If a product is sold on the basis that it reduces jitter, or something like that then it absolutely should be measured and shown to do that otherwise it shouldn’t be able to make that claim.
Whether having your jitter reduced sounds better is up to the listener.
The nefarious part is that the brands who market these products know that they don’t really do anything measurable, so they carefully avoid using terms that would imply that.
Take this marketing text for the EE8 switch for instance:
"In the world of streaming music, there is a belief that digital signals by their nature are immune to the failings of analogue transmission – but it’s not that simple – and even the most hardened digital sceptic may be surprised at what can be achieved with a high specification Ethernet switch.
Listening to streamed music is not just about measurements and numbers. While both elements have important roles during product design and development, the real test for music lovers is how it sounds
The 8switch by English Electric (a Chord Company brand), is an 8 port GbE Ethernet switch modified and upgraded to our specific requirements for high-performance music streaming networks.
Have a listen and test it yourself!"
So it’s magic basically, and anyone asking for proof is easily dismissed on that basis.
That same blurb does actually say that:
“The 8Switch jitter measurements indicate an improved network signal performance of up to 90%”
So that is measurable, whilst the other claims are not.
Personally I’m prepared to believe that some things that affect the sound cannot be measured (although I’ve not found Ethernet switches and cables to make a difference for me). But a claim that an up to 90% reduction in jitter should be substantiated and it’s then up to the listener to decide if they like that.
Interesting, that part wasn’t included on the page from which i copied the text…
It’s just that this means nothing at all. What’s “improved network signal performance”
But the dac sorts out jitter and clocking anyway to what it wants, so can’t see the point
“up to 90%” doesn’t say anything obviously, since 1% also falls into that category.
Also if they use relative percentages, if the jitter before was 0.0002% and with the EE8 it’s 0.0001%, that would be a 100% improvement.
I guess there’s levels of ‘bad’. The best measuring product might not sound the best, but I doubt a product that truly measures badly will sound good.
That occurs sometimes, specially with some tube amps.
Here some measurements, to please the motivated numbers guys.
From Hifi news, lab report on Melco S100 switch.
REPORT
MELCO S100
Melco’s custom dataswitch solution employs additional PSU smoothing, 1.5MB of FIFO buffering and a proprietary clocking system in its bid to deliver a low-noise, low-jitter Ethernet packet stream. Nevertheless, as the music data remains in the digital domain across the network (hubs, routers or switches notwithstanding) any uplift in subjective performance over a conventional NAS or other music library can only be inferred via a third-party DAC. Secondary re-clocking or jitter suppression within the network-attached DAC is also a factor here, so a
DAC with excellent performance may not express a significant difference. Similarly, a DAC that incurs jitter at the chip level will not improve regardless of the S100’s data signal conditioning.
Three very fine network-attached converters were tested with and without the S100 switch, using a 0.5m length of Melco’s own CA1E cable – the Mytek Brooklyn Bridge [HFN Dec ’19], Arcam CDS50 [HFN Jun ’20] and Lumin D2 [HFN Jul ’20]. The Mytek and Arcam DACs exhibit superb jitter rejection by default and the impact of the S100 was at the limits of measurement (~1psec). The Lumin D2 did resolve a repeatable difference, however, with correlated jitter reduced from 15psec to 10psec and uncorrelated noise squeezed by 0.4dB over a 146dB range [see Graph, below]. PM
ABOVE: 48kHz/24-bit jitter spectra from a network- attached Lumin D2 DAC (without Melco S100 switch, red; via the S100’s 100Mb ports, black)
HI-FI NEWS SPECIFICATIONS
Ethernet Ports (RJ45) 4x100Mb / 4x1Gb
Ethernet Ports (SFP/optical) 2x100Mb
Digital jitter (Arcam CDS50) 5psec (6psec without S100)
Digital jitter (Mytek Brooklyn Br.) 5psec (5psec without S100)
Digital jitter (Lumin D2) 10psec (15psec without S100)
Power consumption 6W
Dimensions (WHD) / Weight 215x61x269mm / 2.5kg
“A picosecond is an SI unit of time equal to 10−12 or 1⁄1 000 000 000 000 (one trillionth) of a second. That is one trillionth, or one millionth of one millionth of a second, or 0.000 000 000 001 seconds. A picosecond is to one second as one second is to approximately 31,689 years.”
This means the jitter reduction was from 0.0000000000015 to 0.0000000000010 seconds on the worst performing streamer they had available (Lumin D2). On the other streamers the jitter was either identical with the Melco S100, or only 1 pico (trillionth of a) second improved.
0.4db over 146db noise reduction is way way below the human hearing threshold, which is around 120-130db under the most optimal circumstances (completely silent environment without any background noise).
Remember that we are considering the reports from users who describe significant and audible changes in audio quality from changing switches.
I do very much welcome the addition of the extra measurements, thank you for that!
I posted that because you were saying that no brand communicates on the measurements of their switch. As Hifi news tried to measure something, I was curious to know what you or others could think of that.
Thanks for your feedback.
So nothing of that should be considered as possible to hear by the human ear. Same goes for measuring speaker cables. I’ve never seen someone measure cables and prove by data that they should sound different. Just buy a thick one with little resistance loss and you’ll be fine. Still most of us claim to hear a rather big difference. What is it in cables and equipment that cannot be measured that the ears and brain still can conclude to as an improvement?
It’s a bit different i think, speaker cable is analog and goes between your amplifier and speakers, that is the most messy part of the chain (except perhaps when playing vinyl) where a number of things can have quite a bit of influence on the sound. The digital part in front of the DAC is a lot more precise and controlled, so the changes are much smaller there.
this is how the Silent Angel Bonn N8 (aka the English Electric 8) switch (in red) measure vs a standard switch (in blue) on jitter noise as per ASR
almost identical performance to the Melco S100 above
Yes, from Audioscience certainly.
Yes, and they did a few other tests on noise floor and deliberately ramping up network traffic to see how it would affect performance etc. In the end, 0 difference, and hence their conclusion that expensive “audio grade” switches such as the N8 make no difference.
Really, do you have any examples? I’d be interested.
Bad can of course be relative, for example the B&W speakers which are popular on this forum don’t measure that well (flat) with regards to frequency response, but many people do like that signature. The drivers individually have excellent measurements which probably also helps.
But speaking of tube amps, I looked up your EAR to see if you meant that, but that seems to measure just fine. Perhaps not as good as some solid state designs, but still. (I also use a tube pre BTW)
Don’t remember, sorry…but I often read that on Stereophile or even Hifi Plus ( in his great years).
Stereophile make measurements for all the reviews they publish.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.
