Many companies look at digital audio for aircraft systems as a way of improving system operation in the electrically noisy environment. Thanks to effective consumer advertising, "digital audio" has achieved some mystical properties it does not deserve, and an understanding of its attributes and utility has gotten lost in the process.
This column will provide a quick overview of practical digital techniques and system concepts for the cockpit. The next column will look deeper into the underlying science behind digital audio.
In digital audio design projects it helps to know that no existing ARINC control protocol standard is ideal for audio or audio control systems. While one of the control standards can be used, all of them have limitations and none of them offers compelling advantages.
The audio control bus needs special attributes that are not common to other airframe functions. It therefore does not benefit from a common bus connection. For true digital audio transmission, existing protocols are unable to support the required data bandwidth. A new format and interconnect will be needed to achieve the best performance, especially in the streaming audio model.
Digital audio systems can take many different forms and still be considered "digital." The first-stage digital system is when control functions–especially many remote volume controls and selections–are implemented digitally from a remote user control station. In the next stage of complexity, the audio can be processed digitally at a remote or local box. And in the final or most truly digital form, all the audio can be converted at a central location and then streamed digitally and combined digitally at each user station.
Digital controls can be added to fundamentally analog systems easily by using transmission gates and multiplying digital-to-analog converters (DACs) or similar dedicated audio processor circuits. These are used to select and adjust signals from radios or other sources and combine them via analog summation.
Potentiometers used as voltage dividers can feed analog inputs of a simple microcontroller to serve as level controls. Then any desired "taper" can be implemented in any control through software. Long-life digital shaft encoders can be used also as audio level controls, but they require a more complex interface.
Digital serial communication between a user control station and a remote processing unit allows many control functions to be implemented without contamination of the audio, which usually occurs with extensive analog interconnect wiring. This digital control link also requires little physical interconnect and can be as simple as a single RS-232 path. A powerful technique, it allows considerable functionality to be implemented in installations in which there are only one or two user stations and a remote audio processor is acceptable.
In the intermediate form, a digital signal processor (DSP) and user commands process and combine the audio inputs. (The inputs are converted to suitable digital form by a device called a coder-decoder, or CODEC.) This processing achieves the required composite audio signals. Audio inputs are multiplied digitally by their respective level setting, then summed by the DSP.
The DSP also can be used for active noise reduction, filtering and other signal conditioning in true digital form. A digital-to-analog converter performs the output conversion from this processing and recreates the desired analog signal from the digital data stream. The user can perform the processing remotely or locally. The processing can achieve good isolation between inputs because of the input digital conversion step, which avoids direct analog signal summation and its inevitable audio isolation problems.
Designing an intermediate system is complex and includes significant component availability risks with regard to DSPs and CODECs. It also has a high software component that may not be easily portable between DSP types. In addition, the system may not achieve one desired result: the shift to digital audio by eliminating analog wiring. However, it can reduce wiring considerably if all processing occurs at a consolidated remote unit. With this concept, final composite analog signals are exported to and from the user–or in a more complex implementation they travel as digital streams and are converted at the user station. The intermediate system may be viewed as the model to build. But it is not optimal as a long-term technology and offers only limited performance advantage with considerable risk elements. One should keep in mind with this and all other digital processing techniques is that the analog-to-digital and digital-to-analog integrated circuits for these functions almost never have extended temperature capability.
For the final configuration, which I think will become the dominant digital mode, we need to borrow from telephone technology, using its technique of subcarrier channels to pack many simultaneous audio signals into a composite high-speed digital stream. This stream is like a modified T-1 or integrated services digital network (ISDN) interconnect with increased audio channel bandwidth (minimum of 6 KHz vs. existing 3.5 KHz). In this configuration, all analog radio inputs and outputs go only to a single remote unit that converts them into two digital streams: one incoming for transmit and one outgoing for receive. They then are distributed digitally to as many stations as required.
The outgoing receive audio stream encodes packets for each radio or signal source and has all the available signals, including incoming intercom audio and supplemental data. The simpler incoming transmit stream contains the transmit audio from each user and control data that routes it to the correct radio or into the appropriate intercom bus. The streams are unpacked and summed or routed in either the user station or interface unit, providing the appropriate analog signals after conversion.
In theory, this stream in a star topology could support numerous stations with enough hardware. Certainly the maximum requirements even of large military aircraft could easily be achieved, although some hardware is required to support each distinct user in the remote unit. Using a ring or hub-and-spoke topology for the interconnect also is possible, to get dual physical connections to each station for added robustness. The transmit stream also could be altered to add every user within a single stream as a system variant.
This digital streaming technique sends all audio to every user. A digital mask can then select what is acceptable and requested for that user. Local decoding of the digital stream allows perfect interference-free combining, or generation, of all required signals. If the physical interface to and from user stations is a fiber optic cable, an interference-free system can be implemented, with only the inputs to the remote combining station susceptible to external signals.
While this system has many technical hurdles and some component risk, it is the optimal digital implementation, as it has virtually no physical interconnect other than two fibers to each station. It also has no real intrinsic path for audio cross-talk other than the physical coupling in the cables leading to the main conversion unit, as well as any chance contamination internally. The system avoids the analog source summing and mixing paths that create cross-talk, which increases with each user and audio source added to the system.
Many additional secondary features can be added to this digital network to make a stronger, more practical system, including dual or partitioned networks and fail-safe, fall-back dedicated analog connections to specific radios and a backup intercommunication link. For installation, the system also can self-configure and, at the control station, digitally label each audio source with its correct nomenclature. It can implement a priority or partitioned scheme, as well. The final network can minimize the possibility of cross-talk and interference, provide the minimum physical interconnect and have the most expansion and enhancement possibilities. As a final consideration, it’s critical to remember that in a pure digital system, there is no implicit or possible failure recovery. It will have to be added deliberately as a secondary analog pathway for emergencies and needs to be considered early in the overall system design.
Walter Shawlee 2 may be reached by e-mail at firstname.lastname@example.org.