Author Topic: "A British Audiophile" compares audio interconnects  (Read 18692 times)

0 Members and 1 Guest are viewing this topic.

Offline mansaxel

  • Super Contributor
  • ***
  • Posts: 3554
  • Country: se
  • SA0XLR
    • My very static home page
Re: "A British Audiophile" compares audio interconnects
« Reply #100 on: August 30, 2021, 11:24:48 am »
I wonder what audiophools would think about professional digital audio streaming over IP, for example Dante.

It needs a properly configured network but it works like a charm. Now tell a live sound engineer to buy a 50 metre long
"special Ethernet cable" for his digital snake from Audioquest  :-DD :-DD :-DD

But of course audiophile grade routers do exist!

Hey Borja,

I know you are clever, but when you call Dante "working like a charm" I'd like to question that a bit.

Dante:

  • Will not traverse a router
  • Is limited to 64 channels
  • Will only be routed by a stinking Windows app, also on the same LAN
  • Is using PTPv1
  • Won't easily interoperate with AES67 or Ravenna
  • Is proprietary
  • And, I don't like it.

 :-DD :-DD :-DD

But, seriously. In my application (2000+ sound sources spread over a 2000km long country), Dante is a joke, a toy. Their attempt at world domination would have been spent better in coöperation with other industry forces to build an even better AES67.

Offline MrMobodies

  • Super Contributor
  • ***
  • Posts: 1906
  • Country: gb
Re: "A British Audiophile" compares speaker cables
« Reply #101 on: August 31, 2021, 03:20:55 am »
I bet he’s “sick of experts” too  |O

Hear that said a lot by morons.

EEVblog #29 - Audiophile Audiophoolery
Quote
04:31 anyone with half a brain at all knows that what is this gonna do when
you've got all that crap or crap quality mains cable running in your walls and
through the powerlines

Joke: Then I guess the morons don't have half a brain.

Maybe he'd be sick of this review:
youtube.com/watch?v=KzwAdwvy9l4
Audioquest Thunderbird Zero Speaker Cable Review: 72V DBS Legit Science or Snake Oil?

Quote
26:41 now look look what happens when i plug the audio quest cable in boom wow now
you see our noise floor went up here it's doing some weird stuff over here it
really went up here i don't think it's detectable because it's still at point zero
two percent here but who knows you know um who knows when you're cranking it up if
it can amplify that even more but look at these spikes so whatever kind of noise
was in the room

33:12 this cable does not live up to the marketing claims from my measurements
um it's difficult to work with in my opinion the connector breaks pretty easily
on it as you can see here just by sticking in the amplifier it broke off...

Pretty much sums it up for that Audioquest snake oil cable.

Quote
36:36 tell me if these educational videos are helpful... tell me if you fell victim to some bs markettiing that you now realize was nonsense.
I suppose that may make them sick too when they realize.
« Last Edit: August 31, 2021, 03:24:11 am by MrMobodies »
 
The following users thanked this post: xrunner

Offline borjam

  • Supporter
  • ****
  • Posts: 908
  • Country: es
  • EA2EKH
Re: "A British Audiophile" compares audio interconnects
« Reply #102 on: August 31, 2021, 06:12:42 am »
I know you are clever, but when you call Dante "working like a charm" I'd like to question that a bit.
I guess you have more experience with Dante than me, so I won't argue much. Anyway what I meant (poorly phrased
is that it works without creating any of the "audiophoolic" problems audiophools imagine ;)

Quote
Dante:

  • Will not traverse a router
  • Is limited to 64 channels
Hmm. I think they wanted to play it safe. Running on a single broadcast domain they can do a lot of magical autoconfiguration without
relying on poorly supported multicast stuff.

So, well, it's designed for a building or at most campus level application.

Also I think they haven't done a bad job making sure that it can work with many "enterprise" Ethernet switches and with a reasonable
configuration it can coexist on the same infrastructure.

Quote
  • Will only be routed by a stinking Windows app, also on the same LAN
What I have used is Macos based. I haven't tried Dante Domain Controller but I understand it's available on Macos as well? No Linux, though.

Or what do you mean with "routed"?

From what I understood they avoid a central "router" node per se, relying instead on a Dante Controller application that is more like an orchestrator for the capabilites in each Dante interface. That's not a bad decision.

