Ethernet Switch- what's the point?


I run an Ethernet cable between my router (standard issue from Verizon) and my streaming transport. I note that some use an ethernet switch between between the router and streamer. Assuming I got that right, what is the point- what does a good switch do? I've been into audio since the 70's but when its comes to streaming, I'm definitely a newbie- 

Thanks all!

128x128zavato

Showing 3 responses by asctim

All I can say is that I just love the way my Eero wireless mesh network sounds. I do have a switch of some kind, which is also a modem, which feeds ethernet to the Eero unit and a has a few more outlets that are unused, so I guess I'm listening to that switch too. It sounds great. 

So a good network switch can reduce noise and distortion, some of which comes from jitter. Anything else? How much of a difference should one expect? Like going from -90 dB to -95 dB noise to signal? Is it something very subtle but still worthwhile for some, or something quite loud and obvious to some, more so than typical room reflections and diffraction effects? I get the impression it's something far to subtle for me, and yet I can easily hear the difference between live music and any high end 2 speaker system. 

@tonywinga

 

They’re measuring the analog output. I’m not sure it’s all FFT, may be wavelet or whatever. I think it’s very safe to say that no spurious white noise or random tones or phase or timing errors are sneaking by the measuring equipment, unless they’re 140 dB or more down in level, in which case they fall below the noise floor and will not have any effect on soundstage or tonality for anyone no matter how good their ears or their playback gear.

For imaging, there may potentially be timing issues between the left and right channel. We are more sensitive to that kind of thing, but I think there’s a limit to that. I’m not sure what it is. We tend to notice directionality associated with timing differences between left and right ear starting at about 0.1 millisecond. Jitter involves timing errors vastly smaller than this. I think we can safely say we’re not hearing jitter in any way at all unless our gear is severely malfunctioning.

Might be something else though.

I’ve heard some suggest ringing of digital filters as digital’s big problem. This idea was popular for a while but some audio reviewer got ahold of a DAC that would let him apply and listen to filters he could program himself. He intentionally made some terrible filters in terms of having obnoxious amounts of pre and post ringing. To his surprise and amazement, these didn’t have much of an effect on the sound at all. It’s really not that amazing because the ringing is so far down in level, otherwise the test equipment would easily pick it up as distortion. So I think it’s safe to say it’s not digital filter ringing that we are hearing 99 percent of the time.

Could be something else though.

What could it be? We know that the signal in the range of human hearing is being very faithfully produced at the output of just about any DAC that isn’t broken. A cheap switch normally will provide the digital signal intact, allowing the DAC to function within it’s normal operating parameters. If adding an expensive switch to the signal path is doing something good it has to mean that something was going wrong with the previous switch. What could it be?

LP being able to present us with great sound is proof that we’re really not that bothered by noise and distortion, we may even perceive it as an enhancement. It may help us suspend our disbelief. I did an experiment once with digital photography, taking over 100 pictures of the same scene and then blending them together to eliminate digital noise. It worked. And the result was --- color banding! Without noise it’s impossible for there not to be banding across pixels over gradients. Think about it - how would you show a black to white gradient if you only had a picture 30 pixels wide? You’d see 30 bands of grey going from black to white. What if we created the same scene using a very small piece of photographic film? You’d see film grain and a lot of noise and a sort of blurriness, but no banding. If you add noise to the digital picture you get the same effect. The banding is no longer apparent. You notice the noise but you also perceive a smooth gradient behind it. Your mind can separate the noise from the signal, and the signal is remarkably perceived as being more detailed and natural.

DACs do add noise on the output, and they even shape that noise to make it less audible, so no equivalent of imaging banding is happening on a DACs output. 44.1 16 bit audio is the equivalent of an extremely high resolution photo viewed from an appropriate distance. But, our brains react to easily detectable noise, giving us an impression of the content behind the noise that’s different than when the noise is not noticeable. I find that adding easily visible noise to digital photos can sometimes have remarkable positive emotional effects in how I perceive the photo. I suspect that the perceptual problem with digital audio for some people is it’s lack of noise and distortion, not an excess of something that our ears can pick up on but measuring equipment can’t. Some noise gives our brain wiggle room to imagine, or I should say deduce, what’s not actually being heard.