Avid Pro Audio Community

Avid Pro Audio Community

How to Join & Post  •  Community Terms of Use  •  Help Us Help You

Knowledge Base Search  •  Community Search  •  Learn & Support


Avid Home Page

Go Back   Avid Pro Audio Community > Pro Tools Hardware > Pro Tools | Carbon

Reply
 
Thread Tools Search this Thread Display Modes
  #1  
Old 01-02-2022, 06:38 PM
Phil O'Keefe Phil O'Keefe is offline
Member
 
Join Date: Jan 2002
Location: Southern CA USA
Posts: 2,921
Default Carbon expansion speculation

I know that Avid won't share their "roadmap" and what they have planned for the future - that's totally understandable IMO - but it doesn't mean we can't speculate, right? What would you like to see Avid release in the future in terms of Carbon-related expansions? What do you think they're likely to release eventually?

I have zero insights or behind the scenes information - this is purely speculative on my part.

The "SUR" on the front panel seems to suggest that they'll eventually make Carbon surround sound capable in terms of monitoring. No big surprise there - the clue is right in front of us. I'm sure that would be welcomed by some users.

Similarly, with that second RJ45 jack on the back, I wouldn't be surprised if they eventually also allow two Carbons to be used on a single system to provide increased I/O and DSP. There has to be some sort of reason that they included that second port on the back. What do you think they have planned for it?

With four headphone jacks, I kind of doubt we're going to see any sort of network expansion boxes for headphones or anything like that, but I suppose that is a possibility that might appeal to some, although I suspect only a few, Carbon users.

To me, Carbon has a lot of similarities with some of the UA products. If they were not an inspiration for Carbon, I'd say that UA is definitely at least a direct competitor. I'm not interested in brand wars - I personally own a UA Octo Satellite Thunderbolt (which I really love), and I've loved and used PT for two decades (Carbon is largely replacing my old HD3 Accel rig), so IMHO there's plenty of room for both in the world. I'm not trying to stir up any controversy. But it wouldn't surprise me if Avid eventually released an AVB DSP expander for Carbon to compete with some of the UA DSP expansion options. If they offered an 8 DSP AVB expansion unit for Carbon that cost $2k or less, I'd be all over it. I don't really need more I/O than 24 (even 16 is usually sufficient 90% of the time for me lately), but I have come close to maxing out the Carbon DSP for larger tracking sessions.

What kinds of expansions and add-ons would you like to see Avid release for Carbon?
__________________
Phil O'Keefe

PT 2023.6 Ultimate (Perpetual) | Avid Carbon | M1 Max Mac Studio; 32 GB RAM / 1 TB SSD, macOS 13.4.1 Ventura.

PT 2023.6 Studio (Perpetual) | M1 MacBook Air; 16 GB RAM / 1 TB SSD, macOS 13.4.1 Ventura.
Reply With Quote
  #2  
Old 01-02-2022, 06:56 PM
Darryl Ramm Darryl Ramm is offline
Member
 
Join Date: Nov 2010
Location: USA
Posts: 19,510
Default Re: Carbon expansion speculation

The second Ethernet is either for future expansion or Ethernet redundancy. And it's not clear why an "expansion port" needs to be on the box--it might, but I'd sure hope that any expansion is supported between boxes connected anywhere over AVB.
Reply With Quote
  #3  
Old 01-12-2022, 02:30 PM
LukeHoward LukeHoward is offline
Member
 
Join Date: Mar 2001
Location: Melbourne
Posts: 1,233
Default Re: Carbon expansion speculation

Quote:
Originally Posted by Darryl Ramm View Post
The second Ethernet is either for future expansion or Ethernet redundancy. And it's not clear why an "expansion port" needs to be on the box--it might, but I'd sure hope that any expansion is supported between boxes connected anywhere over AVB.
I think the interesting technical question is whether they'll be able to maintain the same "in-the-box" latency across multiple devices. I don't think DSP expansion for its own sake (i.e. outside a low latency monitor path) has much value these days given how fast CPUs are.

