[Darren G]

Thank you for the reply

I have particular interest in the in measurable differences of inaudible noise which can pass though the USB cable (if that's possible) and how the design and materials of the USB cable could affect this inaudible noise.

There are some good posts earlier in this thread that emphasize a key point. That cables pick up noise isn't news. It's precisely the motivation behind digital signaling, to transfer the information while filtering the noise. There were also comments made about the wires in the cable, the purpose of each, including why there are two data lines, reverse signals, so that common noise can be rejected (a common and well understood technique).

There were also comments about the slight possibility of a ground loop, and maybe a comment about noise from the +5 volt line, with emphasis that any competent design should isolate.

Still, if any of that concerns you, check out the Schiit Eitr. For < $200 it will isolate all 4 wires from the source (usually a computer), including outputting it's own +5 volts. Even so, Schiit still just recommends using a good old (i.e, meets spec) USB cable in, and likewise a good old (meets spec) coax cable out.

 

[Clive101]

There are some good posts earlier in this thread that emphasize a key point. That cables pick up noise isn't news. It's precisely the motivation behind digital signaling, to transfer the information while filtering the noise. There were also comments made about the wires in the cable, the purpose of each, including why there are two data lines, reverse signals, so that common noise can be rejected (a common and well understood technique).

There were also comments about the slight possibility of a ground loop, and maybe a comment about noise from the +5 volt line, with emphasis that any competent design should isolate.

Still, if any of that concerns you, check out the Schiit Eitr. For < $200 it will isolate all 4 wires from the source (usually a computer), including outputting it's own +5 volts. Even so, Schiit still just recommends using a good old (i.e, meets spec) USB cable in, and likewise a good old (meets spec) coax cable out.

So your saying no noise can pass though a USB cable because it is filtered, OK as that was why digital was invented.

Could possibility inaudible noise have been missed because as it is inaudible no manufacturer bothered to filter for it...?

Sorry about so many questions.

 

[ev13wt]

So your saying no noise can pass though a USB cable because it is filtered, OK as that was why digital was invented.

Could possibility inaudible noise have been missed because as it is inaudible no manufacturer bothered to filter for it...?

Sorry about so many questions.

Imagine a beach. There are the big sets of waves coming in, plus a huuuge amount of tiny waves all around. Lets call the big waves the audio signal, and all the small waves "noise".
You are a wave creator. You send big waves to the beach. One per 5 seconds. The height of the wave symbolizes the loudness of the signal. Wave, 5 sec pause, wave, 5 sec, 5 sec, wave, wave = 1010011

In analog ocean, your friend the amp counts all waves and makes them bigger. It does not care about noise or signal, it simply amplifies the entire scene. Big waves, small waves, cross waves. Its now a storm!

In a digital ocean, your other friend Mr. dac input counts all the waves coming in, but only looks for 2 things: Is the wave bigger than "x", and is there a space of 5 seconds where no wave comes in?
It converts the information back to "real waves" and shows the amp the perfect ocean, smooth everywhere but for the clearly visible and smooth rollers coming in. Amp is happy, listener is happy.
Wave, 5 sec pause, wave, 5 sec, 5 sec, wave, wave = 1010011

If the other waves, that represent the noise, are as big as the factor X, digital transmission will "fall apart". Otherwise, you always get a perfect beach scene. The noise waves are simply disregarded entirely.

So as long as the usb cable transmits the signal waves at a "higher volume" than the noise waves, the signal is perfectly recontructed by the DAC.

Any noise smaller than the peak of a "1" is simply a non issue. It disappears. Long version: https://en.wikipedia.org/wiki/Entropy_(information_theory)

 

[Darren G]

@Clive101 - For the data lines, the receiver's job is to take those two data lines, detect the high/low state of the lines, and what comes out the other side really is just 1 or 0. It's worth reading an article or two on the details, but the high/low signals occur at a frequency, and frequencies above/below those don't trigger the 1 or 0 out decision. What comes out is not music, it's not what you are going to hear either, and you aren't listening to the 1 or 0s. Later, the DAC converts those 1s and 0s to analog, and that's what you hear, so the answer really is the noise is filtered.

Some people still worry about voodoo noise slipping through, but can't explain how that happens either. There is something to concerns of ground loops, but the solution to that is isolation (some other outputs, like AES require isolation as part of the spec) or sometimes re-arranging the wall sockets you plug your gear into. Already commented on the +5V, which some DACs don't use anyway. There there is that the DAC has a case to shield it, but anyway...

If someone is having issues, spending more on a USB cable isn't going to solve it. Something like the Schiit Eater actively can address isolation, but the USB cable, I'm in the save my $ camp and spend it on something else.

