Audio Mixing Platform (AMP)

Please login or register.

Login with username, password and session length
Advanced search  

News:

Pages: [1] 2 3 ... 7

Author Topic: AMP Dante is here!  (Read 24749 times)

admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 956
    • View Profile
AMP Dante is here!
« on: May 18, 2016, 11:14:28 PM »

Hey Gang,

We've been working a bit lately on getting AMP working with Dante, and I'm very happy to report that this is an exciting new option!

We will have the AMP Dante version available, which will come with the Dante PCI card, which features a redundant connection.  This provides 128x128 channels, all accessible via the current AMP!

Using a Mackie DL32R equipped with a Dante card (which gives me 32 in x 16 outs), I am seeing a latency of 3.74ms!  This is using a gigabit switch between Dante hardware devices, and setting the Dante latency on both devices to 250us (which is the lowest the Mackie would do).

Using an Intel Core I3-4360 at 3.7GHz, running Windows 7, I just ran a test using:
- 96 Input channels with *everything* turned on and processing:
   Low-cut, 7 bands of EQ, Gate with side-chain EQ, De-esser (with side-chain), and compressor
   24 auxes, each turned on
- 25 output channels, each with *everything* turned on.

CPU was running solid and click-free at 83.7% DSP.

Running the same number of channels, but with 3 bands of EQ, a compressor, and just 8 aux sends (which is more realistic of the typical use of a mixer, as you have to remember that not every channel is going to be up and running all sorts of processing) yielded 33.9% DSP utilization.

Folks, this means high channel counts (well into the hundreds) are possible with very modestly priced hardware!

The above numbers were using the Dante PCI Express card, which appears as 128 ins and outs to AMP.

If anyone is interested in testing AMP Dante, please PM me.  It's not cheap, but it does seem to work rather well!

Bob
Logged

RBIngraham

  • Guest
Re: AMP Dante is here!
« Reply #1 on: May 19, 2016, 09:18:01 AM »

Just to check and make sure I'm reading this correctly you have a package available that would include an AMP license and the Dante PCIe card?

But any Dante I/O devices would be purchased elsewhere?

Looks interesting...
Logged

admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 956
    • View Profile
Re: AMP Dante is here!
« Reply #2 on: May 19, 2016, 11:23:08 AM »

Yes, we will have a package that will contain the Dante PCI Express card, as well as an AMP license.  Still working out the details of the final cost.

Bob
Logged

mattseymour

  • Guest
Re: AMP Dante is here!
« Reply #3 on: May 19, 2016, 12:44:29 PM »

Sweet. I'm already playing with Dante using the Rednet card. Works very well.
Logged

admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 956
    • View Profile
Re: AMP Dante is here!
« Reply #4 on: May 19, 2016, 10:18:34 PM »

Hey Matt,

Yes!  The Rednet card is similar, except it does not have the redundant port.  Ours will have the redundant connection.

I just did a test earlier, disconnected the Dante I/O.  As soon as I plugged the cable back in, it picked right up and started passing audio again.  Nice!

Bob
Logged

gilnavarro

  • Guest
Re: AMP Dante is here!
« Reply #5 on: May 19, 2016, 10:40:10 PM »

Does the Dante card take some load off of the computer?
Logged

RBIngraham

  • Guest
Re: AMP Dante is here!
« Reply #6 on: May 19, 2016, 11:01:36 PM »

Does the Dante card take some load off of the computer?

Unless Bob cooked up something special it doesn't take the load off the computer.  It essentially just creates an ASIO interface that can communicate with a Dante network.  Obviously there is something a bit more than just that going on or Dante Virtual Sound Card would be able to perform at the same level as the PCIe card does.  Of course Audinate doesn't make all that much money selling DVS licenses, it makes money selling manufacturer's chips that are used to make hardware and the licensing fees that go along with that.  So I wouldn't be at all surprised if they purposely limit DVS's latency performance in order to sell more hardware.  And for all the noise they made about Dante VIA it's pretty limited as well.  Certainly far short of what I was hoping for.

