Snake oil

[davelodwig]

A thread for the best snake oil you find.

To start with audio grade network switches. Never mind the complete lack of understanding of how ethernet works consider that even better this guy recommends only mounting your switch on HiFi vibration isolation stands:

https://youtu.be/fQGrkavDXeE

Apparently you can hear the difference.

I must remember when I next do a low level design and bill of materials for a datacentre to include thirty HiFi isolation racks. I mean I wouldn't want to just rack those switches in normal datacentre racks imagine all the corrupted packets if I did that! As a datacentre network specialist I must hang my head in shame that I have missed this essential step in good data transmission for the last thirty years.

It's not even a very good switch, in fact they refer to it as a hub a couple of times in their documentation.

No management interface, no vlans, no anything useful. No SFP port so I can hook the fibre up to the workshop? no I have a 50 quid mikotik switch that can do that.

And a stupid bloody plug that doesn't work with anything else.

[Pickled Onion]

So the guy played an album (apparently streamed through his home router) with and without this switch. He recorded the result, put it on the audio track of a video, it gets compressed, sent across the internet through however many non-hifi switches, out of DAC in people's computers and they can *still* hear the difference?

[Kim]

So the guy played an album (apparently streamed through his home router) with and without this switch. He recorded the result, put it on the audio track of a video, it gets compressed, sent across the internet through however many non-hifi switches, out of DAC in people's computers and they can *still* hear the difference?

What else is he supposed to do?  Use an oscilloscope?

[ian]

I think the issue is getting it to play in sync on several speakers that are networked together.

That's surely just an exercise in clock synchronisation?  Buffer the data, and tell each player when to start...

That seems to be the way Sonos stuff works, there are about two seconds when you first hit play, which I assume the system building a buffer.

[Kim]

I think the issue is getting it to play in sync on several speakers that are networked together.

That's surely just an exercise in clock synchronisation?  Buffer the data, and tell each player when to start...

That seems to be the way Sonos stuff works, there are about two seconds when you first hit play, which I assume the system building a buffer.

It's the obvious solution for something like that, where they don't have to be synchronised with anything except themselves, and a couple of miliseconds here or there is a non-issue (NTP should easily get the clocks tighter than the time it takes for sound to travel between speakers, even over The Devil's Radio).

From experience^[1], you can get acceptable results from just sending a "Go!" to multiple sound players over TCP most of the time, which is the stupid way to do it.


_{[1] Our alerting system makes no attempt to synchronise playback of audio samples (they're stored locally on SD card) other than the network messages arriving at pretty much the same time.  Sometimes one module gets very out of sync (up to a second or so), but usually the delay between rooms just sounds like a bit of echo.  I was going to use a "play this at $time" strategy, but never felt the need.  It's Rod Hull telling you there's someone at the door, not hi-fi audio.}

[Ham]

What they need to do now, is create a different, better network protocol, it's the uncontrolled nature of Ethernet that leads to all these problems with clocking and, clearly, poor hifi quality.

Lets start off sending empty packets around, maybe we could call them "present boxes" and computers wanting to talk to each other could grab an empty present box, fill it and pop on the front the address of the computer it wanted to talk to. That computer could receive the present, change the status to received. Then, when the box came back, the sending computer would know it got there ok, it could empty the box and let someone else use it.


That idea may have a 90% packet drop rate, but I suspect (at least?) 10% success 

[Gattopardo]

So the guy played an album (apparently streamed through his home router) with and without this switch. He recorded the result, put it on the audio track of a video, it gets compressed, sent across the internet through however many non-hifi switches, out of DAC in people's computers and they can *still* hear the difference?

Yep.

It is magic.

(or maybe, it is the placebo effect)

[Gattopardo]

A thread for the best snake oil you find.

To start with audio grade network switches. Never mind the complete lack of understanding of how ethernet works consider that even better this guy recommends only mounting your switch on HiFi vibration isolation stands:

https://youtu.be/fQGrkavDXeE

Apparently you can hear the difference.

I must remember when I next do a low level design and bill of materials for a datacentre to include thirty HiFi isolation racks. I mean I wouldn't want to just rack those switches in normal datacentre racks imagine all the corrupted packets if I did that! As a datacentre network specialist I must hang my head in shame that I have missed this essential step in good data transmission for the last thirty years.

You and your ones and zero's   If there isn't a record player, a valve amp and some sort off BBC designed speaker cabinet and speakers set correctly how can if be real?

[davelodwig]

What they need to do now, is create a different, better network protocol, it's the uncontrolled nature of Ethernet that leads to all these problems with clocking and, clearly, poor hifi quality.

Lets start off sending empty packets around, maybe we could call them "present boxes" and computers wanting to talk to each other could grab an empty present box, fill it and pop on the front the address of the computer it wanted to talk to. That computer could receive the present, change the status to received. Then, when the box came back, the sending computer would know it got there ok, it could empty the box and let someone else use it.


That idea may have a 90% packet drop rate, but I suspect (at least?) 10% success 

Token Ring?

[Kim]

To be fair, there is a certain logic to using circuit-switched networks for audio...

[Canardly]

I often watch Z reviews on U tube of sundry audio equipment (be prepared for profanities) but one thing he does is sound demos of kit via no doubt expensive sound sampling equipment. How can this possibly represent reality given the interwebby and my low level but good, no longer made, Altec Lansing THX speakers or low/mid range Senny or Sivga headphones?

[Gattopardo]

Or as you get older your hearing decays....

