You are here

Q. Why so many digital audio formats, and what are they for?

Published February 2008

There are so many ways of connecting equipment these days, such as S/PDIF, ADAT, AES-EBU and MADI, not forgetting good old analogue. What are all the digital connections for?

SOS Forum post

Technical Editor Hugh Robjohns replies: Back in the '80s and '90s, there were dozens of manufacturer-specific digital interfaces, such as Yamaha Y1 and Y2, Melco, TDIF, ADAT, SDIF2, R-Bus and many more, and none of them could be connected together. It was a complete nightmare!

In order to make 'going digital' a practical option, the Audio Engineering Society (AES) and the European Broadcast Union (EBU) put their collective heads together and came up with two generic, open-source digital interfaces: one for stereo and one for multi-channel audio, the latter of which was originally intended to link multitrack recorders to large consoles. These were called AES-EBU (now more commonly referred to by the AES standards document number, AES3), and MADI (Multi-channel Audio Digital Interface).

A comparison chart showing different types of digital audio protocols. A comparison chart showing different types of digital audio protocols. AES3 was a bodge in the engineering sense, but the use of apparently standard mic cables and connectors made it a familiar-looking interface that reduced the fear and cost of 'going digital'.

The original MADI specification essentially carried 56 channels, made up of 28 AES3 stereo pairs transmitted serially. A later revision called MADI-X catered for 64 channels and is in widespread use today in applications such as connecting stage boxes to digital desks, linking Outside Broadcast trucks, and connecting the infrastructure in digital studio complexes.

Today, AES3 is the preferred interface format for professional stereo applications, although there is a noticeable trend towards the AES3-id format which uses unbalanced BNC connectors and 75Ω video cables rather than balanced XLR connectors and 110Ω cables. AES3-id is a much better-engineered interface, and is far more space-efficient. AES3 digits run with a fundamental frequency of 1.5MHz, with strong harmonics all the way up to 10MHz and more. Video cable and connectors are far better suited to handling those kinds of frequencies than manky old mic cables, and AES3-id works more reliably over greater distances, with less jitter as a result.

Having designed a very versatile and effective digital interface, and all the hardware chips to drive and receive it, Sony and Philips took the opportunity to use the same thing for domestic applications and called it S/PDIF, with coaxial (phono) and Toslink (optical) interfaces. The nitty gritty of the auxiliary information and metadata carried by AES3 and S/PDIF are slightly different, but the basic structure and audio formatting are identical, and you can normally interconnect AES3 and S/PDIF with little problem. S/PDIF is electrically almost identical to AES3-id.

In terms of the actual interface properties, AES3 runs balanced signals with a nominal 7V peak-to-peak swing, feeding a receiver with a minimum sensitivity of 200mV. Because the signal starts so big, it tends to go a long way (more than 100 metres) even on nasty mic cables. Put it into decent low-capacitance 110Ω cable and it will easily travel 300 metres. AES3-id is unbalanced and starts at 1V. The receiver sensitivity is the same 200mV, while S/PDIF is also unbalanced and starts at about 0.5V. The receiver sensitivity is also 200mV. The lower starting voltage is why S/PDIF doesn't travel very far.

There are several eight-channel AES3 interfaces, most using 25-pin D-Sub connectors. Sadly, there are lots of different incompatible pin-outs: Yamaha, Tascam, Genex and Euphonix, to name a few. But the Yamaha and Tascam formats are the most prevalent.

Yet another variant of AES3 is called AES42. This still uses XLRs and balanced cable, and the data is encoded in exactly the same way as AES3, but it is intended for carrying the output of digital microphones. The critical difference is that an AES42 input socket provides 10 Volts of phantom power, and that power is modulated in a specific way to allow remote control and digital clocking of the microphone. It is an agreed format that has been adopted by Neumann, Sennheiser, Schoeps and others, and will start becoming a common feature on digital consoles and professional recording interfaces.

In terms of other digital interfaces, Tascam's TDIF is virtually dead, but ADAT is alive and well and in widespread use. ADAT uses the same Toslink fibres and connectors as S/PDIF, but with a different data-stream structure to carry eight channels.

In addition to MADI for high numbers of channels, we also now have the new AES50 SuperMAC and HyperMAC audio networking interfaces (originally developed by Sony Oxford in the UK and now owned by Telex Communications under the Klark Teknik brand).

SuperMAC provides 48 channels bi-directionally over Cat 5 cable, while HyperMAC provides up to 384 channels bi-directionally over Cat 5 or Cat 6 or fibre. The signal format includes embedded clocks in the same way that AES3 does. 

Published February 2008

Issue navigator