 

[Clive101]

Imagine a beach. There are the big sets of waves coming in, plus a huuuge amount of tiny waves all around. Lets call the big waves the audio signal, and all the small waves "noise".
You are a wave creator. You send big waves to the beach. One per 5 seconds. The height of the wave symbolizes the loudness of the signal. Wave, 5 sec pause, wave, 5 sec, 5 sec, wave, wave = 1010011

In analog ocean, your friend the amp counts all waves and makes them bigger. It does not care about noise or signal, it simply amplifies the entire scene. Big waves, small waves, cross waves. Its now a storm!

In a digital ocean, your other friend Mr. dac input counts all the waves coming in, but only looks for 2 things: Is the wave bigger than "x", and is there a space of 5 seconds where no wave comes in?
It converts the information back to "real waves" and shows the amp the perfect ocean, smooth everywhere but for the clearly visible and smooth rollers coming in. Amp is happy, listener is happy.
Wave, 5 sec pause, wave, 5 sec, 5 sec, wave, wave = 1010011

If the other waves, that represent the noise, are as big as the factor X, digital transmission will "fall apart". Otherwise, you always get a perfect beach scene. The noise waves are simply disregarded entirely.

So as long as the usb cable transmits the signal waves at a "higher volume" than the noise waves, the signal is perfectly recontructed by the DAC.

Any noise smaller than the peak of a "1" is simply a non issue. It disappears. Long version: https://en.wikipedia.org/wiki/Entropy_(information_theory)

Thank you, yes I get point, thought I was on holiday for one moment..!

Yes aware of that but..... can the inaudible noise go though the USB cable though Mr Dac and somehow affect the SQ when it gets into Mr Amp...?

 

[Clive101]

Some people still worry about voodoo noise slipping through, but can't explain how that happens either

Yes can that be the cause, the voodoo effect, well could the noise travel though the electrical circuits of the Dac..? Sort of by passing the processing, it is all interconnected ..?

If the answer is no I have finished with my questions, but thank you all for your patience I will continue to stay and see what happens ...

Edit @gregorio has answered this in the next post thank you

 

[gregorio]

[1] So your saying no noise can pass though a USB cable because it is filtered, OK as that was why digital was invented.

Could possibility inaudible noise have been missed because as it is inaudible no manufacturer bothered to filter for it...?

Sorry about so many questions.

1. No, he's saying no noise (except in very exceptional/rare circumstances) can affect the digital data. Noise can and does affect analogue signal circuits, that's why all DACs have a noise floor.
2. No, manufacturers publish noise specifications for their DACs, signal to noise ratio and dynamic range specs for example. Both of these specs are beyond the limits of audibility and to publish those specs they obviously have to measure them. So inaudible noise is routinely measured and reduced even further, to provide impressive looking specs. Noise floors of DACs today, even relatively cheap ones, are commonly at least 30 times lower than the threshold of audibility and at least 100 times below the noise floor on the recordings themselves.

G

 

[ev13wt]

Thank you, yes I get point, thought I was on holiday for one moment..!

Yes aware of that but..... can the inaudible noise go though the USB cable though Mr Dac and somehow affect the SQ when it gets into Mr Amp...?

You haven't defined "inaudible noise". Frequency spectrum? Amplitude? Light is inaudible noise, but it doesn't affect the amp.



Yes, if it is present on the recording and the audio is not saved in a lossy compresson of sorts.

Otherwise, if there are no ground problems, no. (This to your "it is all connected" reference.)

 

[gregorio]

[1] Yes can that be the cause, the voodoo effect, [2] well could the noise travel though the electrical circuits of the Dac..? Sort of by passing the processing, it is all interconnected ..?

1. The voodoo effect is imaginary and we have not yet learned how to pass imagination down a cable, even an audiophile cable!
2. Noise can and does pass through the electrical, analogue circuits of the DAC, as explained previously. Yes, it is all connected and it's quite complex, which is why circuit designers go to university and study it, but it's all problems which were sorted out years or decades ago and any vaguely competent designer should keep the noise way below audibility.

G

 

[Clive101]

1. The voodoo effect is imaginary and we have not yet learned how to pass imagination down a cable, even an audiophile cable!
2. Noise can and does pass through the electrical, analogue circuits of the DAC, as explained previously. Yes, it is all connected and it's quite complex, which is why circuit designers go to university and study it, but it's all problems which were sorted out years or decades ago and any vaguely competent designer should keep the noise way below audibility.

G

Thank you.

 

[Arpiben]

Thank you for the reply

I have particular interest in the in measurable differences of inaudible noise which can pass though the USB cable (if that's possible) and how the design and materials of the USB cable could affect this inaudible noise.