But like every other manufacturer they make the money on hardware not software, in this case it's Dante chips. 
Logged

gdougherty

  • Guest
Re: AMP Dante is here!
« Reply #7 on: May 20, 2016, 01:51:24 AM »

Does the Dante card take some load off of the computer?

Unless Bob cooked up something special it doesn't take the load off the computer.  It essentially just creates an ASIO interface that can communicate with a Dante network.  Obviously there is something a bit more than just that going on or Dante Virtual Sound Card would be able to perform at the same level as the PCIe card does.  Of course Audinate doesn't make all that much money selling DVS licenses, it makes money selling manufacturer's chips that are used to make hardware and the licensing fees that go along with that.  So I wouldn't be at all surprised if they purposely limit DVS's latency performance in order to sell more hardware.  And for all the noise they made about Dante VIA it's pretty limited as well.  Certainly far short of what I was hoping for.

But like every other manufacturer they make the money on hardware not software, in this case it's Dante chips.

Compared to the DVS, the PCIe card absolutely does take the load off the computer.  It's like server grade NICs that can offload encryption duties from the CPU to the network chip.  In this case the PCIe card handles the audio streaming to and from the network as well as converting that data into an ASIO compatible format for software to interface with.  Similar to how PCIe based interfaces perform better than their USB or Firewire brethren.  There's far less work for the CPU to do, especially with USB where a good chunk of the data transfer involves the CPU.
Logged

admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 956
    • View Profile
Re: AMP Dante is here!
« Reply #8 on: May 20, 2016, 02:02:26 AM »

Yes, so far that I can tell, there's virtually no CPU time eaten by the card or its drivers.

Just did another test, maxing the system out:

128 Input channels
32 Stereo Busses
32 Stereo Auxes

CPU without any processing was in the high 20s.  With all auxes turned on, signal routed to 8 of them, 4 bands of EQ on every channel, and compressor turned on, I'm running 55% CPU.  Remember folks, this is 128 mix busses.  This is higher than the big Avid mixers.  I'm currently using 200 slots in the mixer window.  It's blowing me away.

Bob
« Last Edit: May 20, 2016, 02:03:58 AM by admin »
Logged

mattseymour

  • Guest
Re: AMP Dante is here!
« Reply #9 on: May 20, 2016, 04:58:47 AM »

As an aside, I'm using the Presonus RM mixers as Dante stageboxes. Works out much cheaper than the Mackie, in the UK at least.

The only piece of the puzzle missing is the ability to control the preamp from AMP. Maybe one day AES70 will be implemented by enough manufacturers to make that a practical option.

At least there's proper remote control of preamps and phantom power. Much better than having to run 80 metres to the stagebox to tweak a hot channel.
Logged

RBIngraham

  • Guest
Re: AMP Dante is here!
« Reply #10 on: May 20, 2016, 10:13:46 AM »



Compared to the DVS, the PCIe card absolutely does take the load off the computer.  It's like server grade NICs that can offload encryption duties from the CPU to the network chip.  In this case the PCIe card handles the audio streaming to and from the network as well as converting that data into an ASIO compatible format for software to interface with.  Similar to how PCIe based interfaces perform better than their USB or Firewire brethren.  There's far less work for the CPU to do, especially with USB where a good chunk of the data transfer involves the CPU.

Good point.  I was specifically thinking more along the lines of actually processing audio data and I doubt the Dante card is helping with that.  But yeah I could see where it would help with the processing of actually pulling the data from the network and sending it to the network. 

PCIe vs Firewire or USB has much to do with the bandwidth of the bus and how that bus transfers data to and from the main system bus as it does with the fact that there is a card involved.  There is no reason that a built in NIC couldn't communicate with the main system bus just as quickly as the PCIe bus and to the best of my knowledge most built in NICs do sit on the PCIe or a similar speed bus, don't they?