Quote
  • Is using PTPv1
  • Won't easily interoperate with AES67 or Ravenna
I have used it just for mostly straightforward stuff, so no experience there. Anyway I insist, in the context of this thread the "works like a charm" means it doesn't create "audiophool" issue and I am sure audiophools that have attended live concerts in which Dante was used haven't noticed "imaginary digital issues".

Quote
  • Is proprietary
  • And, I don't like it.
I must agree with this. It should be an open standard instead of a proprietary solution. As for liking, I think they did a good
job for their intended usage (venue level networking, at most campus level).

Quote
But, seriously. In my application (2000+ sound sources spread over a 2000km long country), Dante is a joke, a toy. Their attempt at world domination would have been spent better in coöperation with other industry forces to build an even better AES67.

I think they played it safe taking into account the lack of networking experience in audio professionals and the awful chaos in some implementation aspects of common protocols.

So yes, I saw it's not a "WAN" protocol but still I was surprised at how easy it is to set up as long as you have decent networking knowledge.
 

Offline coppercone2

  • Super Contributor
  • ***
  • Posts: 9234
  • Country: us
  • $
Re: "A British Audiophile" compares audio interconnects
« Reply #103 on: August 31, 2021, 07:11:31 am »
boxed wine at radio shack?
 

Offline mansaxel

  • Super Contributor
  • ***
  • Posts: 3554
  • Country: se
  • SA0XLR
    • My very static home page
Re: "A British Audiophile" compares audio interconnects
« Reply #104 on: August 31, 2021, 01:03:29 pm »
I know you are clever, but when you call Dante "working like a charm" I'd like to question that a bit.
I guess you have more experience with Dante than me, so I won't argue much. Anyway what I meant (poorly phrased
is that it works without creating any of the "audiophoolic" problems audiophools imagine ;)


I unfortunately have experienced Dante, yes. And like all pro and semi-pro system solutions, it is pragmatic w.r.t. 'phoolery. That much is true.

Quote
Dante:
  • Will not traverse a router
  • Is limited to 64 channels
Hmm. I think they wanted to play it safe. Running on a single broadcast domain they can do a lot of magical autoconfiguration without
relying on poorly supported multicast stuff.

So, well, it's designed for a building or at most campus level application.

Also I think they haven't done a bad job making sure that it can work with many "enterprise" Ethernet switches and with a reasonable
configuration it can coexist on the same infrastructure.
...until you have Ravenna and/or AES67 and/or SMPTE2059-clocked SMPTE2110 video. They all use PTPv2 which is an either-or protocol with PTPv1. If you support PTPv2 on your network, the Dante nodes will lose sync, because they can't create their clocking island. We ended up downgrading to a much worse switch for those nodes where Dante can't be turned off.
Quote
  • Will only be routed by a stinking Windows app, also on the same LAN
What I have used is Macos based. I haven't tried Dante Domain Controller but I understand it's available on Macos as well? No Linux, though.

Or what do you mean with "routed"?

From what I understood they avoid a central "router" node per se, relying instead on a Dante Controller application that is more like an orchestrator for the capabilites in each Dante interface. That's not a bad decision.

That is correct. And it is the right decision to make.  I meant that the Dante Domain Controller that actually lets you set the crosspoints (i.e. decide which multicast addresses nodes are going to subscribe to) is a Windows app.

Quote
Quote
  • Is using PTPv1
  • Won't easily interoperate with AES67 or Ravenna
I have used it just for mostly straightforward stuff, so no experience there. Anyway I insist, in the context of this thread the "works like a charm" means it doesn't create "audiophool" issue and I am sure audiophools that have attended live concerts in which Dante was used haven't noticed "imaginary digital issues".

Quote
  • Is proprietary
  • And, I don't like it.
I must agree with this. It should be an open standard instead of a proprietary solution. As for liking, I think they did a good
job for their intended usage (venue level networking, at most campus level).

Quote
But, seriously. In my application (2000+ sound sources spread over a 2000km long country), Dante is a joke, a toy. Their attempt at world domination would have been spent better in coöperation with other industry forces to build an even better AES67.

I think they played it safe taking into account the lack of networking experience in audio professionals and the awful chaos in some implementation aspects of common protocols.

So yes, I saw it's not a "WAN" protocol but still I was surprised at how easy it is to set up as long as you have decent networking knowledge.


All your points are valid. It's a fantastic 1:1 replacement for an analog multicore cable and a small matrix router. The problem is that's there no path upwards. You can't grow, and you can't control things beyond the peephole they give you. 

Also, you're completely correct that they wanted an easy sell for a complex problem. Problem is that this does not always work.. Biggest issue is the discovery of nodes and sources. Noone has solved this correctly in the "complete products" area; there's only skeletal implementations of properly designed node discovery systems, like the BBC efforts with NMOS IS-04

But no, Audiophools they're not!

Offline borjam

  • Supporter
  • ****
  • Posts: 908
  • Country: es
  • EA2EKH
Re: "A British Audiophile" compares audio interconnects
« Reply #105 on: August 31, 2021, 01:26:36 pm »
...until you have Ravenna and/or AES67 and/or SMPTE2059-clocked SMPTE2110 video. They all use PTPv2 which is an either-or protocol with PTPv1. If you support PTPv2 on your network, the Dante nodes will lose sync, because they can't create their clocking island. We ended up downgrading to a much worse switch for those nodes where Dante can't be turned off.
Out of my "pay grade". I do live jazz and I can work with a 10 channel mixing desk, go figure :)

That is correct. And it is the right decision to make.  I meant that the Dante Domain Controller that actually lets you set the crosspoints (i.e. decide which multicast addresses nodes are going to subscribe to) is a Windows app.
Ah! I saw they offer it as a virtual machine appliance and I imagined it would carry some Linux system as the OS.


Quote
All your points are valid. It's a fantastic 1:1 replacement for an analog multicore cable and a small matrix router. The problem is that's there no path upwards. You can't grow, and you can't control things beyond the peephole they give you. 

Also, you're completely correct that they wanted an easy sell for a complex problem. Problem is that this does not always work.. Biggest issue is the discovery of nodes and sources. Noone has solved this correctly in the "complete products" area; there's only skeletal implementations of properly designed node discovery systems, like the BBC efforts with NMOS IS-04

But no, Audiophools they're not!
So, we basicly agree. Now, how about teaming up and creating 9th Circle fixing all the Dante problems using IPv6, eggs, bacon, blockchain, BGP, egs bacon, quantum cryptography, blockchain and  and fuzzy logic and blockchain blockchain instead of discrete bits?  :-DD :-DD I am sure audiophools would love the fuzzy thing and the blockchain would help them feel secure?


 :-DD
 

Offline mansaxel

  • Super Contributor
  • ***
  • Posts: 3554
  • Country: se
  • SA0XLR
    • My very static home page
Re: "A British Audiophile" compares audio interconnects
« Reply #106 on: August 31, 2021, 01:43:33 pm »
...until you have Ravenna and/or AES67 and/or SMPTE2059-clocked SMPTE2110 video. They all use PTPv2 which is an either-or protocol with PTPv1. If you support PTPv2 on your network, the Dante nodes will lose sync, because they can't create their clocking island. We ended up downgrading to a much worse switch for those nodes where Dante can't be turned off.
Out of my "pay grade". I do live jazz and I can work with a 10 channel mixing desk, go figure :)

That's my hobby!

http://vvv.besserwisser.org/Public/Bilder/Skafestplats/desk.jpeg

But no, Audiophools they're not!
So, we basicly agree. Now, how about teaming up and creating 9th Circle fixing all the Dante problems using IPv6, eggs, bacon, blockchain, BGP, egs bacon, quantum cryptography, blockchain and  and fuzzy logic and blockchain blockchain instead of discrete bits?  :-DD :-DD I am sure audiophools would love the fuzzy thing and the blockchain would help them feel secure?


 :-DD

I'm actually content at applying science where there's a need. I run a lab-grade clock system for work, with dual GNSS mainframe clocks, a distribution system and a monitoring system (clock quality, et c.) that's something extra. That's what it takes to build good synchronisation for audio and video..

Online madires

  • Super Contributor
  • ***
  • Posts: 7694
  • Country: de
  • A qualified hobbyist ;)
Re: "A British Audiophile" compares audio interconnects
« Reply #107 on: August 31, 2021, 01:51:26 pm »
Hmm. I think they wanted to play it safe. Running on a single broadcast domain they can do a lot of magical autoconfiguration without
relying on poorly supported multicast stuff.

So, well, it's designed for a building or at most campus level application.

Also I think they haven't done a bad job making sure that it can work with many "enterprise" Ethernet switches and with a reasonable
configuration it can coexist on the same infrastructure.
...until you have Ravenna and/or AES67 and/or SMPTE2059-clocked SMPTE2110 video. They all use PTPv2 which is an either-or protocol with PTPv1. If you support PTPv2 on your network, the Dante nodes will lose sync, because they can't create their clocking island. We ended up downgrading to a much worse switch for those nodes where Dante can't be turned off.

I don't know anything about Dante, but there are two layers of QoS you can tweak to match your requirements and make things work. Ethernet and IP have each their own QoS functionality which can be combined. Additionally, professional switches and routers support several methods of queueing and queue configuration to map QoS classes to specific traffic handling while also supporting filters to classify specific traffic. If you need to build broadcast domain islands you can use VLANs.
 

Offline mansaxel

  • Super Contributor
  • ***
  • Posts: 3554
  • Country: se
  • SA0XLR
    • My very static home page
Re: "A British Audiophile" compares audio interconnects
« Reply #108 on: August 31, 2021, 09:27:45 pm »

I don't know anything about Dante, but there are two layers of QoS you can tweak to match your requirements and make things work. Ethernet and IP have each their own QoS functionality which can be combined. Additionally, professional switches and routers support several methods of queueing and queue configuration to map QoS classes to specific traffic handling while also supporting filters to classify specific traffic. If you need to build broadcast domain islands you can use VLANs.

At times I dream of the blissful times before I knew about the intricacies of IP QOS and Ethernet COS. And queuing in switches.

Anyway, it is moot, in the Dante vs Ravenna case. To get good enough quality of the PTPv2 clock (all nodes on the net synchronised within 1µs ) for the real stuff, you need PTP-aware switches; most implementations that are useful are acting as boundary clocks, where the switch becomes an intermediate master clock to its children.

These are not cheap devices. Our most common one, for audio, is the Cisco Nexus 9348, which costs about 7000€, list, for 48 ports GE, 4 ports 10GE, and 2 ports 100GE.

Now, the implementation (and this is common to all switches we've tried, not to bash Cisco) is such, that when you tell the switch to start doing PTPv2 magic, it also stops forwarding PTPv1. Yeah, it sucks. I blame IEEE. This means that if I want Dante and AES67/Ravenna on the same network infrastructure (and it's a very strong requirement that we follow Metcalfe's Law and don't build one net per application), I will have to use minimum 2 switches. The one for Dante can be PTP unaware, but since I then must make extra effort to put the clocking packets first in the interface queues on that switch, I can't take just any crapola switch; it needs queues and it needs means to fill them with the right packets. So, I now have 2 rack units occupied, 2 switches to manage, even if I only need a handful of ports.

Offline Jr460

  • Regular Contributor
  • *
  • Posts: 142
Re: "A British Audiophile" compares audio interconnects
« Reply #109 on: August 31, 2021, 09:54:54 pm »
Late to this thread.

Let me say I use Dante in semi-pro way every weekend for years.   I have friends that work large venues that have the whole place setup with Dante.   Never had an issue with it ever, it just works.

It doesn't require $7k switches.   I use a pair of lower end Cisco 8 port switches.   Read some of the doc, the switches does have to do PTP, what they do want suggest is a switch with low gitter.  They gave a list of switches from Cisco to stay away from.   I trunk two VLANs between the switches, one has Dante only on it the other has DMX/Artbet and other low bandwidth data.

The switch does have to run or understand PTP, just be predictable in the forwarding of packets.   What you do need to setup onto switch is QOS so that PTP gets priority.   Plenty of docs, show the exact settings to use.


I do question two things stated.   Limited to 64 channels, NO.
Needing some windows only app, NO.   Most times nothing else is needed.   If I want to pull a live recording, then the controller running on MacOs  and also the Dante Virtual sound card driver, that looks like an ASIO driver to any recording app you happen to have.   Controller allows me to map the streams to channels on the virtual sound card.

Mixing and matching withe other new protocols.....   No need.
 

Online madires

  • Super Contributor
  • ***
  • Posts: 7694
  • Country: de
  • A qualified hobbyist ;)
Re: "A British Audiophile" compares audio interconnects
« Reply #110 on: September 01, 2021, 10:45:24 am »
Now, the implementation (and this is common to all switches we've tried, not to bash Cisco) is such, that when you tell the switch to start doing PTPv2 magic, it also stops forwarding PTPv1. Yeah, it sucks. I blame IEEE.

Have you tried switches supporting the configuration of the PTP version per port (e.g. OnTime networks)?
« Last Edit: September 01, 2021, 08:33:47 pm by madires »
 

Offline mansaxel

  • Super Contributor
  • ***
  • Posts: 3554
  • Country: se
  • SA0XLR
    • My very static home page
Re: "A British Audiophile" compares audio interconnects
« Reply #111 on: September 01, 2021, 07:56:36 pm »
Now, the implementation (and this is common to all switches we've tried, not to bash Cisco) is such, that when you tell the switch to start doing PTPv2 magic, it also stops forwarding PTPv1. Yeah, it sucks. I blame IEEE.

Have you tried switches supporting the configuration of the PTP version per port (e.g. OneTime networks)?

No, and having looked at the CM-1600-FC4 from OnTime (I'm assuming that "onetime" was misspelled based on googling for both strings) I see that it would help with protocol translation, but not with the carrying of media traffic, and then it's out. I'm currently running something like 250 ports Ethernet for audio media traffic locally, with some 200 more for remotes, and I need more, to give an impression of the scale we're operating at.

Offline mansaxel

  • Super Contributor
  • ***
  • Posts: 3554
  • Country: se
  • SA0XLR
    • My very static home page
Re: "A British Audiophile" compares audio interconnects
« Reply #112 on: September 01, 2021, 08:01:37 pm »
Late to this thread.

Let me say I use Dante in semi-pro way every weekend for years.   I have friends that work large venues that have the whole place setup with Dante.   Never had an issue with it ever, it just works.

It doesn't require $7k switches.   I use a pair of lower end Cisco 8 port switches.   Read some of the doc, the switches does have to do PTP, what they do want suggest is a switch with low gitter.  They gave a list of switches from Cisco to stay away from.   I trunk two VLANs between the switches, one has Dante only on it the other has DMX/Artbet and other low bandwidth data.

The switch does have to run or understand PTP, just be predictable in the forwarding of packets.   What you do need to setup onto switch is QOS so that PTP gets priority.   Plenty of docs, show the exact settings to use.
You can do it without PTP support, that is the whole point. All you need to do is make sure you have enough bandwidth and some queuing in place.  It is a splendid alternative to an analog snake, which is what people buy it for. It does not work as well for my scale of things, where I have, as stated, perhaps 2000 sources active over a 2000 kilometer long network.

I do question two things stated.   Limited to 64 channels, NO.
Needing some windows only app, NO.   Most times nothing else is needed.   If I want to pull a live recording, then the controller running on MacOs  and also the Dante Virtual sound card driver, that looks like an ASIO driver to any recording app you happen to have.   Controller allows me to map the streams to channels on the virtual sound card.
I'm happy to be corrected here.

Mixing and matching withe other new protocols.....   No need.

Lucky you!

Online tszaboo

  • Super Contributor
  • ***
  • Posts: 7305
  • Country: nl
  • Current job: ATEX product design
Re: "A British Audiophile" compares audio interconnects
« Reply #113 on: September 02, 2021, 11:08:47 pm »
Anyway, what I'd love to know from people smarter than I am is this... He streams all his music to his overpriced premium sound system via a Bluetooth DAC. Does Bluetooth recompress audio and result in degradation of the original source material?
Depends on the BT standard and Codec used.
Some Codecs do not compress, they just use the available bit-rate and drop whatever would push past the limit.
Wireless streaming is always lossy.

Important distinction:
Streaming - Information is NOT stored permanently on the receiving device
Transfer - Information IS permanently stored on the receiving device
Not for long:
https://www.qualcomm.com/news/releases/2021/09/01/qualcomm-adds-bluetooth-lossless-audio-technology-snapdragon-sound
 

Offline mansaxel

  • Super Contributor
  • ***
  • Posts: 3554
  • Country: se
  • SA0XLR
    • My very static home page
Re: "A British Audiophile" compares audio interconnects
« Reply #114 on: September 03, 2021, 11:50:18 am »

Not for long:
https://www.qualcomm.com/news/releases/2021/09/01/qualcomm-adds-bluetooth-lossless-audio-technology-snapdragon-sound

As long as the path from storage to D/A isn't bit-clean (just shuffling octets) there will be recoding from one lossless format into another, and several steps of it at that. Recoding is identified as one major cause of signal degradation in sound transfer systems. It can be reduced in impact by starting out with less reduced bitrates, but is still audible.


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf