Streaming over WiFi versus Ethernet Connection

[Guest rmpfyf]

1 minute ago, Steffen said:

It would have to be a lot of noise in order to be audible through the natural noise floor at your listening position (that's I think at least 20dB in a very quiet spot).

 

This should be easy to test, no guessing or expensive measurements required. If you put your ear right next to the tweeter or midrange driver of your speaker, while everything is turned on, volume up but nothing playing (or perhaps "digital silence" from a test CD), then you should hear some faint hissing. At least if your system is anything like mine . This should be faint and become inaudible once you're a foot away from the speaker.

 

You can now compare this close-up hiss across a variety of conditions – different network cables, Ethernet vs. WiFi, laptop on battery vs. SMPS, different power cables, action figurines holding up speaker cables, and whatnot. If you don't hear an appreciable difference at the very close range then you certainly won't hear it at the listening position. If you do, try to increase the distance between ear and driver, and see how far away it matters. If you hear a difference (or the hiss) at the listening position then I'd suggest something is seriously wrong with your system. Or, alternatively, your ears are so good that the sound of blood rushing through your vessels, or the beating of hearts in the next room must drive you mad...

 

What? No. A noise floor doesn't just pertain to SPL. It's a phenomena applicable to jitter spectra, SPL, anything measurable.

 

It's safe to say with digital the right bits usually get out the cast majority of the time. You're more concerned with timing accuracy.

[Steffen]

4 minutes ago, rmpfyf said:

 

What? No. A noise floor doesn't just pertain to SPL. It's a phenomena applicable to jitter spectra, SPL, anything measurable.

 

It's safe to say with digital the right bits usually get out the cast majority of the time. You're more concerned with timing accuracy.

Right, so you're concerned only with distortions introduced during D/A conversion. Aren't those easily* measurable, quantifiable? Isn't that what null testing reveals?

 

* easily, given the right equipment, that is. Like what they used when they compared a cheap Topping with an expensive Schiit, and found that the Schiit performs like sh*t in the distortion stakes, compared with the Topping...

[Ittaku]

Also be aware that from dithering research it has been shown that we can hear signal up to 25dB below the noise floor. Exactly how that applies in terms of what other noise/distortion effects we can hear on actual signal below the noise floor remains a complete mystery.

[aussievintage]

8 minutes ago, rmpfyf said:

Part of what's inherent in power saving in a Pi involves specific CPU timings that don't favour time dependent code execution.

You mean like the problems running a real time kernel described here  https://www.osadl.org/Single-View.111+M5c03315dc57.0.html.  They seem to have found the problem with the code and patched it.  But, OK,  most Pi systems won't be running this.

 

But is it audible?  I wonder - and yes I am a jitter cynic to a large extent.  I don't think a small amount is a problem.

[Guest rmpfyf]

1 minute ago, aussievintage said:

You mean like the problems running a real time kernel described here  https://www.osadl.org/Single-View.111+M5c03315dc57.0.html.  They seem to have found the problem with the code and patched it.  But, OK,  most Pi systems won't be running this.

 

But is it audible?  I wonder - and yes I am a jitter cynic to a large extent.  I don't think a small amount is a problem.

I can flip different kernels on mine - all realtime - all audibly different.

[Guest rmpfyf]

7 minutes ago, Steffen said:

Right, so you're concerned only with distortions introduced during D/A conversion. Aren't those easily* measurable, quantifiable? Isn't that what null testing reveals?

 

* easily, given the right equipment, that is. Like what they used when they compared a cheap Topping with an expensive Schiit, and found that the Schiit performs like sh*t in the distortion stakes, compared with the Topping...

Nope.

[Guest rmpfyf]

4 minutes ago, Ittaku said:

Also be aware that from dithering research it has been shown that we can hear signal up to 25dB below the noise floor. Exactly how that applies in terms of what other noise/distortion effects we can hear on actual signal below the noise floor remains a complete mystery.

So much null testing assumes a relative SPL difference itself below audible threshold isn't audible when added to an audible signal.

 

Which is wrong. 

[Steffen]

4 minutes ago, rmpfyf said:

So much null testing assumes a relative SPL difference itself below audible threshold isn't audible when added to an audible signal.

 

Which is wrong. 

I would say that's an issue with the interpretation, not the testing. Null testing will show whether there is a difference at all, and if so, how large it is. The difference can then be viewed in relation to other things that vary in a playback system. It would have to be easy to spot in order to justify claims of "massive" or "day and night" differences between Ethernet cables, etc., I'd suggest.

