Time to close the chapter on AVB for the Pro AV industry

In early 2008, a group of top consultants and I were invited to a room. We were briefed on AVB and given a basic demonstration of the technology.

Like any group of professionals, we asked some questions. Unusually, these were not received well at all. We were not criticizing, just wanting more detail. Our hosts couldn’t understand why we weren’t automatically embracing AVB as the answer to all our prayers. We could see the benefits, but there were questions to answer too.

Over the next few years AVB has been presented as the next big thing for the ProAV industry. It was just around the corner or ‘happening now’ and a raft of product announcements were promised, including network switches from the largest vendors. ‘It’s already there in the silicon’ we were told.

The AVnu alliance was formed to promote AVB, now called AVB/TSN. The marketing of AVB was not one of engagement with our industry, more that it was coming and we would simply have to adopt it. Like any industry that spends millions on customer engagement to introduce people to its products and technology, this approach was unusual, some might say arrogant.

Since then dozens of articles have been written promoting AVB and they mostly follow a similar pattern. They highlight the great benefits such as that it is an IEEE standard, that some very large companies are associated with it, that we benefit from it being used in other industries and that it offers features such as bandwidth reservation.

The industry press followed AVB and, in deference to those large companies, often gave it equal billing with other network technologies.

Five years later, in April 2013, I published an article saying that AVB might not succeed in ProAV. I actually wrote it nine months earlier and canvassed people privately for their thoughts before making my observations public. There was an outcry from some corners of AVnu. One company devoted a web page to refuting some of my comments. One dismissed me as ‘going against the views of over 50 major companies’. One responded in a follow up in the same issue of the magazine. None of them addressed any of my concerns as to what was really holding back the success of AVB.

Some founding companies like Harman have practically withdrawn from AVB development in Pro AV. Companies such as Sennheiser announced AVB products but never shipped them as they saw the market reality and the limits of the technology. After ten years, only six manufacturers ship AVB certified audio products.

The fact is that after all this time, the promise of AVB has not materialised for the Pro AV industry. The adoption simply isn’t there and the reasons for that are because the fundamental issues to make it viable still haven’t been solved.

In 2016 Cisco announced that it would ship AVB-enabled switches, and some in AVnu hailed this as the moment it could become mainstream. My white paper a year earlier explained how such a move was unlikely to make a difference and there has been no significant increase in the use of AVB in Pro AV since these switches went on sale. 15 months later, those switches remain non-certified by AVnu. Meanwhile the rest of the networked audio industry has been growing at a colossal rate. Investor details given from Audinate’s (the inventors of Dante) recent IPO gave some further insight into the size of the market. With expected revenue this year around $15m, that’s a huge amount of networking going into products. There are currently over 1,000 shipping products using Dante and over 160 using RAVENNA against just 27 that use AVB. With each of these three initiatives around a decade old, AVB is the clear laggard.

RH Consulting’s research found most uses of AVB were not multi-manufacturer solutions. They are mostly single manufacturer sales, sold because of the benefits offered by each product. My view is that these products would sell anyway because they are very good products, not because they use a particular protocol.

Enthusiasts argue that AVB is part of what makes their solution superior, but even if true, this has to be set against the huge limit in product choice and interoperability over choosing another protocol. At the time of writing, there is no AVB-certified mixer, only four loudspeakers, five amplifiers (possibly seven), no microphones, no intercom, no audio player/recorder. - source AVnu. In other words, no useful choice for the vast majority of projects.

There are no cross-manufacturer software tools and utilities such as Dante Controller, Dante Via or Merging’s ANEMAN.

AVNnu argues that open standards are the answer and some members even warn of the 'dangers'of proprietary technology. However, our research put standards seventh in what users look for when choosing a networking protocol.

Respondents preferred things like choice of product or the expectation that the protocol might be around in a few years. The simple fact is that most people just want to get the job done using technology that works.

It appears that de jure standards don’t really matter to people, as long there is technology that does the job and it is licensed fairly.

Do users even know what is under the hood of most equipment? Some well-known vendors provide complete OEM amplifier modules to dozens of professional audio manufacturers with no fanfare and mostly without the knowledge of the customer.

For technical or commercial reasons many audio products contain some technology either built or developed elsewhere. Private organisations can invest in a technology and are more agile than any standards committee. Audinate’s success demonstrates that an investment lead approach works.

