Following the Digital Audio Chain
by Michael Karagosian
©2004 Karagosian MacCalla Partners All rights reserved worldwide
Edited version published in the July 2004 issue of Pro AV
Click for sidebars to this article.
There is theory, and there is fact. Some might call the pair fantasy and implementation. The words "digital audio chain" conjures up images of both. Yet the fine line between theory and fact is always moving. New technologies regularly find their way into the marketplace that allow design and installation professionals to do things never before possible. Understanding that it can be a challenge to keep abreast of that fine line, this article provides a survey of both the ever-factual present as well as the theoretical future of the digital audio chain.
On the surface, sound systems aren't all that complicated. There is a source, some processing, and a speaker. We as users expect technology to offer regular improvements and innovations in each of these areas. However, true innovation is experienced when new architectures become realizable and affordable. New architectures characteristically distinguish themselves not in the areas of source, processing, and sound reproduction, but in the areas of control and transmission.
Perhaps the most fundamental innovation that digital technology brings to the audio industry is the convergence of the system control path and the audio signal path. Historically, the signal path for system control and the signal path for audio signals have been different. Certainly, with analog systems, there was little opportunity to combine them. But modern digital transmission technologies provide the means for both asynchronous and isochronous data to coexist on the same wire, allowing not only complex control and audio signaling, but video signaling as well.
IEEE 1394, a purpose-built digital A/V bus, is one such innovative technology. Originally introduced by Apple Computers under the name "Firewire" in the 1980's, it was standardized in 1995 by the IEEE. The IEEE 1394 bus offers both high-performance isochronous signal transmission, as well as high-reliability asynchronous transmission. The IEC 61883 standard was later created to define the management and application of these services. As originally introduced, 1394 was limited to a bandwidth of 400 Mb/s and 5 meter runs. In 2001, the 1394b standard was approved, extending the bus bandwidth up to 3.2Gb/s, while also expanding the choices for physical transmission mediums. 1394b can travel up to 100 meters using copper-based Cat 5 cabling, and further distances using optical fiber.
While 1394 was intended for A/V transport, it has made little headway into commercial audio installations. That, however, will soon change. 1394 brings tremendous advantages to the consumer industry, due to its ability to simplify cabling in the home and unify the control of devices connected to the bus. The growth of 1394 in the consumer market will lead to cross-over applications in the commercial market. For control, 1394 employs the AVC control protocol described in IEC 61883. (AVC stands for Audio Video Control - not to be confused with the AVC Advanced Video Compression effort conducted jointly by the ITU-T and ISO/MPEG.) For security, DTCP, Digital Transmission Content Protection, provides a standard means to protect highly-valued digital content. This powerful combination of features gives it the green light by both content industries and silicon manufacturers, which will make it the interconnect of choice for A/V source devices for years to come.
Of course, the popular network technology today is venerable Ethernet. Ethernet has found its way into commercial products of many types, providing both audio transmission and system monitoring and control. For audio-over-Ethernet transmission, Cirrus Logic's Cobranet™ is the most widely used method for networking high-quality digital audio. Competing technologies are also offered by Digigram and Gibson Labs. Notably, each of these methods make use of the MAC layer, or "lower layers" of Ethernet. They do not operate over IP (Internet Protocol), as IP is not optimized for high-quality isochronous transmission. IP support is desirable to users, however, as many products incorporate monitoring and control methods that require the IP layer of Ethernet. In response to this demand, some audio-over-Ethernet implementations provide ways to tunnel IP traffic.
Audio-over-Ethernet has proven itself to be a powerful method for transporting audio. It allows a large number of audio channels to be placed over one Cat 5 cable, substantially reducing installation costs. Most importantly, it builds upon a technology whose advancement is not dependent upon the commercial audio market.
But there are those cases where digital transmission and control may seem beneficial, yet remains unsuitable, such as the cinema. Typically, cinema amplifiers are racked in the projection booth, hundreds of feet from the speakers behind the screen. It would seem ideal to replace the long runs of heavy gauge copper wire with a single run of Cat 5 cable, along with audio-over-Ethernet and power amplifiers behind the screen. Not only would this save dollars in new installations, but the improved speaker damping factor would add to the quality of sound in the auditorium. While a few cinema installations have taken this step, most haven't, due to other concerns. John Wolski, Vice President of Projection and Sound at Loews Cineplex, points out that "equipment behind the screen is vulnerable to damage and theft." He prefers to keep his amplifiers in the booth, where they're safe and easy for maintenance personnel to access.
Those who have labored to install overhead speakers in dropped ceilings may be surprised to hear that the office ceiling is ripe for innovation. In-ceiling amplifiers and sound masking generators are not new. Armstrong i-ceilings® offers such a product, incorporating digital signal processing to implement speech privacy and sound masking. One can imagine a future ceiling speaker panel capable of monitoring ambient sound levels and self-adjusting its volume accordingly. Taking the thought further, RFID technology in the work environment could allow the public address system to locate you and direct your public address pages to the zone you're in. Local DSP can also enable a talk-back function through the ceiling, bringing the office a little closer in public address performance to the Starship Enterprise.
As fanciful as some of this may seem, the enabling technologies for such applications are with us today. IEEE 802.3af defines power-over-Ethernet, providing the ability to deliver 12-13W of power to a network node while retaining backwards compatibility with ordinary Ethernet cabling and distribution. Cirrus Logic utilized 802.3af at this year's NSCA, driving an Armstrong i-ceilings® sound panel equipped with an 802.3af-powered Cobranet™ module and amplifier. Ray Rayburn, Senior Consultant for the Peak Audio division of Cirrus Logic, referred to the demonstration as a "proof of concept", adding that "the use of 802.3af-driven Cobranet™ would reduce the ceiling speaker cabling to one Cat 5 cable per sound panel."