[Guest rmpfyf]

39 minutes ago, Steffen said:

I would say that's an issue with the interpretation, not the testing. Null testing will show whether there is a difference at all, and if so, how large it is. The difference can then be viewed in relation to other things that vary in a playback system. It would have to be easy to spot in order to justify claims of "massive" or "day and night" differences between Ethernet cables, etc., I'd suggest.

It shows if a difference in isolation is below an audible threshold.

 

Not whether a difference applied to an  phenomena within audible SPL is audible.

 

Which is why A/B, blind etc us infinitely more useful.

[Steffen]

I agree, A/B testing is the ultimate proof (the proof of the pudding is in the eating  ), if it's done correctly, i.e. precise levelling, double-blind, etc.

 

However, if measurements and null testing come up completely empty (like in the published cases of testing WAV against FLAC, for example), then the considerable effort to conduct a valid A/B test can be saved. The reverse is not true, of course, as you've stated above.

 

There is also a "reverse-reverse" case with null testing, where a test has shown large, obvious and unmissable differences between the test cases (lossless vs. lossy compression). Subjective tests, some of which were properly conducted A/B tests I'm sure, still missed these obvious differences, at least when the test subjects were casual listeners. I don't think audio enthusiasts were quite as easily fooled.

Edited September 10, 2019 by Steffen

[aussievintage]

8 minutes ago, Steffen said:

I agree, A/B testing is the ultimate proof (the proof of the pudding is in the eating  ), if it's done correctly, i.e. precise levelling, double-blind, etc.

I fully agree, but you so often get huge arguments if you mention it around here, I just gave up.

[Guest rmpfyf]

40 minutes ago, Steffen said:

However, if measurements and null testing come up completely empty (like in the published cases of testing WAV against FLAC, for example), then the considerable effort to conduct a valid A/B test can be saved. The reverse is not true, of course, as you've stated above.

 

No, again a fallacy.

 

Thus suggests you can't hear the difference between 90 and 91 dB because you can't hear a 1dB source from the same location, even if 90 and 91 are very much audible.

 

Null results apply to a null condition only. 

[Steffen]

59 minutes ago, rmpfyf said:

 

No, again a fallacy.

 

Thus suggests you can't hear the difference between 90 and 91 dB because you can't hear a 1dB source from the same location, even if 90 and 91 are very much audible.

 

Null results apply to a null condition only. 

That's not what I'm saying, and not what I'm talking about in the sentences you quoted. I mean that when there is no difference that can be measured then there is nothing that can be heard either (and an A/B test isn't going to reveal anything further).

[Guest rmpfyf]

57 minutes ago, Steffen said:

That's not what I'm saying, and not what I'm talking about in the sentences you quoted. I mean that when there is no difference that can be measured then there is nothing that can be heard either (and an A/B test isn't going to reveal anything further).

 

I know where you're going and I'm not trying to be adversarial.

 

Null testing is a joke for audio development.

 

Measuring jitter is legitimately hard other than bulk measurements. Different jitter characteristics sound different.

[Dankeshon]

That's not what I'm saying, and not what I'm talking about in the sentences you quoted. I mean that when there is no difference that can be measured then there is nothing that can be heard either (and an A/B test isn't going to reveal anything further).

Oh so we believe now that a machine mic can listen better than humans?

I give the machines one point, they are useful tools for what they can do, and they don't have feelings that disturbes or the famous placebo desease.

[Ed Sky]

So I managed to borrow a 5m cat5 cable from work yesterday just to give it a try. Just had one hour to test so my findings are really tentative.

 

So to cut to the chase… I think I am hearing a difference. I’ll need to spend more time to ensure that my results are consistent. Though the results have been convincing enough for me to put down some cash to get a better quality cat8 cable which I’ve just ordered. Should arrive in two to three weeks.

 

So my perceptions of the change are as follows: greater realism in high frequency details, blacker background allowing more ambient sound through, there is a slightly clearer sound, the music sounds more “awake”. Saying this, WiFi performance is also outstanding and I will admit that at this point, it would be difficult for me to pick out the differences in a blind test.

 

Incidentally it does appear that the Moon 390 switches off WiFi circuitry when the Ethernet cable is plugged in, judging by my router’s client list and also that the Moon required about 20 seconds to re-establish a WiFi connection once the Ethernet cable is unplugged.

[Guest rmpfyf]

You'll have internal EMI as long as the wifi hardware is powered. But a good step forwards so far. 

[Dankeshon]

To all sceptical to Ethernet cable.

