A New Dawn for DSP?

In Web Articlesby tfwm

Digital signal processing has already brought untold flexibility to a wealth of install applications. But with full-scale networking likely to become the default, could it be that an even more exciting era is just around the corner?


The Symetrix SymNet Radius 12×8 DSP. An open-architecture, scalable processor with Audinate’s Dante protocol (Dante Brooklyn card shown)

Alongside the advent of more powerful networking solutions, the emergence of digital signal processing (DSP) that is capable of being implemented quickly and (relatively) easily for all manner of fixed install applications has arguably been the defining pro-audio story of the decade. From Acoustic Echo Cancellation (AEC) for video-conferencing to site-wide routing and distribution, high-capability DSPs have heralded a minor revolution in the built environment.

In an industry that has always paid extremely close attention to the bottom line, it’s unsurprising to discover that lower-cost solutions have done much to initiate the New DSP Dawn. Not only are DSP chips “running at considerably higher MIPS (millions of instructions per second) rates than they were ten years ago, today’s faster chips enable designs that use fewer components – meaning lower costs and increased reliability,” observes Dane Butcher, founder and CEO of Seattle, US-based DSP pioneer Symetrix.

Software developments have also helped to slash DSP costs, with advanced memory management features and higher clock speeds paving the way for designs that, notes Butcher, “use only a single DSP chip to exceed the processing power of designs that used to require four chips.”

Highly effective automixing, loudspeaker management and zone mixing/paging are among the other common applications of these new-generation DSPs. But although their superiority to conventional processing solutions is now unquestioned, the sheer quantity of current DSPs can easily leave the integrator – let alone the end-user – with something of a quandary. So what are the factors that need to be borne in mind when selecting a DSP that is truly capable of delivering the processing power you need?

Processor Practicalities

At the most basic, pro-audio customers are faced with one fundamental decision when it comes to DSP – namely, between a fixed and open architecture solution. As their name should imply, open architecture solutions afford the greatest possible scope for custom configurations to complement a specific application.

Doug Windle is audio visual manager of Texas, US-based systems integrator ACE Audio Communications, and explains: “Some hardware has limits with both fixed input/output configurations and processing power; however, these units will still work in 90% of all projects. If you utilise the open architecture with modular input/output configurations, the options are almost endless, especially when these are coupled with networked audio.”

Powerful they may be, but they can also be more costly to implement than their fixed brethren due to the fact that they essentially offer a blank canvas. “The integrator has to provide all the infrastructure inside the processors to fit the intent of the system,” confirms Windle.

Accordingly, lower cost fixed architecture solutions still tend to be more popular for the less demanding applications. With these DSPs, “the price is certainly more attractive, and they are really quick and easy to set-up and operate,” says Windle. Input/output configuration and the layout of the software are more limited than with open architecture DSPs, but some manufacturers are able to deliver significant customisation. Symetrix is among them: not only does its Jupiter solution possess a large library of preconfigured layouts to choose from, it can also be deployed with layouts created especially to match user requirements.

Although the size and budget of a given project will hold great sway when it comes to the choice of a fixed or open architecture solution, there are plenty of other, more specific requirements that need to be borne in mind. Moreover, remarks Symetrix vice-president of global sales Craig Richardson, “the importance of a particular factor will vary depending on the application and scope of the decision. For example, you might want to consider whether it’s for one project or is it to become a new standard for the organisation.”

Integrators therefore need to think very carefully about whether their DSP has the necessary algorithms and processing power to get the job done. Ensuring overall efficiency of implementation is, if anything, even more challenging since it forces the system designer to grapple with the ever-thorny issue of future-proofing. Richardson puts his finger on the central issue here: “If your installation requires 90%-plus utilisation to implement your system, you may want to consider the next ‘beefier’ device, or another manufacturer’s device, to ensure you have some processing headroom.”

Other concerns are more universal, but no less important for that. Among other factors, Richardson highlights audio fidelity and form factor. Then, what we might label the post-installation life of the DSP is to be borne in mind. Getting end-users to feel comfortable with the system in the first place is one thing, but when you have queries or concerns, how easy is to get a knowledgeable person on the end of a phone?

Delivering Dante

In the old point-to-point days, that list alone would have been more than enough with which to contend. But now, as full audio networking brings its flexibility to bear in every sphere of the live and fixed install world, the ability to connect quickly and easily to other digital devices is assuming ever-greater significance.

This is easier said than done because a variety of solutions are continuing to compete for integrators’ attention. Audio-over-standard-Ethernet technology CobraNet retains its advocates, but latency and hardware costs can be problematic. Looking ahead, the Audio Video Bridging (AVB) project championed by the AVnu Alliance is set to provide a comprehensive, standards-backed approach to audio/video streaming. AVB has racked up some successful, high-profile demonstrations, but a shortfall in end-user understanding is arguably inhibiting the technology’s adoption.

It’s perhaps no wonder, then, that proven low latency Audio over Ethernet media networking technology Dante, developed by Sydney-based Audinate, is proving an increasingly popular solution during this uncertain period. Dante ensures a high level of interoperability with third-party devices by an OEM partner list that now totals 105 companies. “There are about 175 products from OEMs available and at least 200 more in the development pipeline. Audinate shipped over 3.5 million Dante channels last year, so it is a proven technology that works on a converged IT network,” says Ervin Grinberg, director of marketing, Audinate.

Symetrix is among the companies to be supporting this rising star of the networking world. “Symetrix DSP devices have the advantage of using Ethernet networks for control and Audinate’s Dante technology, and are typically audio front-end for video conferencing systems or other audio transport networking devices,” explains Richardson, who adds that giving customers a range of options is very much the order of the day: “Because of where our products sit in the network, we support standard IPv4 networking along with Auto-MDIX for 10/100 Mbps Ethernet for data and control, and 1Gbps network interface for Dante that is also compatible with IEEE 802.3. When the Symetrix devices are controlled over Ethernet, we use UDP for command and control; of course, RS232 serial control is also an option.”

Manufacturing Model


The Symetrix ‘legacy wall’, a collection of legacy products spanning the company’s heritage from studio to commercial audio

The fact that customer demand for sophisticated DSP will continue to grow is beyond question; but so too is an increasing pressure on margins. Manufacturers will therefore need to ensure that they deliver products to market at a reasonable price-point if they are to remain competitive – and that calls for a production cycle that is streamlined and efficient.

In the case of Symetrix, that has translated to in-house product assembly at its US facilities and a ‘Lean Manufacturing’ model for the foundation of its production operations. “The Lean Model focuses our efforts on responding to actual customer demand in real-time while reducing product lead times and costs,” says Eric Dies, vice-president of manufacturing, Symetrix. “The direct benefit to our customers is a very high quality product that is readily available at a reasonable cost.”

While the benefits of high-spec DSP in the networked world are now firmly established, the best methods for helping customers to implement their solutions successfully remain a matter of debate. It is this territory, then, that is likely to be the real battleground for manufacturers as DSP enters its next phase of evolution.