AVB and Audinate's Dante: An Update After Infocomm 2015
Update June 30, 2019: Latest thoughts here.
Update June 25 2018: Latest thoughts here
Update June 25, 2016: I have a followup here.
In recent years, Infocomm has become the go-to show for the convergence of Audio, Video, Networking, and Control (all the areas I follow most closely), and this year was yet another great show. My primary focus in my audio work and teaching is live sound, and walking the audio areas of the floor the hottest topic of conversation amongst my live sound-focused friends was Audinate's proprietary Dante audio network transmission system, and of course IEEE's open Audio Video Bridging (AVB) standard.
Work on AVB started back in 2005, and I first saw a demo of it at Infocomm 2009 (my initial writeup with an introduction to the technology here). It seemed like a great idea, and I was a big proponent of it. However, by 2013 I was concerned that the AVB development was dragging behind the needs of the live sound market, and after Infocomm in June of that year I was thinking that it might succeed if it could be integrated with Dante, who by then was well established with major players in our market like Yamaha and Shure. Later that summer, I visited the AVnu testing lab in New Hampshire, and was still somewhat hopeful about AVB's prospects, but after AES in November of 2013, after the release of the AES-67 inter-operability standard, I peered into my crystal ball and wrote, "Audinate's Dante has won the race, while AVB is still being tweaked by the mechanics over in the pit lane. ". That was further reinforced by what I my visit to Infocomm last year, and even more so this year. Here's the products on the AVB/AVnu Alliance stand:
And somewhat telling in that exhibit was the Yamaha console:
That Pivitec expansion card apparently gets 16 channels of AVB in and out of the console (this 16 channel limitation, of course, is due to Yamaha's ancient MY card architecture--not AVB), but notice that the Dante support for a virtually unlimited number of channels is built in. Unless you wanted to interface with a Pivitec monitor system, which apparently uses AVB, why would you buy this expansion card? And I write "apparently" here because looking at the Pivitec product page for that card, it doesn't even mention AVB:
Dante was in so many places on the trade show floor that Audinate was running a scavenger hunt promotion and were giving away T-shirts to people who could find a certain number of products. You can see the Dante booth in the photo at the top of this post, and here's just some of the several racks of Dante gear available today they had on display.
While there is still one big player in our live sound market--Meyer--holding onto AVB, it seems almost all the other big live sound market players have jumped onto the Dante bandwagon. And this includes Avid, with their upcoming S6-L mixer; this is particularly interesting development since Avid was one of the few companies selling an AVB-based mixer system. [Update July 9--Avid's system is based on AVB but they now have a Dante interface card, see comment below]. And it seems that Harman (JBL, etc) is heading towards Dante as well.
At the show, I asked an old friend at Meyer who would know the company strategy about connecting to Dante, and he wouldn't say anything. But it seems to me they have to come up with a way to at least interface with it. Why? Back in May I co-designed and aligned a large, complex all-Meyer speaker sound system for the Youtube Brandcast at the Madison Square Garden theater.
This is how we got signal out of our Yamaha CL-5 FOH and Monitor consoles (black Dante I/O boxes) and into Meyer's Galileo output signal processing system (blue boxes):
Fortunately this system was buzz-free and clean due to the excellence of our crew, but this mess could be replaced with a few Cat 5 cables if Meyer added support for Dante to their Galileo product line (to be fair, those Galileo boxes don't even yet have AVB--only analog and AES). Don't get me wrong, I'm a huge Meyer fan, but when everyone else is using something different, and your users need to connect, we can't stay analog forever. And its just not likely the rest of the world will go Meyer's way and suddenly switch to AVB tomorrow.
At the show, I also saw that some other live sound stalwarts are also now supporting Dante. Digico have long been basing systems on MADI, and at the show they had an Dante bridge (sorry for the photo quality):
SSL also was showing something similar:
And I like the redundancy--Focusrite also now has its Rednet Dante I/O boxes with redundant power supplies.
And I didn't even mention Dante's Via, which Audinate was pushing at the show and introducing soon. It's pretty cool and has a lot of theatre sound-style applications.
So where are we today? I no longer need my crystal ball--Dante is currently the defacto audio networking solution for the live sound industry. And don't take my word for it, take some actual data from an excellent white paper from Roland Hemming's consulting firm (and funded by Audinate). Mr. Hemming kicked off Audinate's AV Networking World sessions with a very bold slide:
Again, sorry for the terrible camera phone quality, but what the slide says is:
Protocols are products
Protocols are ecosystems
Protocols should not be standards
Protocols should use standards as building blocks to allow interoperability
Hmm, "protocols should not be standards" That one got me, because I've long advocated for open standards. But I think he's right at least in the case of audio networking. Dante is doing--and doing well--everything we need today, it's affordable, and as flexible and inter-operable as Audinate wants it to be (and it seems they want to connect it to anything that makes sense). And they were able to get into that position because they were able to be agile and innovative (and no one who knows anything about the standards-making process would use either of those words). Mr. Hemming went onto make a substantive case as to why Dante has come to this position--and much more thoroughly (and with real data) made the case I've been making here on the blog for the last six years.
So where does that leave AVB? I talked to a guy from Extreme Networks at the show, and it seems they are still the only company making AVnu certified switches, and the ability to carry AVB is still an option that costs about $400 per switch. Biamp is often mentioned as an AVB success story, and they had an enormous booth on the floor, but don't really have much presence in my world. And while I heard from a few people at the show "more cool AVB stuff is on the way real soon", I'm afraid I've heard that too long--I just can't even imagine what the product or solution would be that would make everyone drop Dante and go to AVB. For example, we've been using Dante for distribution of audio on our Gravesend Inn haunted hotel for four years successfully. Why would we throw out tens of thousands of dollars of stuff to switch to AVB now?
But AVB still has a potential (and fascinating) future: As I wrote back in January of this year, It seems that AVB and the AVnu alliance has shifted to offer a solution for Time Sensitive Networking (TSN), and this was reinforced on the show floor with an AVB/TSN graphic, which listed only a handful of live sound companies and Audinate itself, who list some of same companies as licencees:
Also the automotive market is apparently very interested in AVB/TSN as well--using special connectors and AVB they can use Ethernet to save lots of weight in cars, which is a big deal.
And so I'll be interested to see what the next year brings, but I'm willing to bet anyone a significant donation to a charity of their choice that Dante will be available in more live audio products on the floor of Infocomm 2016 than any other audio networking solution (beyond that, who knows). And in the end, as I've been saying for a while (and Mr. Hemming substantiates in his paper, which you should read), that's all users who actually buy this stuff really care about. Anyone want to take that bet?
Update July 25:
After the comments closed, Kevin Gross, key developer of Cobranet and now a leader in AES-67 development, wrote in with this:
Although Dante clearly has critical mass in many classes of audio products, the same could have been said about CobraNet 10 or so years ago. Proprietary technology is dependent on its owners to sustain adoption (against the headwind of the license fees they must charge), for maintenance and support and to keep the technology current with respect to market and technological advancement. Over the long term, it is simply not safe to assume that an individual technology owner will be able to negotiate the business and technical changes required to sustain their creation. Although Dante is the current leader in this ongoing horserace, realistically, based on recent history, we have to expect there to be a new contender before the end of the decade.
While a succession of proprietary technologies may move AV technology forward, technology based on open standards offers a more robust alternative. Standards already exist for audio networking in VoIP, ACIP, AVB and AES67. Standards exist and are under development for professional video over IP. Audinate has recognized the importance of standards-based interoperability by implementing AES67. Based on how things play out in IT, it is hard to imagine a future where the means of network audio interconnect is not open standards based. The only question is how long it will take to get there.
Kevin Gross - AVA Networks