One of the first articles that I wanted to write for the Redfish Technologies website was related to what types of digital audio delivery are currently available. Audio delivery is becoming a requirement, not just something that’s cool and nice to have. Why? Well in this modern world having the audio path delivered from end to end without the need to convert back to analogue makes it simpler, easier to manage and more importantly the tyranny of distance is removed.
Many of the leading digital recording vendors are just now “starting” to support digital audio sound cards and devices. However what’s the best technology to use? Below is an article provided by Dan Daley and published in ProAv on the 26th March 2010 that gives a good outline of what technologies exist and the pro and cons. The jury is still out, but CobraNet still is the most prolific technology out there, but is it the best? I really think that AVB will soon become the real deal.
From a justice implementation perspective, very little digital audio delivery methods have been implemented. Even with the multitude of CobraNet equipped mixers that have been installed in many of the courts worldwide. A high majority of these simply output the audio as an analogue feed in either 1, 2, 4 or 8 channels into a equivalent analogue capture card on the recording workstation. If you think about that its pretty daft all-round. Why would you go to all the trouble of using a digital mixer only to convert it back to an analogue output and then re encode it on the recording PC?
Last year I was part of a team at Evidence Technology that implemented two types of digital audio connections to FTR Gold 5.3 recording software. The implementation was conducted at Victorian Parliament and was a landmark installation on a number of front. CobraNet was used on two production recording servers. The backup being analogue. Additionally two committee rooms are currently trialling Audinates “Dante” solution, again with a analogue backup. What’s surprising is that this was the first true implementation of CobraNet with a digital recording vendor in Australia. Its the first time its been tried in a production environment, and its the 1st time that this vendor has support this type of technology. The implementation was a success and continues to be so.
So is this the beginning of a new age or is it just cool technology to try out? Read the article below and form your own view on the technologies on offer. My perspective is digital delivery of the audio is the next step for digital recording in the courts, staying in the analogue world is just old hat now.
Source: Pro AV MAGAZINE
Publication date: 2010-03-26
By Dan Daley
Just when you thought CobraNet was a no-brainer, new audio networking solutions enter the fray.
The world of networked AV is its own kind of Babel. CobraNet, developed by Peak Audio in the mid-1990s and the first successful digital audio networking protocol, had a shot at becoming an industry standard. But while its acceptance by the pro audio industry was fast, it wasn’t fast enough to outpace a host of other companies that thought they had an equally good chance at developing their own widely accepted digital audio networking solutions.
By the turn of the century there were nearly a dozen proprietary networking protocols on the market, each vying for broad acceptance by various sectors of audio, from live to installed sound. The implementation of some, by vertically integrated companies that could make their own mixers, loudspeakers, DSPs, and other audio products more intimately compatible with their particular protocol, helped cement the Balkanization of the nascent networking sector.
However, the diversity of protocols fulfills the same kind of need that diversity in the biosphere does. “They all have their strengths and weaknesses,” says John McMahon, currently executive director of digital products at Meyer Sound and who, when at LCS Audio (which was acquired by Meyer in 2005), was one of the first to employ CobraNet. “Some work very well for low latency on dedicated networks, others have higher latency but use lower bandwidth, which works well for existing IT infrastructure.”
Few would argue that networked solutions are the way to go for AV distribution. We look at four prominent protocols.
Pros: Supports multiple sample rates and latency modes in a standard Ethernet network design and coexists with other network applications.
Cons: New product introductions have slowed as manufacturers opt for more updated technologies. It doesn’t scale beyond local area networks (LANs) and utilizes 100-Mbps devices only, with a maximum 32 channels to any connection.
Suitable Applications: Convention centers, theme parks, airports, churches, digital snakes in touring sound, legislative applications, live theater.
CobraNet is the most established of the digital networking contenders. It’s been around since the mid-1990s and is the most widely used technology for audio over Ethernet. It delivers audio in standard Ethernet packets over 100 Mbps, switched Fast Ethernet on Cat-5 cable at a distance up to 100 meters, or up to 80 kilometers over fiber. It’s fully compliant with the Institute of Electrical and Electronics Engineers’ (IEEE) 802.3u specification and all similarly compliant switches.
The brains of CobraNet are now developed by chipmaker Cirrus Logic, which offers royalty-free versions that provide two, eight, or 16 full-duplex channels of audio input and output at configurable network sample rates of 48 and 96 kHz and sample sizes of 16, 20, or 24 bits. Control, monitoring, and management functions are provided by a high-speed parallel host processor interface or via Ethernet using industry standard Simple Network Management Protocol (SNMP).
There are versions of the CobraNet audio technology that add 32-bit signal processing, as well as interface modules that support up to 32 full-duplex channels of audio (though the latter requires a license).
Audio pros say CobraNet’s advantage is its ubiquity. “You can find more products with a CobraNet interface than any of the others,” says Kevin Gross, an independent systems consultant. “This means many choices as a system designer and a lot of knowledge and expertise out there. It also means that even if it is getting a bit long in the tooth, it is likely to outlive some of the newer technologies.”
Recently, a major $116 million, 407,500-square-foot expansion project at the Oregon Convention Center required an AV upgrade that could flex with the facility’s highly configurable wall systems and integrate with its existing infrastructure. Integrator Delta AV chose to create a CobraNet network over a LAN, incorporating audio, video, and control in any configuration. The audio was based on
Biamp’s Audia platform, which handles systemwide signal processing and routing, and is distributed via
CobraNet using QSC’s RAVE multichannel router.
The project began several years ago, but Jeff Overbo, project manager for Delta AV, says
CobraNet would have been the network of choice had it started last week, citing his firm’s familiarity with it and the network’s ability to scale easily. “We had experience and a good track record with it in complex applications,” he says. The audio in the convention center is limited to paging, announcements, background music, and other types of low-bandwidth content, which Overbo says fits CobraNet well. “It’s probably not the best choice for a high-quality music system because of latency,” he explains.
Pros: Low latency (0.3 milliseconds), high capacity (up to 512 bidirectional channels per interface), fault tolerance, scalable Layer-3 networking support, Wide Area Network (WAN) streaming option, standard Ethernet network design, and coexistence with other network applications.
Cons: Q-LAN is available only on QSC’s Q-Sys products and it supports just one sample rate (48 kHz). The company says multiple sample rates up to 386 kHz will be supported in future releases.
Suitable Applications: Its centralized, high-capacity nature makes Q-Sys networking most suitable for larger installations.
QSC’s Q-Sys provides audio routing, processing, control, and monitoring functionality. The primary elements are the Core and the I/O Frames. The Core is the centralized brain of the system and allows any input to be routed to any output without convoluted signal paths. Although Q-Sys uses a centralized structure, it retains the ability to physically locate the input and output connections near their sources and destinations using I/O Frames. Each I/O Frame can house up to four I/O cards enabling up to 16 channels of input and/or output in a single unit.
Q-Sys utilizes Q-LAN, a proprietary, standards-based, low-latency gigabit Ethernet network implementation that’s responsible for audio routing between devices and ensures that all signals are transmitted from source to destination in less than 0.3 milliseconds. Q-Sys combines Q-LAN with 32-bit floating point distribution and high-performance A/D and D/A converters resulting in very low latency of less than 2.5 milliseconds from any input to any output with up to seven network switch hops. In addition to low latency distribution, Q-Sys also supports long-haul IP streaming of audio over WANs as well as auto-discovery and configuration of end nodes.
Gross says Q-Sys’ Layer-3 support makes the system unique. “Higher layers imply greater interoperability, scalability, robustness and security,” he says. “The Internet is built using Layer-3 IP networking. Your office LAN is built using Layer-2 networking.”
On the other hand, he says, “There is no 100-Mbps option for Q-LAN. At first blush, this might seem to be a strike against Q-LAN until you realize that gigabit Ethernet uses the same Cat-5 cable you use for Fast [100-Mbps] Ethernet, and commercial-grade gigabit switches are now cost-comparable with 100-Mbps equipment. Even if you don’t appreciate the expanded channel capacity of a gigabit connection, you will certainly appreciate the improved latency.”
At AT&T Park in San Francisco, installer ProMedia implemented a Q-Sys system comprising two Q-Sys Core 3000s, 16 I/O Frames, eight mic/line input cards, two line output cards, 56 dataport cards. The company also integrated 102 existing QSC PowerLight and CX amplifiers. Demetrius Palavos, senior sales and design engineer for Promedia, says the company chose Q-Sys in large part because the facility’s existing QSC amplifiers could remain in place.
“Using Q-Sys let us simply replace the CM16 amplifier control devices with nodes on the Q-Sys network,” he says. “That made it a plug-and-play proposition.”
Palavos says performance of the Q-Sys network has met expectations, but emphasizes that the equipment that rides the network plays a significant role in the choice of platform. “If these had been Crown amps, we would,probably have gone with the BSS London [network],” he says. “You’re not going to look at just the network by itself.”
Pros: Supports standard IP over Ethernet network design, coexists with other network applications, and supports a virtual PC soundcard implementation. Dante runs on existing networks, is compatible with PCs and Macs, and offers a migration path to AVB.
Cons: Currently there’s a limited number and type of products on the market, though several new OEM products are expected to be launched in 2010.
Suitable Applications: Entertainment venues, houses of worship, sports venues, live theater and music installations.
Audinate’s Dante is the commercial outcome of work begun at an Australian national research laboratory. Dante uses standard IP messaging on 100-Mbps and 1-Gbps switched Ethernet networks, and allows audio, control data, and other data traffic such as e-mail to utilize the same network simultaneously. This means it can be implemented using an existing data infrastructure.
Multiple sample rates and bit depths can share the same network, and all clocks on the system are synchronized to a master clock independently of the audio data, allowing sample-accurate playback from different devices on the network. Dante uses standards similar to IEEE Ethernet Audio/Video Bridging (AVB) and is considered AVB-ready (see “Whither Ehternet AVB?” below).
Audinate’s software will automatically find Dante-enabled devices on the network. Dante can carry up to 32 channels of 96-kHz/24-bit audio or 96 channels of 44.1-kHz/16-bit audio on a 100-Mbps network with a latency of 1 millisecond; a gigabit link can carry at least 10 times as many channels.
“Dante solves all problems from the most latency-critical to high-channel-count, far-flung monstrosities, [and offers] plug-and-play features including automatic address assignment and routing configuration by name instead of by number,” says Gross. “But I don’t think we’ve seen enough Dante products and installations to know whether Audinate will deliver all of this.”
Lee Minich of Lab X Technologies, a Rochester, N.Y.-based engineering design firm that specializes in digital audio connectivity, likes the fact that the system offers a mechanism for discovering and connecting to Dante devices. “It also lets users affix ‘friendly’ names to devices rather than arbitrary alpha-numeric designations,” he says. However, he adds, “Per-channel royalties make Dante products more expensive, and its functionality is most similar to AVB.”
Audinate itself, however, is quick to point out the differences between its technology and AVB. “Dante’s features [automatic address assignment, routing configuration, name-based configuration management, plug-and-play service discovery; along with redundancy, control, and performance monitoring] are part of the overall solution and are not included in the AVB standard,” says Ashesh Doshi, director of marketing. “These features will build on AVB, which presently covers synchronization and [quality of service] resource reservation, plus transport in the near future.”
Scott Ferguson, systems integration specialist at Peavey, which was an early partner with Audinate, expresses the relationship between Dante and AVB more directly: “Dante will work with any standard network switch; AVB requires an AVB compliant switch, adding cost to an AV install.”
Integrator CTS Audio used Audinate’s Dante system at the Oakwood University Church in Huntsville, Ala., in a point-to-point mode rather than in a comprehensive network mode, using a pair of Dante MY16-AUD cards to connect 32 channels of the church’s Yamaha M7CL 48 FOH console to a digital recording system running on an Apple MacBook Pro laptop. The two cards, three Ethernet cables, and an eight-port network switch cost under $1,500, but economics wasn’t the only reason the company chose Dante.
“We’ve been bench-testing Dante for a while and we’ve found it to be more flexible than CobraNet or EtherSound,” says Brian Pearce, CTS Audio project manager, who also cited its “ridiculously low” latency, making it a good choice for music-intensive applications. But he also cautioned that integrators used to the more ubiquitous CobraNet will have to build their infrastructure slightly differently to take advantage of Dante’s redundancy features. “Unlike CobraNet, Dante does not hook both the primary and secondary cables to the same network switch,” he explains. Instead, Dante’s primary and secondary cables go to separate network switches.
Pros: Very low, fixed latency; daisy-chain topology; fast reconnection and synchronization along the network; single-channel granularity.
Cons: In general, it requires a dedicated network, though it’s compatible with Ethernet equipment that does not also carry control traffic and can be configured to run on a VLAN over a gigabit network. The ring topology for bi-directionality is problematic for large-scale fixed installs because it’s limited to 64 channel inputs (though there is no limit in the number of outputs).
Suitable Applications: EtherSound has been seen mainly in live applications, though it can be configured for fixed installations.
Digigram’s EtherSound protocol is Ethernet IEEE 802.3x-compliant and supports Layer 2 (physical) network peripherals using standard Cat-5 or Cat-6 cables, fiber optic links, switches, media converters, and other standard Ethernet components. It provides low-latency, bi-directional transmission of synchronized audio channels and control data, with end-to-end transmission time of about six samples (125 microseconds at 48 kHz). It can transport 64 channels of 24-bit/48-kHz PCM audio, plus embedded control and monitoring data, via a single cable. Depending on the sampling frequency, other channel counts are possible, e.g. 32 channels at 96 kHz. Latency is independent of the number of channels transmitted and built-in clock recovery ensures low jitter.
EtherSound offers the option–though in some cases it’s a requirement–to wire the network in a daisy-chain configuration. “Some people consider this an advantage that makes for easy wiring racks or speaker clusters,” says Gross. “But it is not necessarily a good way to wire a fault-tolerant network.”
EtherSound has a much larger footprint in Europe than in North America. The Cité Internationale, in Lyon, France, contains a mix of museums, office space (including the European headquarters of Interpol), restaurants, legislative facilities, and a 3,000-seat auditorium. A sizable IP network was installed on a single-mode fiber backbone with audio signal distribution using EtherSound ES8 and ES220 interfaces.
“Using an audio network was imperative due to the long distances in the building and [RF interference],” says Matthieu Chautain, AV engineer for the project. “We chose EtherSound for several reasons: extremely low latency, number of channels available, flexibility, redundancy, and support of this protocol by a number of manufacturers we work with.”
Chautain also pointed out the need for a “clean” network to make sure that points along the way are EtherSound-compatible, citing an unpleasant experience with certain transceivers that resulted in randomly introduced jitter. He advises using an Ethernet analyzer to check for jitter or asking Digigram to vet your transceivers for you. AV
Dan Daley is freelance AV writer and frequent contributor based in Nashville, Tenn.
Whither Ethernet AVB?
Ethernet Audio/Video Bridging (AVB) gives the networked audio market what it ultimately needs: an IT-level traffic cop. Generally speaking, Ethernet protocols on the market can’t transfer low-latency, properly synchronized, high-quality media across a local area network (LAN) without other data traffic interfering.
Ethernet AVB, developed by the IEEE AVB 802.1 AVB Task Group, provides three major enhancements to 802.1-based networks: precise timing to support low-jitter media clocks and accurate synchronization of multiple streams; a simple reservation protocol that allows endpoint devices to reserve the bandwidth in a path to guarantee quality of service (QoS) for audio/video streams; and queuing and forwarding rules that ensure that AV streams will pass through the network within the delay specified. The bottom line: Ethernet AVB fast-tracks designated audio data streams into a higher-priority class of packet that are expedited ahead of standard, “best-effort” data traffic.
On a more technical level, AVB tells the network that audio and video are fundamentally different from other data. “All of the current [proprietary] protocols deal with the endpoints of a network, not the network itself,” explains John McMahon, executive director of digital products at Meyer Sound. “In between, at the routers and switchers, the needs of time-sensitive AV information can find the network to be pretty hostile territory. Ethernet AVB addresses that.”
Ethernet AVB requires all-new hardware but remains backward-compatible with existing Ethernet standards. However, the enhancements enabled by IEEE 802.1 standards require no changes to the Ethernet lower layers themselves and are compatible with all the other functions of a standard Ethernet switch. As a result, the entire Ethernet ecosystem is available to developers.
Rick Kreifeldt, vice president, systems architect at Harman, which has been a staunch AVB backer, says it’s the protocol’s open standards and scalability that will give it the kind of price points needed to bring networking to the vast and expanding pool of smaller AV projects. “Even CobraNet, which has the greatest reach of all of the systems, is still relatively costly and complex and not really scaled for much of the work in the AV market today,” he says. “And without scalability to that degree you don’t get the kinds of economies of scale that can really drive networking.” AVB, on the other hand, he says, is structured in a way that encourages development by numerous third parties to extend beyond the install market and into consumer and automotive sectors.
Ethernet AVB has the potential to become a standard for several reasons. For starters, the protocol is backed by huge consumer audio players, such as Apple and Samsung, which will drive much broader demand for AVB-compliant products. Samsung, Intel, Cisco Systems, and other companies have formed the AVnu Alliance, a trade group to promote the protocol.
Also Up and Running: Aviom A-Net
Among the other proprietary audio networks at an integrator’s disposal is Aviom’s A-Net technology. Like others, it’s based on the physical layer of the Ethernet stack and runs over Cat-5e cabling. A-Net comes in two flavors: Pro16, which supports bi-directional systems up to 32×32 and 48×16; and Pro64, which handles channel counts up to 64×64. With the right Aviom network hub, Pro64 can include fiber-optic transport.
Aviom markets the A-Net platform to a variety of markets, including broadcast, recording, corporate installs, and education. Recently, the largest installation ever of Avion A-Net gear was completed at Cuyahoga Community College’s Center for Creative Arts in Cleveland. The school’s Pro64 network connects studios, control rooms, and more throughout a 75,000-sqaure-foot facility, but it also allows the center to link with AV systems beyond its walls.
Tags :digital connections