If you or your clients don’t like the proprietary approach, AVB is not the only open option. You could freely adopt RAVENNA, it’s all published, or choose to pay for support and naming rights or buy interface cards from several suppliers. AES67 has been published for several years and is an open standard.

No one can predict the future, and perhaps AVB will, in many years, provide some underlying benefit for our industry, but the work on AVB from audio companies is nowhere near as vigorous as some would like you to think. Currently 14 audio related manufacturers are members of the AVnu alliance, with six actually shipping products and we know of at least 11 audio companies who are no longer involved. (See sidebar)

Most importantly, the absolute reliance on generally more expensive, AVB-aware network switches, which have barely been adopted by the IT industry, is likely to take a very long time to overcome. Even when using AVB aware switches there are still the same firmware incompatibilities and difficulties that we all sometimes encounter when deploying networked audio.

AES67 provides interoperability between a number of different audio protocols such as Dante, RAVENNA and Q-Lan, but alas not AVB. There are already over 200 AES67-compliant products and this is growing by the month. Furthermore, AES67 has been picked as the audio element of SMPTE ST2110, the new video-over-IP standard for the broadcast industry.

For all the merits of AVB, there are at least as many other features and benefits in Dante, RAVENNA and AES67 with many more engineers putting time and effort into those and related initiatives.

In my 2015 White Paper, on the development of audio networking, I argued that AVB was also being promoted at the wrong time on the adoption curve as we were still in the integrated phase. However, as we now seem to be approaching the modular phase, where standards come into play, it doesn’t seem at all that AVB is the standard users are gravitating to. It appears that people are buying Integrated networking solutions using Dante, RAVENNA and others and considering AES67 as the overall modular platform - their safety net perhaps?

It’s time to close the door on this. The real-world deployment of interoperable multi-manufacturer AVB solutions in Pro AV is a statistical zero. It is misleading if the industry press to gives equal status to a technology with a tiny and fragile market share. To give further perspective, there are still fewer AVB products on sale than Cobranet, even though support for the latter ceased before AVB was even invented. More products use AES 50 than AVB, but they don’t get the same column inches.

Pro AV manufacturers that support AVB will continue to put it to good use for some projects but these should be considered as manufacturer-specific solutions. Customers need to be aware of the limitation in choice and industry compatibility, matched against the technical benefit of the solution.

I would also ask those manufacturers, is your AVB development work along with the certification time and effort worth it against potential lost sales because you are not compatible with much else? Is the requirement for your products not to contain proprietary technology from others so critical to your success as an audio manufacturer? Does AVB really make such a fundamental difference to product performance and customer satisfaction?

It is wrong for AVnu to ask the audio industry to be at the vanguard of development. Millions of dollars have already been invested in the development of AVB audio products and it is virtually no further forward than when it was first demonstrated to me ten years ago.

AVnu also promotes the use of AVB in consumer, industrial and automotive sectors. It may indeed be beneficial for those industries, but for the next few years at least, it won’t be commercially viable and barely technically relevant to our industry. I have yet to see one instance where work on AVB in other sectors has benefitted a Pro AV project. I have also not yet seen AVB products from another industry being used on the same network as a Pro AV solution.

For Pro AV, any meaningful growth in AVB is several product cycles away because it is significantly dependent on the IT industry choosing whether or not to seriously adopt it. I’m reluctant to put a date on it, but another 10 years would be a safe bet.

In the meantime, the insular approach taken by those AVB manufacturers will have cost them sales because of their lack of compatibility. Furthermore, they could be leading the industry in other ways and helping to solve the many other problems with networking that continue to cause us headaches – it’s not just about audio transport. If or when AVB is ever ready, we can all jump to take advantage of it, just like we have always when other technologies become viable.

Being at the leading edge of audio transport gives you very little competitive advantage when the real key to the success of projects that connect over IP is compatibility with others.

Standards are simply agreements for connections between modules,
they are not the solution in themselves.

The story of AVB in our industry makes for an interesting case study. There are many lessons to be learned. Within ProAV, the project has been a catastrophic miscalculation over the importance of standards and specific technical features, against how projects are actually designed, sold and implemented. On the marketing side, compare AVB’s clumsiness against the sublime approach from Kevin Gross introducing AES67 to our industry in such a favourable manner.

It would take another article to examine the many aspects of AVB’s failure in our industry over the past decade but it’s time to face facts.

Let’s not waste any more time and money on something that is always ‘just around the corner’, when there is plenty of other audio networking development work that can benefit us all now.

Further reading: