Best audiophile switch

[StreamFidelity]

The streamer side is limited to 100 mbps. I don't understand this obsession with high-capacity switching. Last time I looked, and I stream at DSD256, the traffic out of the fibre port to my hifi was 34 mbps.

The required bandwidth can be easily calculated:

DSD256
44,100 x 1 bit x 256 x 2 (stereo) = 22,579,200 bits = 22.6 Mbit/s

The reason you have measured more is that even more data goes over the line than the pure music signal.

DSD1024
44,100 x 1 bit x 1024 x 2 = 90,316,800 bits = 90.32 Mbit/s

The 100 Mbit/s would no longer be sufficient here.

Why the need for 10 gb switches?

The point is that a switch with a bandwidth of > 10G must meet higher specifications. This is logical, as a higher bandwidth must have less latency and jitter for the data transport to work.

A switch that meets the specifications of 10G and more has significantly better components, less jitter and less noise. Of course, it always depends on the implementation in the device.

 

[ssfas]

Now that you have had a few day's with it in your system, what are your thoughts? I'm thinking of ordering one...

Positive. Bear in mind I put my streamer on a set of Stack Auva feet at the same time.

I had a switch with an upgraded clock, a 2-box LPS, but no DC filtering. I also had an inline EMOSytems isolator, which does the same job as the SmoothLAN product that @Republicoftexas69 has in his chain. I took out the switch and the EMO and just put the LAN Regenerator between FMC and streamer.

The SmoothLAN Regenerator has the same functionality and seems to do it a lot better because there is far less of it, it's cleaner, 100 mpbs, and the reduction of boxes is a big deal for me.

The old switch (it was white before I sprayed it), the EMO filter and the new LAN regenerator are pictured together below.


I think Stack comprises a product designer and Josh on sales and marketing. My son is a product designer of consumer products, so I have some idea how it should work. You don't have to be a genius engineer because the technology is not rocket science, and you can buy in tech (most audio companies do). You just need to know how to conceptualise, design, and source the product at the most affordable cost and then market it. All their products are in-house designed with local suppliers. (Rega are the masters of this.) This is the complete opposite of most audiophile network products and cables (with notable exceptions like Buffalo/Melco, SoTM and some others) that are a generic product with some upgrades, a nice case, a veneer of b/s and a crazy price tag. Josh is clear that they see direct selling as a means of keeping prices down (like Schiit, Buchardt, etc), not as an opportunity to charge prices based on inflated profit margins.

 

[ssfas]

The required bandwidth can be easily calculated:

DSD256
44,100 x 1 bit x 256 x 2 (stereo) = 22,579,200 bits = 22.6 Mbit/s

The reason you have measured more is that even more data goes over the line than the pure music signal.

DSD1024
44,100 x 1 bit x 1024 x 2 = 90,316,800 bits = 90.32 Mbit/s

The 100 Mbit/s would no longer be sufficient here.


The point is that a switch with a bandwidth of > 10G must meet higher specifications. This is logical, as a higher bandwidth must have less latency and jitter for the data transport to work.

A switch that meets the specifications of 10G and more has significantly better components, less jitter and less noise. Of course, it always depends on the implementation in the device.

The limiting factor in my system is the Mac Mini M1, which can only do HQ Player reliably up to DSD256.

I appreciate that there are different performance levels (duplex and single mode fibre, CAT 5/6/7/8, 100 mbps, 1gbps, 10 gbps switching) but as far as I understand they only relate to increasing data volumes and distance of travel. Duplex fibre, CAT 5, 100 mbps are plenty sufficient for at least DSD512, and 100 mbps is considered favourable as it produces less jitter than faster switching.

The good thing is that networking is based on performance specifications and you can measure cable transmission loss, filter transmission and return loss, clock phase noise and see your switch traffic in real time (with decent software).

If you want a 1 gbps switch, this product is not for you. Innuos did the same, their servers were always below Roon i3 minimum specification and never got Roon certification, but they work perfectly well and are very popular. Innuos told me years ago if I wanted to run 8 zones with EQ to get a Roon Nucleus.

 

[ssfas]

"I once did a homebrew bandwidth test. I streamed 1080P Netflix from the internet and DSD 512 (8X DSD, 512 times the sampling rate of a CD) from my NAS, simultaneously, through a 100Mbps dual pair ethernet cable. My streamer and TV were both connected to an EtherRegen. Both systems performed flawlessly."
The above is from a post I made previously.

DXD (352.8 kHz, 32-bit) and DSD512 use about the same amount of data. 100Mbps is overkill for audio transmission and may be sonically advantageous, according to several manufacturers. My SOtM renderer can select between gigabit and 100Mbps, and definitely sounds best at the slower rate.

