Air--ubiquitous, free and universally capable of carrying a multitude of languages, dialects and amplitudes-is certainly a great analog for the ultimate digital audio and video network. As a community, we would welcome this utopia because our jobs today, is to employ the less-than-perfect tools we have in our arsenal in an attempt to manage our complex networks in finite, artistic detail.
This oversimplified metaphor does not seek for one second to diminish the art of audio engineering or acoustic design but instead attempts to articulate the vast but simple requirements of audio networking technology and the advantages--if done right--that this much-hyped technology domain offers our profession.
With new networking technologies seemingly popping up at each and every tradeshow, the industry is currently fracturing. While designers of these new solutions claim theirs to be the ultimate, none seems to be gaining critical mass. Words like "open," "proprietary" and "standard" are being bandied about as our community goes through the "format war" for networked audio. Pro Sound News recently had the opportunity to sit down with the Harman Pro Group's Michael MacDonald and Rick Kreifeldt to learn about a seismic shift the company is claiming to be implementing in its approach to audio network technology. It is here that MacDonald draws upon the air analogy: "A successful open network needs to be instinctive, abundant and universally accessible to generate mass usage," he explains. "That's what we're seeking to provide!"
MacDonald, Harman Pro Group executive vice president of marketing & sales, is directing group-wide strategy; and Kreifeldt, as vice president, Harman System Development & Integration Group (SDIG), is the man charged with designing network strategy and then making it happen in the form of actual, usable technology.
"To be successful, a network technology requires intense market and development support in its early days," MacDonald explains. "We invested time and resources in HiQnet, and it set the benchmark for what a dedicated control network protocol could do. Now after four years of solid market traction and consistent development, we are ready to unveil the complementary technology that our SDIG advanced research group has been working on."
Under the direction of Robert Boatright, SDIG director of research, this team has been actively contributing to the IEEE 802.1 Audio/Video Bridging (AVB) Task Group. Together with semiconductor and network technology leaders from every market, the AVB Task Group is developing open standards for the guaranteed delivery of low-latency audio and video. MacDonald and Kreifeldt are quick to point out that to be viable in the long term, a network protocol needs to be truly open, economically and technologically unencumbered with license fees and non-standard hardware.
"It's counter-intuitive that one group would own 'the network' and that the global community would flock to it if it weren't universally open or economically rational," Kreifeldt adds. "Many companies and industry groups have developed excellent technologies but these two fatal flaws have resulted in limited node traction, economically limiting the potential of networked audio systems."
"The existing solutions naturally suffer from performance issues because they have to work on top of Ethernet and cannot bend the Ethernet rules to their own advantage," Kreifeldt continues. "The IEEE 802.1 AVB Task Group is blessed with the ability to enhance Ethernet so that media streaming takes priority. That, and the fact that it will be a gigabit solution, make the questions about latency and channel count a thing of the past."
Harman Pro Group's active participation in the AVB Task Group is not an indication it will abandon its existing and established user-base of Harman Pro CobraNet-equipped devices. Kreifeldt adds, "This move will not impact the current CobraNet customers, because we will certainly continue to support CobraNet for the foreseeable future. It's without doubt the most dominant audio networking technology in our industry today. And because the new technology is very much an evolution of Ethernet, it is fully backwards-compatible and so it and CobraNet can even coexist. It has been designed from concept so that old and new will run seamlessly together. And that would go for other Ethernet-based audio protocols such as EtherSound or Audinate."
Where the new technology does differ from Ethernet as we know it today is at the very heart of the Ethernet switch, which, in effect, polices the traffic and reserves priority bandwidth for audio and video data, up to 75 percent of the bandwidth. Any other information (web traffic, e-mail, etc.) is considered low priority and their packets of data are slotted in between those of the media streams. If you receive an e-mail a few micro-seconds later than you otherwise would have, you're not going to notice. Get a dropout in the audio or a stutter in the video and it's a whole different story. The real-world benefit of this is that one network could be employed for media and regular Ethernet traffic, whereas normal practice today would be to install two separate networks, raising the infrastructure costs considerably.
When connected together the new switches also ensure that an audio or video stream is protected all the way from one end of the AVB network to the other. But while protecting the media data from lower-priority data, the new switches also understand that once upon a time there was an inferior Ethernet technology. If the media data crosses outside the AVB walls it can, but it's on its own. Although this might sound a little bleak, it's actually a considerable benefit? First off, once outside the protection of the AVB network, the data is in no worse a position than it is today, but more importantly, it enables the new technology be used in the expansion of existing networks.
The AVB Task Group comprehensively addresses the challenge of synchronizing audio and video on a single network. The current accepted mechanism of clocking audio data is to use a beat packet, acting like an advanced metronome, which keeps all the data in sync, at 48 kHz, for example. But when an entire network is synced to a single "bpm" it presents significant limitations. "Video clocks just don't relate to audio clocks, so getting them to behave nicely together is very tricky," Kreifeldt says.
AVB actually makes use of a tried-and-tested industrial standard clocking mechanism deliberately designed for multiple clock rates, IEEE 1588. Picture the multiple elements of complex robotic system such as a car assembly line, for example, which all need to be synchronized with extreme precision, and you can quickly picture its heritage. It might sound complex, but IEEE 1588 is simply a mechanism for synchronizing all the devices on the network by "time of day." Each transmitting device will "packetize" its outgoing media data and mark each packet with its own time stamp. The receiving device reads the time stamps and plays each packet of data accordingly. As the packets within each stream are independently time stamped, the adoption of IEEE 1588 thereby enables any number of "clocks" to coexist on the network (in fact, each and every stream could have its own).
As chair of IEEE 1722, the Layer 2 AVB Transport Protocol, Boatright has led the initiative for media interoperability. "A key focus for the IEEE has been to support a wide variety of audio and video formats, raw and compressed. The fact that AVB networks can run multiple, unrelated sample rates and native audio and video formats on the same network presents audio professionals in commercial sound, cinema, nightclubs, houses of worship, performing arts, stadiums/arena, recording, broadcast, tour sound and portable PA with significant performance, integration and economic advantages."
It is within this new network dynamic that an overriding communications protocol like HiQnet comes into its own. Audio professionals previously only dreamt of the vast array of network diagnostics that can be built into the protocol and reported by a device to the front panel or a system monitoring software application like System Architect.
"Technologically, AVB stands on its own and is really the next generation of networking technology for our industry," MacDonald concludes. "But what is truly unique and especially compelling is the cost-per-node economies of scale that come from partnership with IEEE and the semiconductor companies, consumer electronics companies and PC manufacturers. This will finally make networked audio and video second nature to us all! "
While we may never be able to change the properties of the air we breathe, AVB certainly looks primed to change the networking world we inhabit. And as if that's not enough, the really good news? It's closer than you might think.