Mike Law of BCD Audio has written the following opinion piece about the proposal for multichannel audio using AES3. Law writes: AES3 is the standard for interconnection of stereo audio signals between studio equipment. It originated in parallel with the SPDIF format during the 1980s, and has evolved to allow sampling frequencies to 192 KHz stereo, and to transport multichannel compressed audio at the 48 KHz sampling frequency.
An AES standards committee now has an active project, with which I am involved, that would allow more audio signals to be carried with similar hardware to AES3. The project title is AES-X196, “Multichannel Audio in AES3’. The standard is in early draft form, as it is a new concept, and interested parties are welcome to visit http://www.aes.org/standards/meetings/init-projects/aes-x196-init.cfm or to the current project status at http://www.aes.org/standards/meetings/project-status.cfm or to the SC-02-02 meeting reports at http://www.aes.org/standards/meetings/ to access more information on the project.
AES-X196 is based upon the idea that eight 48 KHz signals could be carried across a 192 KHz link and four 48 KHz signals across a 96 KHz link. This idea is attractive, as broadcast audio has standardised on 48 KHz sample rate, but most AES3 compatible hardware is capable already of operating at the higher sample rates.
Currently in a broadcast environment, the transport of surround sound and multichannel audio requires either encoded compressed audio or multiple linear stereo links, and the new idea would simplify this.
Compressed audio requires expensive hardware and introduces considerable signal time delay. The distribution of multichannel audio across multiple stereo links is also unattractive, as many links are required and great care is needed to ensure that the correct audio signals are distributed and routed as a group. AES-X196 would solve both of these problems.
Another advantage of AES-X196 is in point-to-point links. If a single link is available between two locations that currently carries stereo, the new standard would allow that same link to carry more audio signals. This is important in existing installations, as it is often impossible to add more cable without lifting floors, digging up roads etc. Existing point-point transports ( such as ATM) might be able to pass the new signal.
An AES-X196 transmitter would be capable of operating with 2, 4 or 8 audio signals, and would be fully compatible with AES3 in 2 channel mode. Channel status information is carried with the audio, to identify the new modes. An AES-X196 receiver would detect the correct mode, and extract 2, 4 or 8 signals accordingly, that would be fully compatible with existing AES3 2 channel signals.
Completely asynchronous routers, bypass switches, would be directly compatible. Existing AES3 distribution amplifiers might be compatible , provided they can carry the higher rates, and sample rate converters if present are disabled.
Large audio routers would need to be adapted to handle the new signal. If only a few inputs and outputs were required to conform to AES-X196, this could be easy to implement. New designs could be capable of all modes on all ports, including AES3.
There might be resistance to change from the existing AES3 standard, as some of the currently installed hardware would not be directly compatible. Initially it can be imagined that AES-X196 signals would only be used where the equipment was compatible, and this in itself is useful, as in point-to-point links.
Units that combine AES3 signals into AES-X196 feeds and split out AES-X196 into AES3 signals would be available that would make the migration easy to implement anyway.
BCD intend to develop units to implement the standard that will arise from project AES-X196. These units will be based on FPGA technology, to allow for a degree of experimentation. Other parties are expected to implement practical designs too.
The applications of AES-X196 set out here are as seen by BCD and should be read as my opinion, and not the AES. Others will no doubt see other advantages and disadvantages.