Or is it wearing out?

[pcolbeck]

To be fair, there is a certain logic to using circuit-switched networks for audio...

With a wires speed non blocking switch and all end nodes connected to that switch that's kind of what you have anyway. The cross bar switch matrix in the switch connects all ports to all other ports the same as in an analogue circuit switched telephone network. Its when you have multiple switches or switches and routers that things start to get different.

[Davef]

Like the adverts on tv comparing the “definition” of hd and normal tv that are broadcast on normal non hd tv. I can see the difference which is a worry.


Sent from my iPad using Tapatalk

[Ham]

What they need to do now, is create a different, better network protocol, it's the uncontrolled nature of Ethernet that leads to all these problems with clocking and, clearly, poor hifi quality.

Lets start off sending empty packets around, maybe we could call them "present boxes" and computers wanting to talk to each other could grab an empty present box, fill it and pop on the front the address of the computer it wanted to talk to. That computer could receive the present, change the status to received. Then, when the box came back, the sending computer would know it got there ok, it could empty the box and let someone else use it.


That idea may have a 90% packet drop rate, but I suspect (at least?) 10% success 

Token Ring?

Present Box - it has a ring to it

[Beardy]

To be fair, there is a certain logic to using circuit-switched networks for audio...

With a wires speed non blocking switch and all end nodes connected to that switch that's kind of what you have anyway. The cross bar switch matrix in the switch connects all ports to all other ports the same as in an analogue circuit switched telephone network. Its when you have multiple switches or switches and routers that things start to get different.

Point of order m’laud. Telephone networks are not, and to the best of my knowledge, have never been fully interconnect networks. It’s far too expensive and with way too much redundancy to have been implemented. Traffic management and grade of service provision is a whole speciality in telecoms, even in these days of cheap as chips digital switching.

[pcolbeck]

To be fair, there is a certain logic to using circuit-switched networks for audio...

With a wires speed non blocking switch and all end nodes connected to that switch that's kind of what you have anyway. The cross bar switch matrix in the switch connects all ports to all other ports the same as in an analogue circuit switched telephone network. Its when you have multiple switches or switches and routers that things start to get different.

Point of order m’laud. Telephone networks are not, and to the best of my knowledge, have never been fully interconnect networks. It’s far too expensive and with way too much redundancy to have been implemented. Traffic management and grade of service provision is a whole speciality in telecoms, even in these days of cheap as chips digital switching.

This is true. Hence during an emergency if everyone tried to dial at once they couldn't. The connections that could be made were circuit switched though. These days of course a lot of it is digital even if the last mile to your house is analogue.

[Beardy]

To be fair, there is a certain logic to using circuit-switched networks for audio...

With a wires speed non blocking switch and all end nodes connected to that switch that's kind of what you have anyway. The cross bar switch matrix in the switch connects all ports to all other ports the same as in an analogue circuit switched telephone network. Its when you have multiple switches or switches and routers that things start to get different.

Point of order m’laud. Telephone networks are not, and to the best of my knowledge, have never been fully interconnect networks. It’s far too expensive and with way too much redundancy to have been implemented. Traffic management and grade of service provision is a whole speciality in telecoms, even in these days of cheap as chips digital switching.

This is true. Hence during an emergency if everyone tried to dial at once they couldn't. The connections that could be made were circuit switched though. These days of course a lot of it is digital even if the last mile to your house is analogue.

in the old days we had preference keys that could in times of emergency Disconnect all but the telephones of those that were deemed required. The network has gone through a number of upgrades since then, and although i would be surprised if such a system were still not available, I only have the vaguest idea of the structure or the PSTN and no knowledge of the finer management of it. And of course with the mobile networks geographically overlaying the wired network, such things as preference management will re quite a cooperative approach between the network operators.

[Ham]

Here you are, how to spell woooo https://audiobacon.net/2019/11/02/the-jcat-signature-lan-a-1000-ethernet-cable/

[Ashaman42]

Christ. I read the whole thing waiting for the punchline.

[Ham]

It's a good reminder of the old adage: If you open your mind too far your brains can fall out

[Polar Bear]

"I would perform measurements, but it's more efficient for me to listen, ... '

Bolx.  Measurements with properly calibrated kit will not be subjective.

[pcolbeck]

The thing is even if its EMF that's making the difference why not just swap to a using fibre connectors and eliminate the possibility of EMF completely?
Oh yeh then they couldn't sell $4000 cables and switches.

[TheLurker]

> ... it’ll change the sound of your music files stored on disk.
What the actual!?

OK.  I can imagine that a _really_ crap cable might result in a lot of error correction / retransmission requests which might, at a pinch, depending on the buffering or lack of in the playback sw result in stuttering and some loss of information on decoding the noughts and ones but if the noughts and ones are making it from one end of the wire to the other unjumbled then all will be panglossian. However I'm *very* intrigued as to what flavour of magic is being used that allows a cable to alter the information content of a file.

Pterry was right.  There *should* be a theory test before people are allowed to do the practical when it comes to breeding.

[quixoticgeek]

I really want to get a group of Audiophiles and do a blind test of a fancy multiple 10's of grand setup, with all the fancy Ethernet cabling etc...

And one with just every day normal cabling... See if any of them can actually tell the difference....

But must make sure that all systems use the audiophile version of memcpy()... the standard glibc one isn't good enough...

https://www.audioasylum.com/cgi/t.mpl?f=pcaudio&m=119979

J