Demystifying Digital Interface Formats & Protocols
There are various digital audio formats, or protocols, used within the Pro Audio and Broadcast industries, all with wonderful names and acronyms which at first may understandably baffle those new to recording or live applications. The following will explain the most commonly used and referred to digital formats and how they apply to ‘real world’ situations.
Most Common Formats
There are four main digital protocols used to connect digital audio products, namely SDPIF, AES/EBU, TDIF & ADAT. But why use them at all? Well, there are a couple of extremely useful reasons why to keep signals in the digital domain once they have gone through a single conversion stage, i.e. an analogue-to-digital converter. However, there are also synchronization aspects to consider when connecting two or more digital products via any of the following protocols, which will be dealt with at the end of this blog. So why keep it digital?
The first reason is simply to minimize the noise introduced by the quantization process that is used to code all analogue signals into the digital domain. Note that this aspect cannot be removed as it is inherent within the coding of analogue signals into their digital counter parts, although the greater the bit depth to lower the level of quantization noise introduced. This noise simply results from the difference between the actual level of the analogue signal and its digital counterpart at the point in time the signal was converted, or sampled. Therefore, the more conversion stages the greater the induced noise.
Next up, maybe surprisingly, is noise yet again! When analogue signals are transmitted through cables they are subject to interference by electromagnetic radiation, for example radio waves, mains hum, etc. Although there are means to help minimize this, by using balanced wiring, digital signals are simply impervious to this due to the nature by which they are coded.
Last but not least is convenience and connectivity. A single light-pipe can carry 8 channels of audio, as opposed to what otherwise would be an 8-way loom, or multi-core cable. This reduces the cabling complexity of serious studios setups and minimizes the probability of earth loops, etc. which can the nemesis of any pro-audio engineer.
SPDIF is simply an abbreviation for Sony Phillips Digital Interface Format. The SPDIF protocol delivers a stereo digital (L-R) audio signal via a single lead. The connection may be either co-axial in nature, e.g. a phono to phono cable, or optical, which requires a Toslink/Light-Pipe cable. Although ideally co-axial cables with specific impedance are required, I have found from experience that any good quality phono lead will generally suffice. For more analysis of the format, see this video.
SPDIF can be found on a number of products from mastering products such as TC Finalizers to Firewire and USB Audio Interfaces, e.g. RME Fireface 800 and Focusrite Saffire Pro range of products.
SPDIF is essentially a consumer format of AES/EBU. AES/EBU is an abbreviation for Audio Engineering Society/ European Broadcasting Union, which is a professional digital interface format, or protocol and although very similar to SPDIF it actually uses higher signal voltages.
AES/EBU uses a Balanced Co-axial connection and delivers a stereo digital audio (L-R) signal via one XLR-XLR cable.
TDIF is an abbreviation of Tascam Digital Interface format, which is a protocol designed to deliver a 16 channel digital audio stream (8 send and 8 return) via a co-axial connection, which is in the form of one 25-pin D-Sub cable.
Note: Due to bandwidth limitations, one TDIF connection can only support bit widths of 16 to 24 at sample frequencies ranging from 44.1 kHz or 48kHz if passing 16 channels of digital audio. The channel count is halved if running sample frequencies of 88.2kHz or 96kHz, i.e. two TDIF ports will be required to pass 8 channels of I/O at either 88.2kHz or 96kHz.
ADAT, which stands for Alesis Digital Audio Tape, was a protocol designed for the Alesis ADAT digital recorder that recorded 8 channels of digital audio to a VHS tape. This protocol has way outlived the ADAT machines and is commonly used within Pro Audio and Broadcast applications for passing multi-channel digital audio streams. In short, ADAT may be thought of as a Toslink or optical equivalent to TDIF and two Toslink cables are required to send and receive 8 channels of digital audio. A very detailed history of ADAT can be found at http://artsites.ucsc.edu/ems/Music/equipment/digital_recorders/Classic_ADAT.html
Note: Due to bandwidth limitations, one ADAT pair can only support bit widths of 16 to 24 at sample frequencies ranging from 44.1 kHz or 48kHz if passing 16 channels of digital audio, i.e. 8 I/O. The channel count is halved if running sample frequencies of 88.2kHz or 96kHz, i.e. two ADAT ports will be required to pass 8 channels of I/O at either 88.2kHz or 96kHz.
Users will rarely encounter AES/EBU unless involved in professional broadcast or studios. Similarly, TDIF is an extreme rarity these days and with the demise if the DA-range of digital recorders for which it was developed (Tascam’s version of ADAT), this protocol may soon vanish. SPDIF still remains a popular protocol by which to pass stereo digital signals, via lightpipe or co-axial cables.
However the ADAT protocol has gone on to enjoy a life of its own and numerous products from various manufacturers still support this format. For example, many audio interfaces, digital mixers and 8-channel preamps still use the ADAT protocol. The main reason for this is that it allows for convenient transfer of 8-channels of audio within the digital domain and offers easy expansion and connectivity between various units.
The word clock signal that is imbedded with the ADAT protocol also facilitates easy to setup and adds another significant strength to this format, as it allows users to easily sync products together. For example, expanding the number of microphone inputs on a firewire audio interface by adding an 8-input microphone preamp, or connecting a digital mixer to a PC or Mac for 16-way communication via two ADAT ports.
If you’re still struggling with these concepts (most musicians certainly do), talk to someone at Soundslive, who’ve been advising on and selling these bits of gear since they arrived on the scene.
Finally, A Note on Word Clock
Whenever two or more products are connected via any of the formats listed above (i.e. SPDIF, TDIF, AES/EBU or ADAT), there must be one word clock master, which all the other products must slave to. Note that every digital product has its own internal clock, which drives all internal processing and its converters. This factor in turn determines point in time the unit expects to process data, or a sample.
If clocks are not synchronized, units essentially take audio samples at random times and as a result may expect data when none is present. A classic symptom of this is hearing pops or cracks within audio on playback. If the clocks of connected units are synchronized, data is sampled at the correct point in time and audio playback occurs seamlessly.
For reference, all of the digital audio protocols covered within this blog actually contain word clock information within the digital audio stream, so simply setting connected units to slave to the appropriate format and sample frequency normally suffices. However, within more complex setups, users may need to use BNC word clock connections to control overall synchronization, with one device set as the master word clock (Word Clock Out) unit and all other unit synchronizing to this via their Word Clock In port.
By Andy Atkins
This piece was written Andy Atkins, a long-time contributor of articles and reviews on music gear. Andy’s written on behalf of manufacturers like Fender and Korg and for retailers such as Soundslive. He’s spent much of the past 15 years working in tech and an even longer time pretending that he’s part of a band (he is, but they only play every few years!). Find him on Google+