down to lower order gear.
The aim is to harness lower noise performance but marketing pimps more readily apparent accuracy numbers.
Ah so when you talk about higher order equipment you’re talking about transmission speeds. Gotcha.

Do manufacturers have to test a whole load of clocks themselves then, or do the clock manufacturers publish noise measurements? I was struggling to think of any sectors other than audio where clock noise would be more important than accuracy, but I guess maybe medical?
 

JustinTime

Wammer Plus
Wammer Plus
Jan 18, 2023
699
1,397
148
HiFi Trade?
  1. No
Any application requiring a clean timing source
Any app requiring accurate frequency control
radio transmitters
Cellular base stations
Military comms equipment
Precise frequency measurements
High accuracy test and measurement equipment

The clocks are standard off-the-shelf items offered by a variety of manufacturers so audio manufacturers have plenty of choice when it comes to selecting clocks

I do get the point that a network shouldn’t need super accurate clocks in order to function correctly. Its only when that network stream is converted into an analog signal and used to drive an audio system that we find that certain things that ought to have zero affect on the data have a major affect on the music produced. We know that affect isn’t caused by the data itself, because its bit perfect and any change or deviation would be a move away from accuracy. So it comes down to the fabric of the bit stream i.e how ‘perfectly‘ each bit is formed, how many retransmissions are required, how much error correction is required, how much noise is included, how much noise is generated, how much noise is picked up, how much latency is introduced, the speed and shape of the voltage transitions between 0v and 2v, how much power supply noise is generated, how much jitter and therefore phase noise is included etc. A computer network is built to ignore all that ‘stuff’ and just look at the actual voltage transitions. A DAC is designed to ignore all that stuff, but unfortunately that stuff leaves an imprint on and changes the final sound signature. It has an impact on the creation of the analog voltage from the bit stream.
If networking was indeed as robust in the audio domain as it is in the IT domain, then nothing should change how the resulting music sounds as long as we start with a bit perfect stream. But we know that‘s not the case and that many things can and do change how that music sounds. All I’m saying is that when you remove any and all those variables and stick as closely as possible to the ideal spec for noise, timing, vibration levels etc. the effect on the final sound is nothing short of spectacular.
I took network optimisation to an obsessive level because with each step, I got very clear rewards and those individual rewards compounded into a musical presentation that was way beyond my wildest expectation for streamed audio.
Companies like Taiko and Innuos listen to music, identify shortcomings, postulate where those shortcomings may be coming from, make measurements to confirm or deny their postulations then design circuits and products to address the shortcomings. From that work we know that power supplies, RAM, network traffic, EMI, clocks, storage devices, wiring, vibration control, DC cabling, network cabling, processors, mains supplies, connectors, switches, bridges, routers, modems etc etc. all affect how the final music sounds. Clearly these things don't effect the data itself, because its already bit perfect but they do affect how the bit perfect data sounds when its finally converted to music.
I just think it’s important to separate out the above into pre-streamer ethernet domain and post-streamer digital (for want of a better differentiating label) domain, that’s all. I can accept far more of the tweaks you mention affecting digital domain than I can ethernet domain.
 
  • Like
Reactions: tuga

Cable Monkey

Moderator
Staff member
May 16, 2006
9,069
2,089
158
Birmingham, UK
AKA
Henry
HiFi Trade?
  1. No
Ah so when you talk about higher order equipment you’re talking about transmission speeds. Gotcha.

Do manufacturers have to test a whole load of clocks themselves then, or do the clock manufacturers publish noise measurements? I was struggling to think of any sectors other than audio where clock noise would be more important than accuracy, but I guess maybe medical?
The few people I know who are at all original read specifications and know what to look for.
 
  • Like
Reactions: JustinTime

tuga

. . .
Wammer
Aug 17, 2007
14,341
7,001
173
Oxen's ford, UK
AKA
Ricardo
HiFi Trade?
  1. No
Any application requiring a clean timing source
Any app requiring accurate frequency control
radio transmitters
Cellular base stations
Military comms equipment
Precise frequency measurements
High accuracy test and measurement equipment

The clocks are standard off-the-shelf items offered by a variety of manufacturers so audio manufacturers have plenty of choice when it comes to selecting clocks