Let me try this way starting with some basics.

A transmission line, a pair of wire in our case is nothing more than:

dx: unit length

When dealing with USB certified cables there are constraints for R/L/G and C. You may change material and it will slightly change thoses values but they must be kept within specifications.
A transmission line is greatly dependant on the load or characteristic impedance at entry Port A and output Port B. A will be the load from player and B the input load from DAC

.

USB also provides specifications for transmitters (Port A ) and receivers (Port B)

In order to cancel electromagnetic interferences from external sources data is transmitted in differential mode and very often through a twisted cable pair.
The USB receiver will eliminate the EMI by Common Mode Rejection ( CMR cf @pinnahertz posts)
USB20 specifies twisted pairs for Full Speed and High Speed Rates.

Twisted pairs: https://en.wikipedia.org/wiki/Twisted_pair

There are different ways to twist the pairs for increasing the noise rejection. Data pairs or even power pairs may be shielded for improving not only EMI but also for reducing their own radiation. Shielding provides an electrically conductive barrier to attenuate electromagnetic waves external to the shield, and provides a conduction path by which induced currents can be circulated and returned to the source, via ground reference connection.
Then again some shields/foils are more efficient when dealing with Low Frequencies some with High Frequency.

Anyhow in normal environment, a compliant USB cable is more than enough for dealing with those external interferences.

Things can become more complicated when facing unproper grounding connections or management at DAC or Source level.
We may add more complexity by adding that external interferences are not only entering DAC via its USB cable/receptable or plug parts. In fact all Ins/Outs may act as entry point even your own body when listenning through headphones or touching metallic parts of your listenning chain...
We may still add more complexity by adding that interferences may be internaly generated inside player or DAC and then cable may carry those noises through ground/power wires.

That is the reason why, we keep saying, that it is not to USB cable to deal with design issues from players or DACs.

Adding inductance L to USB cables via ferrite chokes or beads in order to reject HF will in most cases reduce the SNR (Signal to Noise Ratio) of the data signal at the risk of having data integrity issues. Hereunder eye patterns examples ( left correct / right not compliant)

With 'specialized' audiophie USB cables, we have not the chance to have access to any measurement: eye patterns at various audio PCM rates for example, EMI rejection,AC rejection, etc....

Hope it helps

 

[ev13wt]

Thank you very much for the time!

I feel like I understood a lot now, but a new area has opened up with so much more complexity. I love a challenge. Eye diagram. Seen it before, no idea what it actually shows. More googling for papers is in order!

 

[Arpiben]

Thank you very much for the time!

I feel like I understood a lot now, but a new area has opened up with so much more complexity. I love a challenge. Eye diagram. Seen it before, no idea what it actually shows. More googling for papers is in order!

Try this application note:
https://dl.cdn-anritsu.com/en-us/te...cation-Notes/Application-Note/11410-00533.pdf

 

[bigshot]

The important thing that always seems to get overlooked is the just detectable difference threshold for human ears. It's great to talk about noise in cables in theory, but it doesn't mean anything if you can't hear it. And the kinds of noise being discussed here is so far below the threshold of perception it isn't even worth talking about. As long as a cable is designed and manufactured without any glaring errors, it's as good for the purposes of transferring signals representing music as any other cable.

Audiophiles waste most of their energy worrying about things that aren't even an issue. There are much more important things to attend to that actually will improve the sound of their system. And perhaps even more importantly, there are much more elegant ways of arriving at the same place that don't require a bunch of expensive equipment and complete lack of ergonomics.

 

[Darren G]

Two key facts are so well explained above -

1.) The whole point of digital data transfer is to transmit information reliably, while filtering out analog noise on the wire.
2.) ALL electronic gear has a noise floor, and that noise floor is never 0. This is nothing new, it's well understood, and the goal of most manufacturers is to lower that floor well below the level of what us humans can hear (that is what the word 'inaudible' means).

--

If anyone is still having nervousa about USB, there is the Schiit Eitr, though even then, those who have it write about a 1-2% difference (i.e., minuscule), and not in blind, or double blind testing, 1-2% subjective impressions. Schiit says up front, USB Nervousa solved (in their usual joking manner). It checks these nervousa boxes -

USB processing moved outside of the DAC? Check.
Ground isolation? Check (Coax out to your DAC is transformer coupled).
+5 volt isolation? Check (it doesn't even connect the source +5, and uses it's own linear power-supply).
Do you fret over jitter? Check (uses 44.1K and 48K on-board crystals to emit a well-clocked SPDIF stream out over coax).

But again, Schiit is good about making no claims you will hear any difference.