Wifi vs hardwired for streaming

@pcolvin @nigeltheflash 

The literature and other content on this topic is probably in its infancy. The basic problem is that few audiophiles actually understand the network layer and the electrical properties of the media and the analog transmission of digital bits.  Surely we all grasp that what is being transmitted are "1"s and "0"s, but it the "HOW" that is the key when it comes to minimizing or eliminating co-transmitted analog noise with that digital content. Good luck getting a coherent and through understanding of that.  First, the network guys (unless they are national security) don't care as long as the connection passes testing/verification.  The data is good, the error rate is low, the retransmits are low, ...All is good they would say.  But we are audio guys and we are hardly ever electrical engineers who understand the analog noise co-transmitted with the digital content.  And who does?  As I said, probably in the early days of understanding the limits of removing that analog noise... The last leg conversion to fiber demonstrably improves the situation remarkably -- but does it get everything? DO these devices themselves need "innovation"? I think yes...

Repost with edits:

@vicweast  You’re usually right on the point that most network guys don’t understand analog noise.  And, I believe, that in most cases, many audio guys who work with studios don't understand networks and network noise. I, on the other hand, am an audio guy who is also a network guy, like yourself. 

Honestly, I’ve never really analyzed analog noise from an SFP. I think it’s because we simply buy what we need (mm, LC, SC, etc.) and go with it. There’s just too many suppliers providing SFPs in almost any price range. Also, where do you find the equipment or the published processes to do this? I think that we’d have to start with popping open SFPs and spec’ing the various optical converters. And even if you do build a list of the various chips, and find those that truly give you low analog noise from the DAC, there’s no way to create a master listing of what SFP uses what chip that’d be consistent.  Also, I’m sure, the problem would also be exacerbated by the SFP receptacles and supporting circuitry.   

So we simply punt. 

On the flip side, isn’t analog noise transmitted digitally something that’s encoded at the transmitting end? 

@pcolvin Thanks for that, it’s not very difficult to measure how noisy something in the network is, and its not hard shooting a space for EMI. You can buy a cheap handheld that will expose nasty behind your audio stack and along various cable runs. So, I think that’s the key, not at the sub-product level. Except for their power supplies... These are likely the worst offenders that we will find at the end of this exercise in cleaning the digital inputs from their analog electrical attached noise.

I think the voltages at the fiber media converter’s ethernet output are pretty limited and small, but yes that perhaps is a place of noise that progress will run up against after this kind of signal clearing becomes something to chase... 

Noise isn’t encoded, it is induced into the copper wires that electric voltages transit. The 1’s and 0’s are the data sent via such voltage. The rest of the wires manage the process, more or less at the fabric level -- the physical layer and the data-link layers in the OSI model. The protocols etc sit above these two.  It is the electric on the physical layer we get noise on.  

The concern isn’t how it might "corrupt" the data because a 1 and a 0 are the only two possible values and should one be zapped into the other, the TCP/IP protocols higher up will detect an error and a correction or retransmission will be guaranteed. It just works like that.

The concern for audio is: Is that attached electrical noise effecting the behavior of the audio device by somehow polluting some internal-to-the-audio-appliance process....?  

Thus: We wish to clean the link (not the data) into the audio appliance. The: an audio grade network switch is NOT what is needed. And what we need to do we only have to do the last leg...

It is the analogous problem with noisy interconnects and power cords that are not audio friendly.  What’s behind the wall ...who cares? How dirty the power is ...who cares?  As long as we can somehow provide clean power TO the audio appliance, "we good baby". (unsolicited plug for PS Audio Power Plant)

Wireless does this to a degree because the noise stops at the radio. But wireless issues are real so it’s not ideal. Media conversion to fiber does it and it is really cheap for what you get.

@vicweast Re the post in which you namechecked me: spot on, well said.

@pcolvin No, analog RFI noise is generated/transmitted/conducted/radiated/whatever as analog noise not converted. To be encoded digitally somehow would imply the data changes or is added to do and it doesn’t/isn’t, which is why corporate networking guys don’t see it as a thing.

@vicweast Re audio grade network switches not being "needed"; it really depends on what is incorporated into the design of a switch which makes it "audio grade". If it is primarily about ensuring the switch puts out less noise alongisde the output signal than it received alongside the input signal then it deserves the badge.