I looked very seriously at a complete SoTM system to the point of receiving an invoice. I cancelled because of the box-count.

Their design seems very focused on the job in hand, and all proprietary circuits and design. Their switch with clock option was very tempting.

 

[Republicoftexas69]

I gave Josh a bit of feedback today, firstly on the website and the fact that you should receive what looks like a $5 5v wall wart that is in fact a good quality German medical grade power supply that costs about $40 trade. There is no literature included, I suggested he includes some form of note to the effect that there is no need to get an after-market replacement.

There is a biggish forum in the UK where there were a load of ignorant non-customers over the last few days saying 100 mbps is insufficient and the whole idea is nonsense. My switch was saying 34 mbps for DSD256 and he told me they've not had it above 50 mbps, presumably 24/192 PCM.

Well you said it, Ignorant Non Customers on a FORUM.

 

[ssfas]

Well you said it, Ignorant Non Customers on a FORUM.

It wasn't entirely me who said it!

Someone decided 100 mbps was too little bandwidth because some AI thing (ChatGPT?) told him.
It was to some extent the ones and zeros brigade not comprehending the phase noise issue in digital streams, discussed extensively above. For me it was a unified and more focused device than what I already had.
Josh said these negative forum chats are quite helpful in revealing prejudices, even if uninformed. I joined the forum simply to say DSD256 does work because I plugged it in and it did.

So I may not be entirely ignorant of the aim of this product to work on the data stream, whereas the SmoothLan works on non-signal noise, but I remain entirely ignorant of the merits of audiophile ethernet cables as opposed to any others that meet IEEE spec. I was thinking of getting a pair of the Neotech ones you have, given I have Neotech usb and analogue cables (silver UP-OCC), I can then say I've tripped the audiophile-ethernet-light-fantastic and at $100 each I can't feel too ripped off, but that's my limit.

 

[Republicoftexas69]

It wasn't entirely me who said it!

Someone decided 100 mbps was too little bandwidth because some AI thing (ChatGPT?) told him.
It was to some extent the ones and zeros brigade not comprehending the phase noise issue in digital streams, discussed extensively above. For me it was a unified and more focused device than what I already had.
Josh said these negative forum chats are quite helpful in revealing prejudices, even if uninformed. I joined the forum simply to say DSD256 does work because I plugged it in and it did.

So I may not be entirely ignorant of the aim of this product to work on the data stream, whereas the SmoothLan works on non-signal noise, but I remain entirely ignorant of the merits of audiophile ethernet cables as opposed to any others that meet IEEE spec. I was thinking of getting a pair of the Neotech ones you have, given I have Neotech usb and analogue cables (silver UP-OCC), I can then say I've tripped the audiophile-ethernet-light-fantastic and at $100 each I can't feel too ripped off, but that's my limit.

I am not disagreeing. 1G for Audio think not.

 

[Di-fi]

100Mbps is overkill for audio transmission

Agreed, but managing local libraries with a 100 Mbps switch is a pain. So, it's best used exclusively for streaming purposes.

 

[Republicoftexas69]

Positive. Bear in mind I put my streamer on a set of Stack Auva feet at the same time.

I had a switch with an upgraded clock, a 2-box LPS, but no DC filtering. I also had an inline EMOSytems isolator, which does the same job as the SmoothLAN product that @Republicoftexas69 has in his chain. I took out the switch and the EMO and just put the LAN Regenerator between FMC and streamer.

The SmoothLAN Regenerator has the same functionality and seems to do it a lot better because there is far less of it, it's cleaner, 100 mpbs, and the reduction of boxes is a big deal for me.

The old switch (it was white before I sprayed it), the EMO filter and the new LAN regenerator are pictured together below.
View attachment 146827

I think Stack comprises a product designer and Josh on sales and marketing. My son is a product designer of consumer products, so I have some idea how it should work. You don't have to be a genius engineer because the technology is not rocket science, and you can buy in tech (most audio companies do). You just need to know how to conceptualise, design, and source the product at the most affordable cost and then market it. All their products are in-house designed with local suppliers. (Rega are the masters of this.) This is the complete opposite of most audiophile network products and cables (with notable exceptions like Buffalo/Melco, SoTM and some others) that are a generic product with some upgrades, a nice case, a veneer of b/s and a crazy price tag. Josh is clear that they see direct selling as a means of keeping prices down (like Schiit, Buchardt, etc), not as an opportunity to charge prices based on inflated profit margins.

"EMOSytems isolator, which does the same job as the SmoothLAN product that @Republicoftexas69 has in his chain"
Well no as I own them both and the SmoothLan is more refined and imaging is better.

 

[ssfas]

"EMOSytems isolator, which does the same job as the SmoothLAN product that @Republicoftexas69 has in his chain"
Well no as I own them both and the SmoothLan is more refined and imaging is better.

