After reading up on this phenomenon I wanted to test this out in the real world. I first thought of optical converters from switch and into my streamer (I know some streamers got optical network support directly) - with fiber cable between them. Then I though of the cheaper alternative - to use wifi. I also read multiple places that wifi were no good; due to often bad implementation of wifi in most streamers which actually lead to more noise internally in the streamer than the noise coming from a switch with noisy power supply.
Regardless - I first started to detach the ethernet cable from my streamer while playing (it got a buffer, so it plays for while without any connection). Here I found to my big surprise that the quality changed to the better. So I dug up the wifi-antennas for my streamer (Moon MiND2 Network Player) - and gave it a go with wifi and no ethernet cable connected. Same effect here; much better sound. And since there are no dataloss over TCP/IP - this can only be because of noise over the ethernet cable (either from environment picked up by cable or from other devices connected to the same local network).
Recommend everyone with a streamer to test this out. The difference in audio quality is big; with ethernet cable connected the sound got “muddy”. Sound quality is really great with ethernet, but so much better when using only wifi.
Please share experience you already have; or you find while testing this. My guess is that it will depend on equipment used. Some are far better over ethernet, other via wifi.
During testing I used a fairly high quality shielded 1.5 meter CAT6 cable. I also tested a selection of other ethernet cables; same result.
LAN is galvanic isolated using isolation transformers so there’s no direct connection including grounding between the router and the streamer. Personally I’ve tried both WiFi and LAN, I don’t hear much a difference. I use LAN due its robust and stable connection.
My understanding is that even it is galvanic isolated noise can still coupled through the isolation transformers. The CMRR of twisted pair LAN is around 40 to 50dB per channel so some noise can still pass through.
At the streamer side, depending how well the LAN receiver is designed to reject those noise. If the streamer has built-in DAC, then additional isolation is needed to prevent noise coupled to the DAC section. There are optical LAN coupler which can completely isolate noisy sources but they are expensive.
Likewise for WIFI streaming, the WIFI receiver at the streamer side must be well shielded to prevent EMI interferences to the nearby circuitry, especially if the DAC is built-in to the streamer.
I think it may me a difference if you are using a Raspberry pi as the USB and Ethernet share a bus.
I found using WiFi improved the sound in this instance. I also found it very close in sound to the Allo Usbridge, but gave the added benefit of working all the time and not needing to manually update Dietpi.
I have the exact same experience as the OP. I am a bit sceptical about SQ changes these kind of tweaks are supposed to bring, but using WiFi instead of a cat7 shielded cable actually made a huge impact. It was really easy to blind test the difference in my case. It adds more details in the mid en lower high range.
I am using an Aries G1 which, as later turned out, is manufacturer recommended to be used over WiFi instead of Ethernet for the exact same reasons as described in this topic.
This is not tweaks; it is just to check which input give the best audio
I know several brands recommend using WiFi, but they also highly recommend having really good and stable WiFi connection since if the WiFi receiver struggles that can actually generate more noise internally in the device than what you get over ethernet cable connected. So this is not something new; I just did not belive in the difference.
I recently replaced my TP-Link optical devices, that resided between the Aqvox switch and the Sonore Signature Rendu SE in my system, with a Gigafoil V4 from DJM Electronics. I am happy with the change and regard it as an upgrade. I use Sonore lps device to power first the TP-Link and now the Gigafoil.
I am wondering if anyone has put any ethernet filters between the server (in my case a Nucleus+) and the ethernet switch. Would it have any beneficial effect, and would it make a difference if the link to the bridge/dac was already well filtered?
While I was waiting for my GigaFoil v4 to arrive I placed the Audiquest Diamond ethernet cable between my soninTransporter and the switch just to initiate break-in. I swear I heard improvement in my SQ for the three days I waited for the GigaFOIL. I then moved the Diamond ethernet cable and inserted it between the GigaFOIL and my dCS Upsampler. I was very impressed with what the combination of GogaFoil and AQ Diamond did to my overall SQ. I may pick up another Diamond ethernet cable to place back between the i5 and swith just in case. Today I added a Keces LPS to power the GigaFOIL. So far so good, I’'ll know more after about a week of break-in.