Digital audio devices are in home and project studios everywhere these days, and connecting them requires an understanding of the various communication protocols and interface formats that send and receive digital audio signals. These protocols and interfaces include AES/EBU, S/PDIF, ADAT Optical, and TDIF, among others. It's important to understand the features and capabilities of these formats in order to make connections correctly and avoid the problems that can crop up if you don't.
To transmit digital audio, both the sending and receiving devices have to agree on the rate at which the data will be sent. To achieve this, all digital connections incorporate a clocking signal-called word clock-which provides a common timing reference. Word clock also defines the number of individual samples (words) sent per second. Word-clock signals are either carried on a separate cable from the digital audio data cable or embedded in the digital audio data itself.
FIG. 1: If word clock is carried on the same cable as the digital audio data, designate the transmitter as the clock source by setting its clock to “internal.” To sync other devices in the chain to this signal, set their clocks to “external.”
If word clock is embedded in the audio data, the transmitting device will become the clock master, and you should configure the device to use its internal clock as a timing reference. Then configure the receiving device as a slave to the transmitter's clock signal by setting its clock parameter to "external." You can then daisy-chain and synchronize additional devices to the transmitter's clock signal within the digital audio data stream (see Fig. 1).
FIG. 2: In this system, a master clock sends word clock over a separate cable to all devices.
Word clock can also be sent by a master-clock source on a separate cable. This clock signal can be sent by one of the devices in the system or a dedicated device whose sole function is to generate stable word clock. Such a device normally includes the connectors necessary for hooking up all your digital gear (see Fig. 2). A dedicated master clock is more common in complex professional systems.
One of the most common digital audio interfaces is called AES/EBU, developed by the Audio Engineering Society and the European Broadcast Union in 1985. The AES/EBU format transmits two channels of digital audio serially (one bit at a time) over a single cable at resolutions of up to 24 bits per sample. It is not restricted to any particular sampling rates, although typical rates include 32, 44.1, and 48 kHz. Along with the bits allotted for audio information, eight extra bits are used to carry subcode information: the sampling rate, error correction, and other information about the digital audio signal. However, not all AES/EBU devices use all of these subcode bits.
FIG. 3: The Apogee PSX-100 includes AES/EBU connectors (3-pin XLR) and word-clock BNC connectors (labeled WC In and WC Out).
The standard means for transmitting an AES/EBU signal is a balanced, 110-ohm cable that is terminated with XLR connectors (see Fig. 3). Balanced XLR cables can transmit signals over distances of up to 100 meters without interference, at a level typically between 3 and 10 volts. But not all AES/EBU cables are XLR; you might encounter an AES/EBU device that uses balanced 1 1/4 4-inch connectors or even a 75-ohm unbalanced video cable terminated with BNC connectors (those "push and turn" connectors you find on TV sets, VCRs, and cable boxes).
It's important to use the proper cables when connecting AES/EBU devices. Analog XLR cables, such as microphone cables, aren't right for the job. These have variable impedance ratings (typically 30 to 90 ohms) and can cause the digital data to be "reflected" back when it reaches the receiving end. This can result in timing errors (jitter), data corruption, and signal dropout. In addition, you should avoid passively splitting a digital signal.
The AES/EBU format is self-clocking: word clock is embedded in the digital audio stream. As a result, a single cable carries two channels of audio plus word clock. However, master-clocking AES/EBU devices is also possible and even recommended in larger, more complex systems. In this case, the master-clock signal is typically carried on a 75-ohm coaxial cable with BNC connectors (see Fig. 3). Keep in mind that not all AES/EBU devices have these BNC connectors, and some don't work with a word-clock signal other than the one in the incoming digital audio signal, so you might not be able to configure your own gear this way.
The Sony/Philips Digital Interface Format (S/PDIF), which is very similar to AES/EBU, can handle various sampling rates and resolutions of up to 24 bits, and also includes 8 subcode bits. However, the data in the subcode bits is somewhat different from the data AES/EBU subcode bits carry. The most notable difference is that some of the subcode bits are used by the Serial Copy Management System (SCMS), a copy-protection scheme that prevents multigeneration copying via S/PDIF. AES/EBU's corresponding subcode bits carry entirely different data.
