[camrector]

That describes my system. NAS > copper > Silent Angel N8 > copper > oMD > 15 meters fiber > ER > copper > PlayPoint

Wrong. The ER isolates via the moat.

So you just use the ER as a FMC or do you have network components plugged into the LAN inputs?

 

[camrector]

Wrong. The ER isolates via the moat.

And again what needs a moat when using a FMC?
Maybe the noise from the cheap accusilicon femto clock needs a moat.
If you are only using the ER as a FMC then, I highly suggest you upgrade to something with higher quality internals.

 

[Psilonaught]

The ER is designed specifically for this situation. Connect to it via fiber, connect to the streamer via copper, which is completely isolated by the ER moat.

The Hifi Rose RS130 doesn't have an ethernet port. SFP only.

That's why I refuse to add a copper + switch chain to my simple router > fiber > streamer setup. Utterly pointless.

What I would like to do is play around with cables and transceiver to see if I can improve on the CommScope/ Finisar 1318 combo.

 

[cpurdy]

In analog audio, you would be correct. With network audio, a switch can clean and reclock the data stream, which will improve the sound. Adding a high quality audiophile switch between a WAP and network player will sound better.
https://www.alpha-audio.net/review/a-very-deep-dive-into-network-switches-listening-and-measuring/

This is incorrect. An Ethernet switch is not allowed to "clean the data stream". There is no such concept to begin with. Don't take my word for it; you can actually read the specification for yourself: https://standards.ieee.org/ieee/802.3/7071/

DACs simply translate bytes in RAM (not "on the network") to analog signals. This may seem like magic (ref Arthur C Clarke), but it's not. And the network is nowhere in the picture by that point: Those bytes are just sitting there in DRAM, waiting to be read by the DAC. And they're often not even the bytes that were passed (with 100% verified correctness) across the network, since they go through a series of transformations, starting with TCP streaming (de-packetization), then decompression (for compressed data formats), etc.

(Not that it makes any difference, but it is theoretically possible -- with DMA or RDMA -- for those bytes to have been placed there directly by the NIC, and then read out via DMA by the DAC hardware. I used to work with Intel on RDMA-based protocols on 40-160Gbit networks, which was a bunch of fun.)

Prediction: This message will self-destruct (with help from a moderator).

 

[griff500]

An Ethernet switch is not allowed to "clean the data stream"

... if it is to meet those published standards. I would assume there's nothing to stop someone from producing one that differs from the standards required for data networks if they feel it would produce benefits for a specified purpose? I'm not saying that is being done - don't know and don't particularly care.

 

[audiobomber]

What I would like to do is play around with cables and transceiver to see if I can improve on the CommScope/ Finisar 1318 combo.

Head-Fi.com is the best site for headphone audio. The premier site for advice on network audio is Audiophile Style: https://audiophilestyle.com/

 

[monf3307]

... if it is to meet those published standards. I would assume there's nothing to stop someone from producing one that differs from the standards required for data networks if they feel it would produce benefits for a specified purpose? I'm not saying that is being done - don't know and don't particularly care.

these are 0s and 1s. they are right or wrong. if the packet does not pass checksum it is simply rejected. in this manner i suppose you could say that all tcp/ip "cleans" the stream.

also there are no timing issues like there are in SPDIF, because the transport carries orders of magnitude bandwidth than the richest audio stream. it is up to the receiving component to slow down and buffer as appropriate

 

[griff500]

these are 0s and 1s. they are right or wrong. if the packet does not pass checksum it is simply rejected.

That has been discussed ad nauseum. Feel free to look back through the thread.

 

[bfreedma]

these are 0s and 1s. they are right or wrong. if the packet does not pass checksum it is simply rejected. in this manner i suppose you could say that all tcp/ip "cleans" the stream.

also there are no timing issues like there are in SPDIF, because the transport carries orders of magnitude bandwidth than the richest audio stream. it is up to the receiving component to slow down and buffer as appropriate

Posts documenting how the technologies being discussed actually work are not popular in this thread.

 

[camrector]

Posts documenting how the technologies being discussed actually work are not popular in this thread.

Similar to the Master Clock thread.
Try to tell someone how things actually work and they loose their minds!
Marketing has gotten good at brainwashing.
One guy over there literally said “it’s Chinese secrets”
Same vibe here

 

[griff500]

@camrector It’s nothing to do with that. It’s down to the strawman argument that data isn’t changed or improved, this being something that nobody has ever claimed.

The person you replied to is on my ignore list and so I cannot see what he said, but I can guess.

 

[camrector]

@camrector It’s nothing to do with that. It’s down to the strawman argument that data isn’t changed or improved, this being something that nobody has ever claimed.

The person you replied to is on my ignore list and so I cannot see what he said, but I can guess.

There is no such thing as “clean the data stream” lol
That is make believe and marketing brainwashing.

 

[griff500]

There is no such thing as “clean the data stream” lol
That is make believe and marketing brainwashing.

It seems a rather vague choice of words and I would want to know exactly what is meant by that. It’s not my choice of words and it’s not a phrase that I’ve ever used in connection with this topic.

 

[cpurdy]

... if it is to meet those published standards. I would assume there's nothing to stop someone from producing one that differs from the standards required for data networks if they feel it would produce benefits for a specified purpose? I'm not saying that is being done - don't know and don't particularly care.

Those published standards are why every stream of data that goes over a network is absolutely 100% bit perfect. Sure, you could build a network switch that "cleans the data stream" (which might mean purposefully corrupting the data in random ways, but in reality I think it's just "do absolutely nothing different from the $10 switch"). If you even get one bit wrong, chances are pretty good that the audio player will segfault (which tends to get caught and then the player will just automatically move to the next song in the list). The reason that this happens is that audio formats often have indexes (basically, pointers) encoded in the stream, and those indexes are referencing into fixed size arrays that were encoded earlier in the stream, so getting even one bit wrong will often cause the index to be completely illegal, e.g. turning the number 7 into some giant negative number, which will either generate a hardware protection fault (or OS protection fault, depending on the combination of hardware and OS), or just yield invalid data if the resulting pointer is accidentally valid within the current process. So if you're starting with bit-perfect data, it's probably a really, really, really bad idea to "clean" it. (Not that I can even imagine what "cleaning the data" means.)

I don't understand the claims about electrical line noise, since those are all supposedly accounted for in the spec (galvanic isolation and all that), but who knows, maybe someone took a shortcut somewhere. It shouldn't be possible for this to be an issue, but I'm not an electrical engineer. And manufacturers have been known to cut corners from time to time.

I really have no problem with people overpaying for audiophile gear. I pay way too much for this stuff, and often make my decisions based on how something makes me feel. The beauty of equipment or a cable means a lot to me. It does make my music sound better, because a dramatic portion of our appreciation of sound is based on our mood, and my mood is better when my stuff looks awesome.

But I don't go around making up technical reasons why my unidirectional oxygen free cables are somehow better. Instead, I just accept that I'm easy to manipulate (thank God!) and I really like the way some artist (and let's be honest: they're artists) made that cable for me, so I overpaid for it because it magically results in my music sounding better.

 

[audiobomber]

Worth watching:


Counter comment from Uptone Audio:
https://audiophilestyle.com/forums/...ng-on-page-92/?do=findComment&comment=1256277