I recommend that you try a specced shielded high bandwidth cable.

If you absolutely don't want to buy from hifi cable companies like cheap Supra or blue jeans.

Try a industry cable Siemens or likevides that has performance warranty.

[Guest rmpfyf]

5 hours ago, Dankeshon said:

To all sceptical to Ethernet cable.

I recommend that you try a specced shielded high bandwidth cable.

If you absolutely don't want to buy from hifi cable companies like cheap Supra or blue jeans.

Try a industry cable Siemens or likevides that has performance warranty.

 

Hifi ethernet cables are bulls***. Seriously.

 

Whether it's machine control, medical, data backbone or whatever else the need for reliable high bandwidth transmission and according shielding exists long before some obtuse audiophile wankery decided it was a thing.

 

There is nothing new in Ethernet cables.

 

Just get a decent cable from a reliable place. If you're paying more than $50 to solve that problem (and $50 is ridiculously high end) then you're doing something seriously wrong. CAT8 cables exist, are real and cost less. If you're worried in roof, RS components will sell you CAT7 rolls. The audiophile industry has nothing of any serious value to add here.

 

Same goes for optical converters, isolation, low noise switching... The lot.

 

There indeed are parallels to the power cable debate in that we've a tendency to think the problem is new or expensive. Neither true.

Edited September 11, 2019 by rmpfyf

[Guest rmpfyf]

And here we are, the ultimate for those wanting to go nuts... A bulk roll of CAT8.2

 

https://www.primuscable.com/store/p/15036-b-CAT8-2-Bulk-Ethernet-Cable-40G.aspx

 

No audiophiles were involved in the making of this product - though I'm sure within a year someone will stick lipstick on it, charge 20x and we'll have people here convinced it feels better.

 

Or you could just look at the data...

[Dankeshon]

And here we are, the ultimate for those wanting to go nuts... A bulk roll of CAT8.2  

https://www.primuscable.com/store/p/15036-b-CAT8-2-Bulk-Ethernet-Cable-40G.aspx

 

No audiophiles were involved in the making of this product - though I'm sure within a year someone will stick lipstick on it, charge 20x and we'll have people here convinced it feels better.

 

Or you could just look at the data...

 

We talking about from 15-20 bucks USD or ca 100 for a 20 meter. 

You don't need to break the bank

 

 

I checked your suggestion out! I absolutely recommend it. I have a cheap one myself for my TV box.

 

Shortened on demand Viasat from 2 minutes to 0 seconds.

I had to wait only to get to a soccer match, this upgrade

Changed it all to 0 seconds.

 

 

 

[Dankeshon]

Especially if you use freebees like my yellow Cat5 or some wallmart dogcats...

 

Price is not important at all, I have many cheap cables that works perfect.

 

What we that have experience promote, is that you choose well built specced cables...

 

Not the "Don't care, everything sounds the same! Use aluminium!"

 

 

 

 

 

[Addicted to music]

57 minutes ago, rmpfyf said:

And here we are, the ultimate for those wanting to go nuts... A bulk roll of CAT8.2

 

https://www.primuscable.com/store/p/15036-b-CAT8-2-Bulk-Ethernet-Cable-40G.aspx

 

No audiophiles were involved in the making of this product - though I'm sure within a year someone will stick lipstick on it, charge 20x and we'll have people here convinced it feels better.

 

Or you could just look at the data...

500ft is 152.4m at $699 a roll!    $4.50/m   Around $6.60AUD/m

Edited September 11, 2019 by Addicted to music

[Ittaku]

1 hour ago, rmpfyf said:

 

There indeed are parallels to the power cable debate in that we've a tendency to think the problem is new or expensive. Neither true.

My favourite is litz cable for power cords... That's not even a case of no science to support it, it's for exactly the opposite of what a power cord needs.

[Ed Sky]

I believe that CAT8.2 is not backwards compatible as it does not take RJ45 plugs. The same company does however stock CAT8.1 for less at $349US per 500ft: https://www.primuscable.com/cat8-1-bulk-ethernet-cable-40g-cmr-23awg-solid-copper-dual-shielded-sftp.aspx

 

There's also an Amazon AU option at $20 for a terminated 5m CAT8 cable: https://www.amazon.com.au/gp/product/B07JD8K7C7/ref=ox_sc_act_title_1?smid=AT2KEEZGY7CCU&psc=1 (They do have other lengths.)

 

I ended up going with an ELECOM CAT8 cable which is being imported from Japan so will take a while to get here. It's a bit pricier to import but still reasonably priced.