In terms of your latter point, I suspect that's because AVB switches are still rare and, even if they're not, it's one less device to qualify. It took Avid a while to qualify AVB switches for the S6L and even then, I think it's only the Luminex one that is officially supported.

What would make it interesting to me is a soft upgrade that allowed it to also support Dante. (Or, an AVB card for the MTRX. I'll take either.)
__________________
https://lukehoward.com/
Reply With Quote
  #4  
Old 01-12-2022, 02:57 PM
Darryl Ramm Darryl Ramm is offline
Member
 
Join Date: Nov 2010
Location: USA
Posts: 19,510
Default Re: Carbon expansion speculation

Completely agree, it's all about that monitor latency and for some folks are tight on DSP resources already and I think the add ADAT IO for more channels but run out of DSP is will cause folks issues.

Not sure about implementing Dante... has anybody looked at what is on the Carbon motherboard? (post photos ). I have no idea if there is enough general purpose compute power in Carbon to run Dante Embedded. But with so many general purpose cores in modern FPGAs it would be great if this box has CPU to spare and it's in the network path.

It's a very interesting system, and the Avid AVB-Dante split-brain is kinda interesting. And AVB support on Windows and old-Macs just seems to limit adoption here. I'd have probably gone with an I/O daughter card with Thunderbolt, Dante and maybe AVB options (don't think Digilink is an option unless you want to disable DSP and have it emulater a dumb box). Stymying early adoption because of AVB compatibility just seems such a bad idea.

Last edited by Darryl Ramm; 01-12-2022 at 03:32 PM.
Reply With Quote
  #5  
Old 01-12-2022, 03:39 PM
LukeHoward LukeHoward is offline
Member
 
Join Date: Mar 2001
Location: Melbourne
Posts: 1,233
Default Re: Carbon expansion speculation

Quote:
Originally Posted by Darryl Ramm View Post
Completely agree, it's all about that monitor latency and for some folks that's tight already and I think the add ADAT IO for more channels but run out of DSP is will cause folks issues.
Right - DSPs could be confined to I/Os local to the system, but that falls down if you have a single monitor path because ultimately needs to be routed through the Carbon you monitor through. A hardware implementation of AVB with cooperating devices (remembering that the presentation time offset is set by the receiver, and defaults to 2ms), should be able to achieve sub-ms latencies: if Dante can do it, with the "overhead" of IP, surely AVB can too. (In quotes as I doubt it's much overhead for UDP implemented in hardware.)

(UAD is the obvious competitor, but as far as I understand Apollo DSP in console mode is also limited to local I/O. The difference is their DSP plugins are useful for mixing, because they're not available natively, i.e. the hardware is a dongle.)

Quote:
Not sure about implementing Dante... has anybody looked at what is on the Carbon motherboard? (post photos ). I have no idea if there is enough general purpose compute power in Carbon to run Dante Embedded. But with so many general purpose cores in modern FPGAs it would be great if this box has CPU to spare and it's in the network path.
There's a hi-res picture here. In addition to the TI DSPs, looks like there's a AM387x ARM CPU and a XC7S50 FPGA. I don't have the information or knowledge to speculate on there are enough spare cycles on either to run a Dante implementation :)

Quote:
It's a very interesting system, and the Avid AVB-Dante split-brain is kinda interesting. And AVB support on Windows and old-Macs just seems to limit adoption here. I'd have probably gone with an I/O daughter card with Thunderbolt, Dante and maybe AVB options (don't think Digilink is an option unless you want to disable DSP and have it emulater a dumb box). Stymying early adoption because of AVB compatibility just seems such a bad idea.
It is confusing. The most charitable interpretation is that AVB is plumbing within the vendor ecosystem (notwithstanding Milan), and Dante is about interoperability.
__________________
https://lukehoward.com/
Reply With Quote
  #6  
Old 01-12-2022, 04:51 PM
Darryl Ramm Darryl Ramm is offline
Member
 
Join Date: Nov 2010
Location: USA
Posts: 19,510
Default Re: Carbon expansion speculation

Thanks for the photo link. In this case the TI Sitara ARM Processor, a single core A-8 32-bit processor running at up to 1 GHz, has on-board dual Ethernet and that's what's driving the dual Microchip Ethernet PHYs in the photo. Sitara has hardware support for AVB, and I'm assuming the software networking stack is just fully running on that Sitara processor. With the Xilinx FPGA handling audio routing/switching. Be interesting to go poking around and find where the embedded OS console port is and see what it says is running and is doing. That SD-Card drive looks interesting, it will be the OS boot drive and other goodies. Anybody want to pop it out of their Carbon and see what is on it If it's running Linux then would be interesting to see what versions of network stuff is running to deliver AVB.