Still, would it really surprise anyone that Audinate purposely hold back the performance possibilities of DVS or VIA in order to sell hardware chips?  Doesn't seem all that unrealistic to me.   ;)  And that's perfectly acceptable, it's their protocol and licensing to do what they want with.  It's the good and the bad of a proprietary protocol over an open one.
Logged

admin

  • Administrator
  • Hero Member
  • *****
  • Posts: 956
    • View Profile
Re: AMP Dante is here!
« Reply #11 on: May 20, 2016, 10:57:54 AM »

Richard: I don't think DVS is purposefully "held back" - it does work, but there is obviously a lot of stuff that has to happen behind the scenes, especially with clocking.  This is much easier to do quickly on the hardware level than with software that has to work on everyone's PC.  This is why DVS works, but with added latency, and I believe reduced channel count.

RME offloads a lot of the work onto the chip on their ADAT interface cards, and is why their cards offer probably the lowest latency on that regard.

Matt: Exactly what products are you using?  My goal is to add code to the host to talk to various hardware for that control of phantom and HA gain.

Also Matt: have you had a chance to measure the latency of your system?


Bob
Logged

RBIngraham

  • Guest
Re: AMP Dante is here!
« Reply #12 on: May 20, 2016, 11:43:08 AM »

Richard: I don't think DVS is purposefully "held back" - it does work, but there is obviously a lot of stuff that has to happen behind the scenes, especially with clocking.  This is much easier to do quickly on the hardware level than with software that has to work on everyone's PC.  This is why DVS works, but with added latency, and I believe reduced channel count.

RME offloads a lot of the work onto the chip on their ADAT interface cards, and is why their cards offer probably the lowest latency on that regard.


Fair enough.  I can see how that would make it possible to process the same data more quickly and there by reduce latency.

For what it's worth, MOTU's PCI and PCIe cards were faster (less latency) than anything I've worked with from RME.  Or any other manufacturer for that matter.  I trust RME to be more solid and reliable but MOTU did have some of the lowest latency and ASIO buffer sizes available for a while there.  But they seem to have thrown that away and gone the AVB route, which for the vast majority of users make sense.  I wish they would make a nice AVB PCIe card.  But they are so heavily Mac influenced that I would highly doubt they would make anything that you could only use on a PC.  (or you could only use on the Mac by buying a Thundbolt to PCIe card cage, etc..)

Just like the theatre sound community is often taking gear and using in ways it wasn't necessarily designed or intended for, the software live sound mixing world is doing the same and pressing things primarily meant for recording and playback into service as live mixing consoles.
Logged

RBIngraham

  • Guest
Re: AMP Dante is here!
« Reply #13 on: May 20, 2016, 12:16:47 PM »

So I did the math for on this side of the pond.  The Presonus is cheaper for 32x16 I/O for Dante.  But only by $100. 

Bob,  is the Mackie DL32R only controllable by an iPad?  That is what it seems like at first glance but maybe I just didn't dig deep enough into their web site?

If that is the case I would go with the Presonus over the Mackie because at least the Presonus can be controlled by an iPad, Mac or Windows PC.  So you wouldn't need to pay an Apple tax just for a device to allow you to set up the box.   :)    (Sorry I plan to avoid an iPad until I simply can not live without it.. which hasn't happened yet...  lol)

Of course this assumes the Presonus would have the same latency performance.  Which I wouldn't count on without someone doing some specific testing.

Logged

mattseymour

  • Guest
Re: AMP Dante is here!
« Reply #14 on: May 20, 2016, 03:58:05 PM »

I've got one of each of the Presonus Rm16ai and Rm32ai. I've yet to do a latency check. Will do that soon.

The flexibility of the Presonus is what I liked about it. Though I think the Mackie app is better, the Presonus UC app runs on Windows so it can sit on the same machine as the Amp gui.
Logged
Pages: [1] 2 3 ... 7