Audio-over-Ethernet, however, may have a competitor for driving new ceiling applications forward. Most office paging is integrated with the telecommunications switch, an area where IP technology has already established itself through the use of VoIP (Voice over Internet Protocol) and the IP PBX (Private Branch Exchange). As the name implies, VoIP, in conjunction with the IP PBX, is used to transmit voice over office IP networks. By adding an 802.3af-enabled data switch, the IP PBX can transmit public address audio directly to powered ceiling speakers. Adding DSP capability for our fanciful features is not so far-fetched.
Digital signal processing technology is the core for most installations today. While DSP-based audio processors have been on the market for over 10 years, their sophistication and power continues to increase. The leader for flexible, complex audio signal processing has long been Peavey's MediaMatrix, soon to be upgraded by the new Nion generation of digital signal processors. Other products in this class include the BSS SoundWeb and Biamp Audia signal processors. This class of product combines the unique ability to provide a user-defined signal processing path with the capability to create sophisticated custom control screens. Their flexibility is demonstrated in their breadth of applications, as many of these products have found their way into stadiums, meeting halls, courtrooms, theme parks, and sound studios.
But flexible general-purpose signal processing is not always what the application calls for. IED (Innovative Electronic Designs) combines the power of DSP with purpose-built hardware in their popular 500ACS series announcement control system, used in 80% of the top airports in the US. Mark Lewellyn, Director of Sales & Marketing for IED, describes the system as "having evolved through the years into the current all-digital, software-driven generation." The 500ACS is feature-driven for sophisticated paging applications, combining pre-recorded digital announcements, digital announcement storage and scheduling, and advanced zone control.
Power Amplifiers and Speakers
Most of the power amplifiers in use today are based on Class A/B designs. Even the so-called Class G and H amplifiers are Class A/B designs at their core, but with switched or audio-modulated power rails. Silicon power devices for Class A/B amplifiers, however, may be a dying breed. Certainly, the selection of power devices available to manufacturers has lessened somewhat over the years. The economics of power amplifier manufacturing has already begun to point to true digital Class D-type technology, where the output devices are not linear in nature, but simply turn 'on' and 'off'.
Digital amplifiers are not new, but they also have not always been successful. "Most digital amplifiers have been plagued in the past with power supply and other design problems", says Kevin Belnap, a product manager for Texas Instruments. To gain a jump on the design issues, TI acquired Toccata Technology for their proprietary version of a Class D design. Toccata's technology is best known in the TacT Millennium digital amplifier, a respected amplifier among audiophiles. Other proprietary versions of the Class D amplifier design are the 'Class I' technology marketed by Crown International, and the 'Class T' technology licensed by Tripath Technology. According to TI's Belnap, "the next 5 years will witness a major shift in amplifier technology as improved digital design techniques are introduced to the market."
Class D amplifiers and their offspring offer the advantage of high power efficiency, allowing more power to be delivered from smaller boxes. They also allow direct digital audio inputs, without digital-to-analog conversion at the input. But Class D amplifiers do not save the digital audio signal from D/A conversion. The amplifier itself is a converter, integrating the digital audio signal into an analog signal at the amplifier's output. To achieve what could be the ultimate digital experience, the human ear must perform the integration. For this, digital speakers are needed. While some products come close, no loudspeakers on the market today are truly digital in nature. However, the technology of choice for a true digital loudspeaker appears to be the unary-driven pressure transducer array, and companies such as 1 Ltd and Texas Instruments could be the leaders in developing this. (See side bar on Eliminating D/A.)
Having explored many of the digital audio technologies on the horizon, what could possibly lie beyond? Perhaps the most innovative approach to sound reproduction will be made possible through digital synthesis of the sound field. One way to achieve this is through the use of multi-speaker arrays and beam steering. Line arrays have already proven popular for beam steering applications, although the problem of unwanted lobes with steered line arrays requires the system designer to model the system in advance. More recently, a technology for beam steering with 2-dimensional arrays has been made available for licensing by 1 Ltd, in Cambridge, UK. Pioneer introduced this technology for the home with its Digital Sound Projector product. The technology may also find interesting applications in commercial venues.
Another technique for digital sound field synthesis is Wave Field Synthesis, which many researchers have been positioning as the processing technology of the future. WFS is well known for its ability to realistically localize sound in 2 dimensions over a broad sweet spot. The IOSONO WFS technology from Fraunhofer Institute in Germany is a prime example. (See sidebar on WFS.)
Last but not least, the issue of security will eventually affect commercial digital sound applications. Consider, for example, the situation where an installation includes a DVD-Audio player, and it can distribute a direct digital multi-channel signal from the player to various locations in the facility. In this particular digital sound system, the audio from the disc is transmitted digitally across an unsecured audio-over Ethernet technology to various locations in the facility. Anyone clever enough with a compatible audio-over-Ethernet receiver can make a pristine copy of the copy protected disc. As the copy protection issue becomes more prominent in the consumer market, it will most certainly spill over into the commercial market. That's a big hint to commercial product developers.
Ahh, we'll remember the good ol' days when quality audio meant shielded cables, differential inputs, and heavy gauge wire to analog speakers!