... and the LANRegenerator seems more effective that my switch with upgraded clock. Suggests Stack's focused approach to re-engineering these components is working.

The good thing about 10 gbps ethernet is that if you go deaf you can open a data centre.

Agreed, but managing local libraries with a 100 Mbps switch is a pain. So, it's best used exclusively for streaming purposes.

Agreed, which is why this unit only has one input. Music libraries can go anywhere on a network.

 

[Republicoftexas69]

Agreed, but managing local libraries with a 100 Mbps switch is a pain. So, it's best used exclusively for streaming purposes.

No problem here.

 

[audiobomber]

I am perfectly fine with having my audio chain limited to 100Mbps, and that is how I run, but I will not limit my entire home network to that speed.

 

[PYP]

I'm interested in what difference grounding and footers make with the various switches.

Grounding the Synergistic switch had a very positive effect. Much more weight and more dynamic while also having an even better sense of effortlessness (great combination of qualities). What was more surprising was how using Stack Audio footers resulted in an even greater experience of the recording space. I've listened to He for She (En Vivo) with Magos Herrera and Javier Limon many, many times because I enjoy the vocals, the music and sense of camaraderie. One might not expect much from streaming 44/16, but what I just heard placed me there in a way that I haven't heard previously. Stunning.

 

[ssfas]

I am perfectly fine with having my audio chain limited to 100Mbps, and that is how I run, but I will not limit my entire home network to that speed.

100 gbps is only in relation to components just before the streamer, not the whole network. I think 1 gbps is plenty for a home network. When I looked at the audiophile offerings, there were a couple based on a QNAP switch that had some 10 gb ports. QNAP obviously design for enterprise level networks.

What was more surprising was how using Stack Audio footers resulted in an even greater experience of the recording space.

I put some Stack Auva footers under my streamer as well and they seem to pay big dividends. They seem to move in three directions rather than just up and down, much the same as Townshend products, but they are more practical and a cheaper as well.

 

[Superdad]

I think Stack comprises a product designer and Josh on sales and marketing.

I’m 95% certain that Stack commissioned John Westlake to engineer this piece. I can spot his layout technique from space.

 

[Republicoftexas69]

I’m 95% certain that Stack commissioned John Westlake to engineer this piece. I can spot his layout technique from space.

[I’m guessing that part of why Stack keeps quiet about it is that Westlake, while a sharp engineer, has a good bit of controversy following him over long-delayed and then abandoned altogether projects for which he collected a lot of money from users in advance. You can Google that.]

A little insight into Mr. Westlake. https://en.wikipedia.org/wiki/John_A._Westlake

 

[Synaxis]

I put some Stack Auva footers under my streamer as well and they seem to pay big dividends. They seem to move in three directions rather than just up and down, much the same as Townshend products, but they are more practical and a cheaper as well.

I agree.

I've been experimenting with vibration control under my switches for about two years and have tried at least 20+++ products and combinations. .

I can do better than the Stack feet on my switches, but at a much higher price. Stack is reasonably priced and have an in home trial which I wish ore stuff had. In my system they are FAR superior to the IsoAcoustics which actually made my sound worse.

As I showed in my video on the NA Tempus, I am currently using a combo of Well-float, Stack Audio and Graphite Audio on my switch with great success.

 

[Republicoftexas69]

I agree.

I've been experimenting with vibration control under my switches for about two years and have tried at least 20+++ products and combinations. .

I can do better than the Stack feet on my switches, but at a much higher price. Stack is reasonably priced and have an in home trial which I wish ore stuff had. In my system they are FAR superior to the IsoAcoustics which actually made my sound worse.

As I showed in my video on the NA Tempus, I am currently using a combo of Well-float, Stack Audio and Graphite Audio on my switch with great success.

I have my switches in a hyperbolic chamber suspended in mix of Helium and Nitrogen by aramid fibers. Temp is 69 degrees F

 

[MarkusBarkus]

hyperbolic chamber

...I think most of this website is in a "hyperbolic chamber" some days.

 

[NigelB]

I’m 95% certain that Stack commissioned John Westlake to engineer this piece. I can spot his layout technique from space.

[I’m guessing that part of why Stack keeps quiet about it is that Westlake, while a sharp engineer, has a good bit of controversy following him over long-delayed and then abandoned altogether projects for which he collected a lot of money from users in advance. You can Google that.]

Is this a problem? I am massively familiar with the long-running controversy (and, by the way, amazed by the faith some of the investors STILL have that one day they will see product!), but this is a very different situation and I think it's unfair to associate Stack Audio's offerings (which actually exist!) with the man... if indeed he is the man behind their creation. Let's cut them some slack, as feedback based on the deployment of their products is overwhelmingly positive.