Shortcuts To The Bus

Where has the audio gone? We used to be able to place a volt meter on an audio cable and see the instantaneous level changing in sync with the sound we heard. Now, we may need computer networking tools to troubleshoot the audio signal. As our signals are adapted to the digital world, new problems and new possibilities are opening up. In the long run, this transition will affect every aspect of audio technology.

If we look at the audio system as a network of data streams, then the signal can be viewed as a series of time-stamped packets of data. The data must be reassembled at the far end of the network in the correct order using the time-stamps. What we do with the audio in between is going to be limited by the processing power we can access and the amount of latency (signal delay) that can be tolerated. We can reduce latency by increasing the speed of the signal processing and by decreasing the distance the signal must travel. If we have complete control over the audio signal path, we could assign priorities to the audio based on the needs of a particular signal. A microphone that is being reinforced in a room would need very low latency to prevent echoes from distracting the performer and audience. A pre-recorded sound could cope with much longer latency, as there is no fixed reference (acoustical source) that must match with sound from the system.

We can look at these audio systems in a new light. A digital microphone could produce a signal that is assigned a specific address (such as mic 1 in room 342) and then processed and ultimately sent to the loudspeaker address (such as loudspeakers 1 to 10 in room 342). The processing that occurs in between is accessed by instructing the system to direct data stream M1-342 to be processed by equalization, compression and mixing. The resulting stream of data M1-342P is then routed to loudspeakers S1-342, S2-342, and so on. The process of pre-amplification, processing and power amplification becomes a part of handling the data. The current practice of processing mic signals only after they have passed through a pre-amp will eventually be overcome, as methods of sampling acoustical energy directly in the digital domain become practical. The same fate will befall the loudspeaker; where the power amplifier will become an acoustical displacement controller that uses the driver in the final stage of converting the data back to sound. When the transducers have in-built addresses, it becomes possible to define the audio system to suit individual signals throughout the chain, rather than the current practice of modifying the mic signal, then combining all the signals, then modifying the signal for the loudspeaker.

How the audio streams are processed is already changing. There are manufacturers that currently produce processing hardware that can be interconnected and share the DSP resources fluidly. As processing resources are used in one DSP box, the signals are networked to other DSP boxes to perform the balance of the processing requirements. Each audio signal path draws DSP resources from a common pool. If more processing is required, the hardware (pool) is increased incrementally to meet the demand. Currently, the system designer must assign processing priorities to the signals and keep track of which signals may have too much latency. Soon, it will be a matter of assigning properties to each input and the processor will optimize the processing based on these properties (such as: sound reinforcement mic – lowest latency; recorded announcement – highest latency). Ultimately, the source device (digital mic) will include these properties and have a unique name that accompanies the audio data stream to ensure that all the processing is handled with the appropriate priority access to DSP resources.

When the streams of audio data include their unique names and properties, the entire system can be connected using signal buses. It will not be necessary for a mic to have a dedicated cable to a dedicated mic input. Instead, the signal is placed on the nearest audio network bus and the assigned processing will be applied to the data that is identified to be part of that stream. The nature of the network (such as Gigabit Ethernet) would determine how many mics, processors and loudspeakers can be connected to a data bus. In effect, the mic signal stream will reach the loudspeaker input directly, but would only be picked off the bus if no processing was assigned to it. If the signal was labeled for processing, the loudspeaker would only be looking for the stream that included that processing. The actual wires and cables connecting these devices would bear no relation to the signal path. Just don’t expect to do troubleshooting with a volt meter.

United Entertainment Media Inc.
460 Park Avenue South, 9th Floor
New York, NY, 10016
Ph. 212-378-0400
FAX 212-378-2160
by e-mail

page construction