The S/PDIF interface typically uses unbalanced, 75-ohm coaxial cable terminated with RCA-style connectors. A video-dubbing cable terminated with RCA connectors is fine for making transfers, but an analog-audio cable is not. The S/PDIF signal's amplitude is about 0.5 volts-much lower than the AES/EBU signal's.
S/PDIF is also often implemented in an optical format called Toslink, which uses a small fiber-optic cable made of plastic or glass. Such optical interfaces avoid the problems associated with electrical connections, such as cable capacitance and grounding issues.
As with AES/EBU, S/PDIF word clock is carried in the digital audio bitstream. However, S/PDIF devices don't typically come with separate BNC word-clock connectors, so they can't be configured for a master-clock system.
AES/EBU and S/PDIF use similar data formats, so you might think that getting the two interfaces to work together is a simple matter of using an XLR-to-RCA adapter-but this isn't the case. The signal levels are very different (about 5 volts for AES/EBU and 0.5 volts for S/PDIF); the subcode bits contain different data; and mismatching a balanced, 110-ohm interface with an unbalanced, 75-ohm interface can corrupt the data. It is therefore a bad idea to make direct connections between AES/EBU and S/PDIF equipment. Instead, you should use some kind of dedicated format-converting device.
FIG. 4: As you might expect, the Alesis ADAT XT20 includes ADAT Optical connectors.
The Alesis ADAT 8-track digital tape recorder comes equipped with a proprietary digital-connection format designed to carry eight channels of digital audio on a single fiber-optic cable (see Fig. 4). This format is known as ADAT Optical (sometimes called Lightpipe). You can also find the ADAT Optical interface on digital mixers, computer interfaces, synths, and effects devices.
Each channel can accommodate digital audio at sampling rates of 44.1 or 48 kHz with up to 24 bits of resolution, plus 64 bits for subcode information. However, ADAT Optical devices might be able to transmit only 20 bits of digital audio resolution. For example, all of Alesis's current recorders have a resolution of 20 bits or less. If you own a device with an ADAT Optical interface, you should check with the manufacturer to see whether the interface sends all 24 bits of digital audio.
The fiber-optic cables used for ADAT Optical are the same as those used for Toslink, but the data format and transmission rate are completely different. Therefore, you can't plug one directly into the other and expect results. Typically, ADAT Optical cables work well at lengths of up to 10 or 15 meters, and they can run even longer with glass cables.
The ADAT Optical format uses an embedded clock signal that does the same job as standard word clock. The timing signal is also transmitted on a separate 9-pin sync cable that carries additional transport-control information. Some devices equipped with the ADAT Optical interface require both connections to work properly, whereas others require only the optical cable to be connected.
Tascam's DA-88 family of 8-track digital tape recorders has its own proprietary digital audio interface called Tascam Digital Interface Format (TDIF). Unlike the other interfaces I've discussed here, this is a bidirectional interface: a single cable carries eight channels of data in both directions. TDIF cables are multiwire, unbalanced cables terminated with 25-pin D-sub connectors, and their recommended length limit is 5 meters. This interface is implemented on Tascam recorders in the DA-88 line, Tascam digital mixers and peripherals, and third-party products such as interface cards for digital mixers and computers.
As with most digital-interface formats, TDIF supports multiple sampling rates and resolutions of up to 24 bits. It is intended to operate as a master-clocked system with a separate 75-ohm cable terminated with BNC connectors that carries standard word clock. However, each pair of channels within the TDIF interface also carries a clock signal called Left-Right Clock (LRCK). This signal runs at the same rate as standard word clock and defines the odd and even channels within the pair. As a result, LRCK can often be used instead of an external word-clock signal if the TDIF device supports it.
Other digital audio interfaces are out there, but the ones I've covered here are the current crop of widely used formats. They all have similarities, so a basic understanding of these formats and how they handle the issues common to all digital interfaces will help you use your digital equipment effectively and make good digital interconnections.
Jeff Baust is a professor of music technology at the Berklee College of Music.