Expanding Your Audio Setup With ADAT

Despite being over 30 years old, ADAT Lightpipe remains the number one way to add more inputs and outputs to your DAW.

Audio interfaces fulfil multiple functions. They amplify signals from mics. They convert analogue signals to digital signals. They transfer these digital signals to the computer. And, on the way out, they accept digital audio signals from the computer, convert them back to analogue and perhaps amplify them for headphone listening.

Only one of these functions absolutely has to be done by the interface itself. That is the transfer of digital audio to and from the computer. The other functions can be delegated to other devices — and, since these secondary features cost money to implement, they often are.

Manufacturers want their products to be usable straight out of the box, so interfaces have mic preamps and A‑D conversion built in, along with line outputs and a headphone socket or two. But not everyone needs the same number of preamps and line‑level I/O. The solution is to provide extra digital paths to and from the computer, and allow users to add other devices later on to take advantage of these.

There are several different types of digital connection that can be used to expand an audio interface in this way. At present, Ethernet and MADI are mainly found on high‑end professional equipment, catering to situations where large numbers of inputs and outputs are required. On home and project‑studio gear, you’ll find ADAT, AES3 and two types of S/PDIF connection — but as AES3 and S/PDIF are only two‑channel systems, the focus of this article will be on ADAT Lightpipe.

Ups & Downs

A digital audio signal is encoded as a string of ones and zeroes. There are different conventions as to how this encoding should take place, and they are mostly incompatible with each other. There are also different ways in which a given string of ones and zeroes can be represented physically. In order for digital audio to be transferred between two devices, they need to support both the same physical connection, and the same encoding scheme.

Since it’s expensive to design and manufacture new connector and cable types, new formats often recycle existing physical infrastructure. This makes sense from an economic perspective, but can lead to confusion! For example, AES3 employs the 3‑pin XLR connector that is so widely used for analogue audio in studios, while the coaxial version of S/PDIF uses the RCA phono connector that is ubiquitous on hi‑fi equipment. This makes it possible to accidentally connect a digital output in either format to an analogue speaker input — a mistake most people are unlikely to repeat.

AES3 and S/PDIF actually use the same encoding scheme, with trivial differences. The same piece of audio will be represented using the same series of ones and zeroes in both cases. Moreover, this scheme is realised in the same way in both AES3 and coaxial S/PDIF. The ones and zeroes are defined as high and low voltages in an electrical conductor; a cable connecting two XLR or phono sockets can, in principle, carry either type of signal.

By contrast, the other optical version of S/PDIF encodes the audio into exactly the same pattern of ones and zeroes, but uses a very different physical representation of this data. In optical S/PDIF, they are not high and low electrical voltages, but pulses of red light directed along an optical fibre. This is terminated at either end by a connector called TOSlink, originally developed by Toshiba. There’s no viable way to terminate an optical cable in an XLR or phono connector, and even if you could, it would be spectacularly pointless.

In this case, though, there’s also potential for confusion, because when the engineers at Alesis developed the ADAT Lightpipe protocol (see box), they decided to reuse the same optical connectors and optical‑fibre ‘cables’. It’s thus physically possible to send S/PDIF data to a destination that’s expecting an ADAT signal, and vice versa. But doing this will be just as futile as trying to make an XLR input accept an optical signal, because S/PDIF and ADAT use different encoding schemes. It would be like expecting a German speaker to follow instructions issued in Mandarin: they would be heard, but not understood.

Streaming Services

From a technological point of view, AES3, S/PDIF and ADAT Lightpipe all have much in common. They are all synchronous formats, meaning that they transfer data in a continuous stream which keeps pace with the analogue signal it’s derived from or turned into. None of them has robust error correction, so if something goes wrong even momentarily, you’ll hear it! They are all fundamentally one‑way streets: a single cable can carry signals only in one direction. If you want to pass signal in both directions between two devices, you’ll need two connections. Likewise, the connectors on the units themselves are doomed to be either inputs or outputs forever, and can’t switch roles. This usually means that if, for example, your audio interface has only a single optical port, you’ll only be able to use it to add additional inputs; its quota of outputs is fixed.

All of these formats also have a fixed, and limited, data bandwidth, which translates into fairly strict limits on the amount of audio that can be piped down them. AES3 and S/PDIF support stereo (two‑channel) audio at sample rates up to 192kHz. ADAT Lightpipe supports eight channels of audio at sample rates of 44.1 or 48 kHz, though if you need to operate at higher sample rates, you can often do so using a kludge called S/MUX (see box), at the expense of reducing the channel count. In all of these formats, the audio is sent in a 24‑bit, fixed‑point format. This has a theoretical dynamic range of 141dB (accounting for dither), which is far greater than any real‑world signal, but is not as foolproof as the 32‑bit floating‑point format that is becoming popular in some circles.

Not just for the home studio: many flagship interfaces from premium manufacturers also offer ADAT I/O, either as standard or via expansion cards.

Ins & Outs

By far the most common use case for all these formats in the home studio is to add extra preamps or analogue inputs to an audio interface. Sometimes this is necessary in order to increase the total input count, for instance if you want to record a multi‑miked drum kit and don’t have enough inputs to plug all the necessary mics in, or you’re running out of inputs for your growing synth collection. In other circumstances, you might want to do this in order to add a better preamp, or one that has features not available on your audio interface. For example, some ‘voice channel’ products that incorporate analogue EQ and compression as well as preamplification have an S/PDIF output that allows them to connect digitally to your audio interface.

There are also good reasons why you might want to use digital expansion to add more analogue outputs to your system. Perhaps you want to integrate external effects units into your DAW mixes, or send stems out of your software to mix on a hardware console. Possibly you wish to send multiple CV and gate signals directly from your computer to control analogue synths. Or maybe you have invested in an immersive monitoring system and need extra outputs in order to feed all those loudspeakers! I’ll consider each of these use cases in a moment, and look at how ADAT expansion can help meet these needs. But before that, we have to look at one more topic that’s fundamental to digital audio. Clocking is one of the most important, yet most poorly understood aspects of digital audio.

Putting The Clocks Forward

In a digital audio signal, the ones and zeroes that represent the audio waveform itself are interspersed with additional ones and zeroes that fulfil what might be called administrative functions. In all of the formats we’re looking at here, for example, the actual audio data is packaged into a series of bundles or ‘frames’ of fixed size and duration, with extra data marking the beginning and end of each frame.

But how does the receiving device know whether a particular one or zero belongs to the audio data or the frame definition? It needs to be able to count the pulses and frames coming in with absolute precision, and this is only possible if both the sending and receiving device have exactly the same understanding of the duration of a frame. If there’s even the slightest discrepancy, the two units will drift apart over time. This will cause errors, which usually manifest themselves as clicks and pops in the audio.

The Focusrite Clarett+ OctoPre, ART TubeOpto 8 and Audient ASP880 feature, respectively, switchable insert points, valve preamps, and variable mic input impedances.

The solution to this problem is for both devices to adopt the same timing framework. In essence, this is done by putting one device in charge of defining time, and making others follow it. This device is known as the clock ‘master’ or leader.

The clock signal can be distributed on its own, using separate cables, most often in the ‘word clock’ format. This repurposes yet another connector from the analogue world, in...