In my recent post about AES67 I referred to the typical path technologies take when establishing themselves:
Adoption - Standardisation - Convergence
There are many examples of widely used technologies which follow this pattern, Telephone systems are a popular example but I have also used railways as an equally valid example in the past. The thing they all have in common is a technology which is useful to a single user but which becomes more useful as more people use it, and not in a linear fashion - its usefulness grows exponentially as its user base grows. This is known as the “network effect”, sometimes referred to as Metcalfe’s Law, and it is probably most clearly illustrated by the example of online social networks. The success of a social network depends on its popularity. Facebook is popular because it is popular. It has an overwhelming advantage because the network effect reinforces its popularity in a positive feedback loop. Things can of course change (Ahem - Myspace…) but the network effect remains very powerful.
When it comes to computer network technologies, it’s hard to remember that there was a time when Ethernet was only one of several competing network systems but Ethernet has become so overwhelmingly pervasive that to most “network” and “Ethernet” are virtually synonymous. In fact the progress of Ethernet has been so unstoppable that it gave rise to the (apocryphal) Metcalfe’s Second Law - "Never bet against Ethernet"!
When talking about laws of computing it’s appropriate to mention the most famous of them all - Moore's Law. This remarkably resilient prediction on the progress of speed in computing, while it's had its wobbles, is testament to the fact that current limitations based on speed will change, and will change soon.
Given that where there is potential for networking it appears inevitable and unstoppable, it would be a brave person who would assume that they don’t need to be aware of its potential and its consequences. This is a big subject and one I will be addressing over the coming months in some detail but for now here are 5 things to consider about life in an audio networked world:
This is a network’s key strength. If you want to duplicate, share or send audio to multiple or remote locations then networks offer so many advantages over traditional point-to-point techniques that these advantages outweigh the disadvantages many times over. This is why live sound and broadcast have been comparatively early adopters of Audio over IP (AoIP) whereas post and music production have been comparatively slow on the uptake. After all, if you record in a single room then the benefits are less apparent but the disadvantages remain. However when factoring in the cumulative “network effect” when AoIP reaches critical mass and the inevitable performance increase over time as hardware becomes more capable, I think people not using AoIP will be the exception rather than the rule.
There are two (largely) distinct audiences who will need to learn about Audio over IP. Audio engineers are going to need to learn about IT infrastructure and IT specialists are going to have to learn about professional audio (and video). This presents an interesting dilemma - All of us are going to have to address just how “front loaded” our respective fields are. Try explaining what you do to someone with no knowledge of audio production in the way you would to a fellow professional, you’ll see just how much knowledge you assume. That’s exactly how it is when IT people speak to you about IT! Unless you are one of the lucky few who understand both audio and IT to a suitable level you’ll have some work to do to exist independently in the future.
3. Generalised Infrastructure
We’re all used to computers in studios. I’m (just) old enough to remember computer-free studios and the idea of having no digital equipment beyond a hardware reverb in a studio is strange indeed. Of course there are people who run all analogue studios as a choice. That is great. Old technology doesn’t just disappear when it is superseded and nor should it, but the fact that the job of specialised hardware can be done by generalised IT equipment, and the fact that any modern facility will have to have Ethernet cabling and switches installed but doesn’t have to have specialised audio and video cabling installed, combined with the low cost and convenience of Ethernet means that there is an overwhelming drive towards Ethernet in all but the smallest of new installs. Because of this, there is an uncomfortable truth in the future life of anyone working in commercial audio - your infrastructure will no longer belong to you. You will become a priority user of a shared resource and to get the best out of it you are going to have to speak to IT people.
4. Managing Delays
Networked audio is about managing delays. All digital audio systems introduce delays to signals. Sending digital audio across networks introduces further delays and the job of the audio engineer of the future will be capturing the waveform using the traditional skills of an audio engineer, but then ensuring it is delivered across the network in the shortest possible time and most importantly with related signals playing out at the same time as each other.
5. Circuit Switched vs Packet Switched
To someone who hasn’t had to deal with this before it can seem like so much hard work to fix a problem which simply doesn’t exist in an analogue system or traditional digital systems. This is because we are in effect trying to use a packet switched network as a traditional point to point system. We want the flexibility of a network without the timing issues that packetisation introduces. If you don't know what packetisation is stay tuned as that is one of the many terms we'll be de-mystifying over the coming months.
If you record in your bedroom with Pro Tools and you can't see how this affects you, you're probably right. It won't affect you. It depends on whether you want to come out of that bedroom?
Many thanks to Mark Yonge, Standards Manager at the AES for permission to use content from his address to the Audio Networking Forum, Dec 2014.