I do get the point that a network shouldn’t need super accurate clocks in order to function correctly. Its only when that network stream is converted into an analog signal and used to drive an audio system that we find that certain things that ought to have zero affect on the data have a major affect on the music produced. We know that affect isn’t caused by the data itself, because its bit perfect and any change or deviation would be a move away from accuracy. So it comes down to the fabric of the bit stream i.e how ‘perfectly‘ each bit is formed, how many retransmissions are required, how much error correction is required, how much noise is included, how much noise is generated, how much noise is picked up, how much latency is introduced, the speed and shape of the voltage transitions between 0v and 2v, how much power supply noise is generated, how much jitter and therefore phase noise is included etc. A computer network is built to ignore all that ‘stuff’ and just look at the actual voltage transitions. A DAC is designed to ignore all that stuff, but unfortunately that stuff leaves an imprint on and changes the final sound signature. It has an impact on the creation of the analog voltage from the bit stream.
If networking was indeed as robust in the audio domain as it is in the IT domain, then nothing should change how the resulting music sounds as long as we start with a bit perfect stream. But we know that‘s not the case and that many things can and do change how that music sounds. All I’m saying is that when you remove any and all those variables and stick as closely as possible to the ideal spec for noise, timing, vibration levels etc. the effect on the final sound is nothing short of spectacular.
I took network optimisation to an obsessive level because with each step, I got very clear rewards and those individual rewards compounded into a musical presentation that was way beyond my wildest expectation for streamed audio.
Companies like Taiko and Innuos listen to music, identify shortcomings, postulate where those shortcomings may be coming from, make measurements to confirm or deny their postulations then design circuits and products to address the shortcomings. From that work we know that power supplies, RAM, network traffic, EMI, clocks, storage devices, wiring, vibration control, DC cabling, network cabling, processors, mains supplies, connectors, switches, bridges, routers, modems etc etc. all affect how the final music sounds. Clearly these things don't effect the data itself, because its already bit perfect but they do affect how the bit perfect data sounds when its finally converted to music.
Audio data packets are transmitted through the network just like any other data.
Such data transfer does not require a clock better than what is provided by a basic run-of-the-mill router or switch. If it did there be loss of information.

A renderer then converts these data packets into a digital audio data-stream which is sent to the DAC over USB, coaxial, optical or AES.
A higher performance clock in the renderer might be beneficial if AES or S/PDIF is used, particularly with older DACs, because the clock is transmitted embedded in the signal.

A higher performance clock is definitely beneficial in the DAC.
Unfortunately clocks in DACs are not user replaceable so people recur to external clocks in hope that they will improve performance (highly unlikely in most cases).
 
Last edited:

Blackmorec

Newbie
Wammer
Feb 6, 2019
90
65
23
Audio data packets are transmitted through the network just like any other data.
Such data transfer does not require a clock better than what is provided by a basic run-of-the-mill router or switch. If it did there be loss of information.

A renderer then converts these data packets into a digital audio data-stream which is sent to the DAC over USB, coaxial, optical or AES.
A higher performance clock in the renderer might be beneficial if AES or S/PDIF is used, particularly with older DACs, because the clock is transmitted embedded in the signal.

A higher performance clock is definitely beneficial in the DAC.
Unfortunately clocks in DACs are not user replaceable so people recur to external clocks in hope that they will improve performance (highly unlikely in most cases).
Hi tuga,
I don’t disagree at all with what you are saying. The standard clock provided in a run-of-the-mill router or switch is all that is necessary. I started out with exactly that AND the cheap-as-chips £6 SMPS power supplies, and the music played fine. No problems, no drop outs, no skips, no distortion.

But that just takes us back to where we were a few days ago. I’m not saying and have never said that you need better clocks for the audio network to function correctly. In terms of having ‘as much as necessary and as little as possible’ your approach in bang on and will deliver a bit perfect data stream that will play music without fault or flaw. However it will not deliver anywhere near the finest sound quality. Not even close. Refine the network in order to achieve as close to the ideal network specs as possible….ie. No noise of any sort, no jitter/phase noise, no vibration, no cable losses, perfect voltage, minimal traffic etc. and the sound quality i.e the presentation of the music will take a giant leap forward in terms of everything an audiophile values and that without changing the bit logic, just the bit structure…the bit fabric.

Let’s use your car analogy. It doesn’t matter whether you drive a Lamborghini or a VW UP at 30mph, you will arrive where you’re going at exactly the same time. It’s just that, if you are a car fan, you’ll enjoy the trip far more in one than the other. That’s what this is about. Both clocks produce fault free music. It’s just that one version is a lot more enjoyable to listen to than the other. A LOT MORE!
It’s no different than say a turntable. 2 TTs run at exactly the same speed, so a record takes exactly the same time to play on either TT. But on one TT the music sounds a lot quicker, a lot pacier, a lot more rhythmical and has a much better sense of rhythm than the other. That’s exactly what happens when you improve the ethernet network. The improved version becomes a lot more entertaining, fun, and involving.
 

tuga

. . .
Wammer
Aug 17, 2007
14,341
7,001
173
Oxen's ford, UK
AKA
Ricardo
HiFi Trade?
  1. No
However it will not deliver anywhere near the finest sound quality.
There is no evidence that the better clock in a network will improve performance.
And you can’t even test it either unless you compare two switches where the only difference is the clock.
You may accept Innuos word at face value but that doesn’t make it true.
 

Latest posts

Forum statistics

Threads
113,444
Messages
2,451,263
Members
70,783
Latest member
reg66

Latest Articles