pi3 sd card vs Pi4 wifi or ethernet sq

[Gate45]

I would like to contribute to this interesting discussion.
I have a house with four rooms wired with CAT6 ethernet cable and the entrance to the house has the collection of the four lines where I have the modem coming out with a 4 port switch. In two rooms I have two "Tp Link" switches to connect various devices including the Raspberry PI 3+.
To store my music I use a QNAP NAS that I kept on the Switch in the room where the Raspberry PI was not, because it makes noise when the NAS reads the disks and so I chose to move it, but so the data path gives the The NAS on RPi was long and passed back and forth between three Switches, but I figured, being digital, there would be no problem.
I was wrong, I hadn't considered the noise that is added to the data by switching between 3 sets of switches.
So I decided to put the NAS on the switch where there is also the RPI. The music is much improved, much cleaner, more dynamic, etc. So I thought that with every switch from an electronic device, the situation got worse.
The noise where it comes from, largely from the power supplies and seeing that the flying power supplies of the switches are not of great quality, I replaced the 9V power supply with a very good low noise LDO power supply and very good PSRR, which has further improved the I listen.
All comments are welcome.
Thank you all

[the_bertrum]

I was wrong, I hadn't considered the noise that is added to the data by switching between 3 sets of switches.

As I said in my response to @WuBai, this is the claim that needs explaining.  No-one has described any way in which noise can alter the information held in the digital description of the information that is transmitted as the signal.
In an analogue system, the signal is the information, noise degrades the signal and therefore the degrades the information.
In a digital system, the information is put into a code that is transmitted as a signal, noise can degrade the signal, but as long as the code is not degraded, it can still be decoded to reveal the information.  If the noise affects the signal so much as to change the code, then the code will no longer be capable of being decoded and the information is lost.
What is required here is a plausible method by which the code can be changed by noise added to the signal in such a way that the information described by the code is altered in a coherent way.  By coherent, I mean that the alteration must fit in the context of the information transmitted such that the decoder still reads it as a valid message - not something random noise is likely to be able to do.  Anything incoherent will actually break the code rendering it unreadable.

So, I entirely believe that you do hear a difference, but I think the most likely cause of the difference is one of the well known and explained sources of listening error to which our brains are prone, and is not down to some as yet unexplained mechanism by which the data in a digital signal can be altered by noise.

At the end of the day though, the mechanism is irrelevant.  If having more switches between your disc and your DAC causes you to have a poorer listening experience, then you need to reduce the number of switches regardless of why you hear a difference.

[Jandu]

I was wrong, I hadn't considered the noise that is added to the data by switching between 3 sets of switches.

As I said in my response to @WuBai, this is the claim that needs explaining.  No-one has described any way in which noise can alter the information held in the digital description of the information that is transmitted as the signal.
In an analogue system, the signal is the information, noise degrades the signal and therefore the degrades the information.
In a digital system, the information is put into a code that is transmitted as a signal, noise can degrade the signal,

Only correct if the info is not time sensitive. However, music on a CD or a digital file has to be converted to music with a clock signal. In the conversion process, what is known as jitter is often introduced.

Have a read of the following article and you may have a sense of what gets altered in the process:

https://www.stereophile.com/content/cd-j...gic-page-3

[the_bertrum]

I was wrong, I hadn't considered the noise that is added to the data by switching between 3 sets of switches.

As I said in my response to @WuBai, this is the claim that needs explaining.  No-one has described any way in which noise can alter the information held in the digital description of the information that is transmitted as the signal.
In an analogue system, the signal is the information, noise degrades the signal and therefore the degrades the information.
In a digital system, the information is put into a code that is transmitted as a signal, noise can degrade the signal,

Only correct if the info is not time sensitive. However, music on a CD or a digital file has to be converted to music with a clock signal. In the conversion process, what is known as jitter is often introduced.

Have a read of the following article and you may have a sense of what gets altered in the process:

https://www.stereophile.com/content/cd-j...gic-page-3

Yes, agreed.  "In the conversion process", not in the transport from the storage medium to the convertor, in the case of digital files transmitted over network protocols at least.  Reading a CD is a different case entirely.

[philrandal]

Yes, agreed.  "In the conversion process", not in the transport from the storage medium to the convertor, in the case of digital files transmitted over network protocols at least.  Reading a CD is a different case entirely.

The only other thing I can think of which could affect sound quality is rf noise and power supply modulation affecting the analogue bit of the D/A process.