Those Sitara processors may get clocked down in many uses so would be interesting to see what frequency it's running at, not a powerhouse processor by any means but maybe able to run Dante embedded at large enough #IO here.

And for your Dante wishes... hopefully it's running Linux since Audinate has Dante on Linux on ARM embedded Dante IP available. I'm going to guess the Xilinx FPGA here is mostly audio routing/switching and ancillary staff. Audinate has Dante IP-cores for Xilinx FPGAs, but unlikely that helps here.

From an adoption of Carbon viewpoint and focus on the Carbon project studio market, I would guess that having Dante on Carbon is more useful than having AVB on MTRX. And also not clear that the way to do this is to have AVB and Dante out of separate ports, it might make more sense to switch the box to be an AVB box or a Dante box. May not want to support trying to use it as a AVB/Dante router, may not have the power to do that, and you may not want to make it easy for/encourage folks to mix AVB and Dante on the same physical network switches? Although I guess this is not super high number of channels/bandwidth. But dual link for redundancy or for expansion makes more sense.

I'm a big fan of RME products, and was interesting to see their MADI focus open up to AVB/Milan with more than a few public shots at Audinate/Dante for being proprietary and expensive to license.... and then of course RME come out with their Dante products, but yes more structured around the edge of their core MADI and AVB systems. Oh well I'm so horribly old-school I actually like MADI, nice antique 100Mbit FDDI technology from the 1990's.

Last edited by Darryl Ramm; 01-12-2022 at 05:04 PM.
Reply With Quote
  #7  
Old 01-12-2022, 05:04 PM
LukeHoward LukeHoward is offline
Member
 
Join Date: Mar 2001
Location: Melbourne
Posts: 1,233
Default Re: Carbon expansion speculation

Quote:
Originally Posted by Darryl Ramm View Post
Thanks for the photo link. In this case the TI Sitara ARM Processor, a single core A-8 32-bit processor running at up to 1 GHz, has on-board dual Ethernet and that's what's driving the dual Microchip Ethernet PHYs in the photo. Sitara has hardware support for AVB, and I'm assuming the software networking stack is just fully running on that Sitara processor. With the Xilinx FPGA handling audio routing/switching. Be interesting to go poking around and find where the embedded OS console port is and see what it says is running and is doing That SD-Card drive looks interesting, it will be the OS boot drive and other goodies. If it's running Linux then would be interesting to see what versions of network stuff is running to deliver AVB.
I poked around the S3 a while back. It appears to have a Linux kernel implementation of AVB (with hardware support for time-stamping). So it wouldn't surprise me if the Carbon is using this stack or an evolution of it.

Quote:
Those Sitara processors may get clocked down in many uses so would be interesting to see what frequency it's running at, not a powerhouse processor by any means but maybe able to run Dante embedded at large enough #IO here.

And for your Dante wishes... hopefully it's running Linux since Audinate has Dante on Linux on ARM embedded Dante IP available. I'm going to guess the Xilinx FPGA here is mostly audio routing/switching and ancillary staff. Audinate has Dante IP-cores for Xilinx FPGAs, but unlikely that helps here.
Interesting. Yes, Dante on Linux on ARM has a minimum latency of 2ms. Not too bad, but not as good as the FPGA IP core.

Quote:
From an adoption of Carbon viewpoint and focus on the Carbon project studio market, I would guess that having Dante on Carbon is more useful than having AVB on MTRX. And also not clear that the way to do this is not to do AVB and Dante out of separate ports, it might make more sense to switch the box to be an AVB box or a Dante box. May not want to use it as a AVB to/from Dante router, may not have the power to do that, and you may not want to make it easy for/encourage folks to mix AVB and Dante on the same physical network switches? Although I guess this is not super high number of channels/bandwidth.
Right, it's not encouraged to run AVB and Dante on the same switch, but in practice I haven't had a problem with it. (I run both on an Extreme X460-48p). Transparent Clock (TC) mode likely won't be supported with both PTP variants (Dante uses PTPv1, AVB gPTP). But probably not an issue in practice.
__________________
https://lukehoward.com/
Reply With Quote
  #8  
Old 01-12-2022, 07:52 PM
Darryl Ramm Darryl Ramm is offline
Member
 
Join Date: Nov 2010
Location: USA
Posts: 19,510
Default Re: Carbon expansion speculation

Quote:
Originally Posted by LukeHoward View Post
I poked around the S3 a while back. It appears to have a Linux kernel implementation of AVB (with hardware support for time-stamping). So it wouldn't surprise me if the Carbon is using this stack or an evolution of it.
Out of interest which NIC are they using there. (sadly?) I understand parts of PTP, having spend time with folks developing some really amazing distributed clock technology at Stanford University that goes beyond PTP but does not need switch support (but does leverage the NIC PTP hardware timestamps), and not meant to replace PTP either, maybe more an interesting microscope to look at distributed timing and to deliver ~ns precision network time in demanding corner cases. (and unlike basic NTP and PTP stuff I sure *don't* understand Dante and AVB at any depth).
Reply With Quote
  #9  
Old 01-12-2022, 08:17 PM
LukeHoward LukeHoward is offline
Member
 
Join Date: Mar 2001
Location: Melbourne
Posts: 1,233
Default Re: Carbon expansion speculation

Quote:
Originally Posted by Darryl Ramm View Post
Out of interest which NIC are they using there. (sadly?) I understand parts of PTP, having spend time with folks developing some really amazing distributed clock technology at Stanford University that goes beyond PTP but does not need switch support (but does leverage the NIC PTP hardware timestamps), and not meant to replace PTP either, maybe more an interesting microscope to look at distributed timing and to deliver ~ns precision network time in demanding corner cases. (and unlike basic NTP and PTP stuff I sure *don't* understand Dante and AVB at any depth).
I'm not sure which NIC it uses as it appears only a virtual NIC is exposed to the non-realtime Linux kernel (it's using the Xenomai RT extensions).

NB, support for PTP in the switch itself is optional (well, maybe not for AVB but generally speaking). It does have some advantages though, at least in larger networks, but I imagine a poor implementation is worse than no implementation.
__________________
https://lukehoward.com/
Reply With Quote
  #10  
Old 01-12-2022, 08:25 PM
Darryl Ramm Darryl Ramm is offline
Member
 
Join Date: Nov 2010
Location: USA
Posts: 19,510
Default Re: Carbon expansion speculation

PTP switch support is optional and lets be polite and say varies widely in quality, including PTP failure under congestion. It's unsafe if this is being really relied on (e.g. in financial trading) to assume it works, it's got to be tested under load. Some folks seem to assume because its an IEEE standard and the switch vendors says they support PTP that it will really work
Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
One more for the speculation fans! drenkrom Pro Tools TDM Systems (Mac) 0 09-13-2010 01:54 PM
A speculation on 8.0 Smithcok 003, Mbox 2, Digi 002, original Mbox, Digi 001 (Mac) 16 11-07-2008 07:56 AM
A New Round of Speculation cfb25 003, Mbox 2, Digi 002, original Mbox, Digi 001 (Mac) 4 03-03-2007 10:15 AM
Speculation on the upcoming upgrade basschair 003, Mbox 2, Digi 002, original Mbox, Digi 001 (Win) 3 05-23-2002 04:48 PM


All times are GMT -7. The time now is 08:33 AM.


Powered by: vBulletin, Copyright ©2000 - 2008, Jelsoft Enterprises Limited. Forum Hosted By: URLJet.com