Getting The Signal

Digital signal processing is probably one of the most enabling and disruptive audio technologies to come along since RCA debuted sound for motion pictures, sending the bells and whistles of the mighty Wurlitzer organ into a premature and pretty much unplanned obsolescence. Literally, Wurlitzer organs were on loading docks, ready to ship to theaters when the bad news came-in the form of cancelled orders.

While Saturday matinee organists joined the ranks of the unemployed, sound experts discovered a booming opportunity and have never looked back. Other than weekend nostalgia buffs who populate small community theater houses and centers to slip into that big-pink-champagne-glass of show tunes and black-and-white silent films, most people agree that pictures with sound are better than pictures without sound. Only time will tell whether life after DSP is better than life before DSP. It's quite possible that the telling truth could be in what used analog audio gear ends up in community centers or some other nostalgia driven project, via eBay.

While DSP is powerful and perhaps as undeniable as the changing of the tide, the phasing of the moon or the sound of the human voice, I still miss all the fiddled-with knobs and controls that can save or end jobs and create or eliminate service calls. I miss all of the distinctly labeled boxes from every little and large audio manufacturer from the rain forests of Washington State to the fog-enshrouded hinterlands of the United Kingdom and everywhere in between, like China, that have just the right sound quality, faders, switches or knobs. I miss the security covers. I miss the acres of racks, wall-warts and connectors. I miss 600 ohm balanced audio intra-rack wiring. Do you? I guess it's mostly just a matter of whether or not you, or your customers, have a knob or control fetish.

Digital signal processing is everywhere. It's in amplifiers and speakers, in mixing consoles and snakes, in codecs and echo-cancellers, in microphones and preamplifiers and in cameras and video processors. DSP has replaced dozens of black boxes in a system with just a few black boxes, or even just one box. All magical and colorful on the inside and typically unadorned on the outside, DSP has brought to us the promise of one-wire systems from the microphone all the way to the speaker; all borne of ones and zeros, meshed with packets and stirred in the cauldron of mathematical algorithms.

It's quite amazing to think of what is really possible with DSP; feedback elimination, EQ and delay on every channel; matrix routers with cross-point levels at any point in the system; remote-control integration (albeit limited compared to the knob rich environment of analog gear); automatic mixing; repeatable presets; steerable loudspeaker polar patterns-every block diagram symbol imaginable under the sun and then some. These and other functions are the new definition for a blank sheet of paper that is the DSP engine. Truly at some point the currently faux technical abilities of a pseudo audio device, say a "sheen" control, just might come into reality. The sky's the limit. Or is it?
DSP is mostly just another way of doing the same things we have done for years, only cheaper-in boxes, wire, connectors, bulk and price for a given system-and in some ways better. Audio and video tools and functions that may have been left out of a design are now there in the DSP box, awaiting our beck and call. And the tendency of most humans is that if it's there, we'll use it, right or wrong. This calls to account the real skills of the AV designer, salesperson or technician.

As designers, the human tendency for efficiency against billable hours may lead to an oversimplified block diagram, in which sources go into an under- or non-defined DSP box symbol on the functional diagram and then out to the amplifiers and speakers. The thought is that the installer will be able to use all the colorful magic on the computer screen GUI to build the right functional system, so why limit them with definition of a signal chain. This shifts the burden of design responsibility to other shoulders, including the integrators and even the manufacturers. Ensuring that a device has the right capabilities to satisfy the project requirements is the key to success. And when all of the elements appear to be magically available upon demand, it's easy to make assumptions that can lead to failure, or at least a change order.

For sales people, the hard-learned pitch-grooving of this particular EQ over that or this compressor/limiter over another often gets distilled down to DSP input/output counts, graphical look, algorithm-quality, analog-to-digital/digital-to-analog conversion, a few knob-like controls and a relatively reduced price.
Franchises will never be more important as a differentiator as they are with DSP. There just currently aren't enough manufacturers to populate the sales landscape with as many discreetly different DSP devices as exist in the analog domain, except as to what functions may be within a given DSP. The more that is crammed into a DSP box, the better, seems to be a rule. At least until a manufacturer needs to add some part numbers to the catalog due to all of the analog gear they obsolesced with DSP, then DSPs just get repackaged with limited functionality. I/O counts, aimed toward a specific application, like speaker processing or other specialized usage, most likely are doomed to misapplication. The ability to put everything into one box pressures every manufacturer and ultimately every salesperson out there.

For the technician (or designer who has to truly define the system), the real risk/reward of DSP comes into play. The reward comes from being able to create, wire, tweak and test a very complex system chain, without ever picking up a tool, other than a computer mouse and keyboard. The risk comes in the depth of what a box can or can't do. To really plumb the depths of a modern DSP device, especially one that has a lot of I/Os and functions, you generally have to invest in very specific training and certifications. You do this to learn all the cool things you can do (like using a button push on a control input to fire serial code to a projector on an output).

You also take the training to learn that unless the device has a fixed architecture, many great pieces of equipment do have a fixed architecture, putting them in a class separate from the flexible architecture devices. You just might not be able to have all those delays, feedback eliminators and routers you thought were possible, due to running out of DSP resources. Issues of latency, routing limitations or lack of complete GUI control of certain signal functions, like initial gain stage adjustment, can hobble or confuse system setup and operation.

Then there's that "common point of failure" issue, in which if a box fails, the entire system fails and the yearning for a 600 ohm patch-around rises like a head of hot steam. Redundancy is mandatory in larger systems, as is a back stock of DSP units and a ready laptop. Did you remember to download the configuration and store it in a secure place? Of course you did, just like you used to write down the EQ settings on a graphic. New game; same old rules.