Phil

[Gate45]

I have carefully read John Atkinson's article published in "Stereophile" very complete, but with many doubts and truths.
I think we still have a lot to learn about digital signals and how to transmit them via cable or wireless without audible damage.
I have personally verified that it becomes difficult to convey a digital signal for a few tens of meters, with cable, passing through some switches.
The following statements have convinced me that we don't know much about it yet,

John Atkinson writes:

1) It is that if the ones and the zeros coincide, the sound must be identical.

2) "bits are bits"

3) Furthermore, it can be shown that a digital master tape and a digital copy of multiple generations have identical data (in a bit-for-bit comparison), even if they do not sound the same.

5) Clearly, some other phenomenon is at work. The most likely source of audible changes in digital audio is jitter.

6) The minimal amount of jitter in the word clock that synchronizes the D / A converter affects the musicality of digital playback. A shift of only 100 picoseconds (0.1 nanoseconds or 1/10 of a billionth of a second) in the word clock causes audible conversion timing errors.

For many years I have been concerned with getting good music reproduction and digital is the best solution for me. I have now verified that the signal passing through multiple switches deteriorates and I think due to the Jitter, but I have neither the tools nor the skills for such measurements.
Finally, because by replacing the DC power supply of the single switch I get so much difference. On the latter topic I am in tune with "philrandal" which finds the cause in the very high frequency noise that passing through the power supply interferes with the digital data and therefore an excellent power supply with high PSRR eliminates the problem.
Will there be anyone out there, a digital sound professional, who knows more about such interesting topics?
Thank you all

[the_bertrum]

The only other thing I can think of which could affect sound quality is rf noise and power supply modulation affecting the analogue bit of the D/A process.

Now we are talking.  There is a plausible mechanism there certainly.

I think we still have a lot to learn about digital signals and how to transmit them via cable or wireless without audible damage.

Nope, we know very well how to transmit digital signals without damage. You never send a £50 payment for something only to find the message damaged and it becomes £500, or £55. We send hi-res sounds halfway across the world from TIDAL and never once is the signal damaged.

I have personally verified that it becomes difficult to convey a digital signal for a few tens of meters, with cable, passing through some switches.

Finally, because by replacing the DC power supply of the single switch I get so much difference. On the latter topic I am in tune with "philrandal" which finds the cause in the very high frequency noise that passing through the power supply interferes with the digital data and therefore an excellent power supply with high PSRR eliminates the problem.

So here we begin to see a mechanism that works. If noise is being introduced by a switch and transmitted along with the signal to the DAC, then the noise can interfere with the information in the signal during and after the conversion. In this case there is no need to invoke some hitherto unknown method by which the information in the digital signal is altered in some subtle and non-destructive way simply by travelling through a different medium.

[Jandu]

Have a read on the following, it is about USB's way of transmitting digital info and audio signal. 



https://www.edn.com/fundamentals-of-usb-audio/

USB errors are fairly common, there is a good discussion:

https://www.head-fi.org/threads/new-audi...2/page-485

[philrandal]

Have a read on the following, it is about USB's way of transmitting digital info and audio signal. 



https://www.edn.com/fundamentals-of-usb-audio/

USB errors are fairly common, there is a good discussion:

https://www.head-fi.org/threads/new-audi...2/page-485

Thanks for posting that.

I once had an audio engineer explain these problems with USB audio. He said that, in his experience, the quality of USB cables did make an audible difference.

Cheers,
Phil

[Mellowship]

Have a read on the following, it is about USB's way of transmitting digital info and audio signal. 



https://www.edn.com/fundamentals-of-usb-audio/

USB errors are fairly common, there is a good discussion:

https://www.head-fi.org/threads/new-audi...2/page-485

Thanks for posting that.

I once had an audio engineer explain these problems with USB audio. He said that, in his experience, the quality of USB cables did make an audible difference.

Cheers,
Phil

I once tested a DAC with a USB cable that came with a zip drive and it was noisy as hell. Changed to another USB cable that came with a hard drive and no noise. Interesting thing is that the first cable had a good and tight external metal shielding and the second was very thin. 

My pi4 is directly connected with a normal usb cable, with just two ferrite beads, one at each end, to a smsl m500. I usually stream from tidal using mconnect upnp. Never noticed sounds degradation, distortion, noise, whatever. Using toslink from a mi box 4s, or coaxial from a NAD CD player, the sound is just the same when playing the same stuff. 

Also, I only use wi-fi, and the modern is just 2 feet away from the pi4. 

At 42, I'm starting to get less and less worried about these very small effects a certain tweak or change has on sound quality and more concerned about music itself.