are NET Switches worth considering?

Always wanted to separate audio network from the regular LAN, finally bit the bullet and got it working. All copper cabling, nothing special, one $120 Router, EtherRegen, Lumin U1, Neo stream, Eversolo A8. The rest of the network can't access the audio network at all due to routing rules and config of a JCAT NIC in the Music Server. 

Access to the internet for updates is provided by the router which obtains access through the main router. No switches needed, other than the EtherRegen. Wifi is switched on to allow the remote to work for Audirvana. HQPlayer plays to the A8 as NAA, it all works.

The sound has made a very good, 'organic' improvement, can now hear drumsticks hitting cymbals first, rather than just the decay of the metal. Vocals are rock solid, take on great front to back depth, suspended in mid-air if you like.

Most of my components are not new by all means, well the A8 is, the rest are 10 years plus old, made a big difference to what I hear. Noise reduction as far as possible is the goal, noise will always be present when it comes to networking, it's lessening the impact with a smaller network to begin with.

 

@tcutter I'm getting the impression that all one needs is a fibreoptic setup anywhere in the pathway between modem\router and the DAC. Correct?

Generally, fiber moats should lower the noise floor, but some have reported it reduces the “life” of the music.

isn't it measurable?  Whether the switch makes a difference in the data delivered to the DAC?

Generally, fiber moats should lower the noise floor, but some have reported it reduces the “life” of the music.

if this is the sonic result, you have not done it correctly

@jjss49 

+1, Indeed! 

The proof is in the implementation, so let’s re-iterate for anyone following along.

A basic SFP + media converter chain often recommended here because it’s inexpensive; typically looks like this:

Router → RJ45 → media converter → SFP module → fiber → SFP → media converter → RJ45 → streamer

On the surface, it achieves optical isolation. But it also introduces multiple conversion stages, generic clocks, switching power supplies, and a fair amount of uncontrolled variables at each junction.

In other words, it’s a networking workaround not an audio-optimized solution.

That distinction matters. Because once you move to a purpose-built device like the Matrix Audio SI-1, you’re not just isolating…you’re controlling power, clocking, signal integrity, and noise within a single, coherent design. That’s where the real gains come from, and why two ‘optical’ solutions don’t necessarily sound the same in practice, IMHO.