How did we end up with so many ‘alternatives’ in AVoIP? 

I’m old enough to remember when ‘High Definition’ video was introduced – we all said, “thank goodness, now we can have a single standard and eliminate that analogue ‘nightmare’ of NTSC vs PAL vs SECAM” (and indeed the various subsets of each that manifested in the broadcast space – e.g. PAL-B, PAL-G, PAL-I, …). And before long there were, if I recall correctly, around 20 versions of HD. 

So much for a ‘single standard’.  

What typically happens in any high-tech environment is that different manufacturers and groups take ‘a good idea’, work on it in parallel but independently, and ultimately propose different things. Then they seek to defend their investment. And we end up with what many call ‘protocol wars’, as they each seek to outdo each other – by spec waving or market-segment share comparisons. 

And that’s exactly what has happened with AVoIP too – many schemes have been introduced, and most have been successful by some measure. Some have specialized, some have stumbled, some have fractured. Whilst some have sought to broaden their appeal, others have remained resolutely proprietary. An interesting side note is that some have made domination of the control plane the main part of their business strategy – important yes, but fundamentally unrelated to video performance. 

So that’s the landscape. It’s tricky for the user to pick something that won’t leave them at significant risk of being in a proprietary dead-end.  

IPMX, the Internet Protocol Media Experience, is set to change that. But it isn’t ‘just another in a long line’ of AVoIP methodologies. 

The difference makes the difference. 

What is different about the IPMX approach is that it started from a true standard for moving audio and video over IP – SMPTE ST2110 – which has been around since 2017 and is accepted by the top tier of the international broadcast community. Let’s face it, that industry knows a lot about different video and audio flavours globally, so that acceptance counts for a lot.  

It is not just SPMTE behind IPMX. The Audio Engineering Society (AES), the Video Services Forum (VSF), the Advanced Media Workflow Association (AMWA) and the European Broadcasting Union (EBU) are all supporting it; they are all bodies without allegiance to a particular manufacturer. And all of them want a low latency, high quality, super reliable method of moving audio, video, and metadata across networks. 

 How does the difference show? 

The ST2110 evolution started from the opposite end of viewer experience to where all the other AVoIP folks started. It started from ‘our best video is uncompressed, we have to support that, so what infrastructure do we need?’. Everyone else started from ‘let’s start with the infrastructure that we’ve got, and then see what we can do with it’. And to be fair, low bandwidth (i.e. in the 50-200Mbps ball-park) is where the mass market (consumer and video conferencing) is, so it’s no surprise that many approaches started there. That’s the bottom of the ‘streaming market’ pyramid; highest performance is of course at the top of the pyramid.  

It’s a truism that it’s much easier to go downhill than up a slope – so extending an approach that was designed for the bottom of the pyramid will struggle as it meets increasingly difficult tasks. A glaring example is that none of the other AVoIP approaches can do uncompressed 4K60 video at 8-bit colour (and certainly not at the 10-bit demanded by HDR). 

ST2110 brings a well-developed system with proven scalability across increasing numbers of users and existing and demonstrable inter-brand interoperability. IPMX inherits those features and extends ST2110 into the world of professional AV by adding an array of AV specific features like EDID, HDCP, lower spec network architectures, different picture shape factors, different compression regimes, and a myriad of connectivity that does not exist in the broadcast space. 

And where is the future? 

With the advent of Artificial Intelligence, it would take a brave individual to predict what users will want or need to do in a year’s time. That makes it tricky to be precise about what will be expected of an AVoIP system in that timescale. 'Minimum viable system' seems distinctly risky. Flexibility sounds like a very good plan! 

One thing we can say for sure is that more bandwidth is becoming available. We see around us the bandwidth to our homes increasing from a few tens of Mbps to over 1Gbps. This is not the final chapter of domestic connectivity. Cable modems are driving domestic bandwidth ever higher. 

The same trend applies in corporate LANs too. If you asked a typical AV integrator what network they suggest for AVoIP to their enterprise clients, many today would say ‘1Gbps edge, 10Gbps core’. They think they are at the leading edge. At one point they were. Increasingly, that description is looking ‘limp’ in comparison to what those same clients have in their homes and what is used elsewhere in the client's own organization. I noted the other day, a city-suburb rubbish incineration plant that had a 10G network for running their industrial process control! 

If you buy a ‘commercial’ 1Gbps network switch today, the odds are good that it is ‘multi-gig’ and can also do 2.5Gbps. 10Gbps switches are falling in price. Many enterprises today – for example the broadcasters mentioned earlier – have 25Gbps ‘to the edge’ at least for some users. The 'core' network within enterprise is expanding too – 40G and 100G switches are no longer the preserve of the network provider. More bandwidth means many things, but for AV folks it means one or more of:  

• more endpoints,  
• more streams, 
• better-quality video. 

So IPMX, by supporting all of those bandwidths, is prepared for the future. Including a future that probably includes 8K AVoIP but doesn’t shed the need for lower resolutions. Mixed needs on the same network. And IPMX can scale to the inevitably increasing size of corporate networks. 

By supporting multiple compression types from the offset, IPMX has the flexibility to cope with user-needs scenarios we haven’t thought of yet. And the near future is going to have some of those scenarios. 

In a nutshell  

IPMX is not like any of the preceding approaches, and that is why IPMX is the future of AVoIP.

IPMX is: 

• tractable, it is based on published standards and fully open. There is no ‘entry fee’, subscription or obligation to use any proprietary chip, code or software.  
• scalable, increasing the number of endpoints does not cause its performance to crumble. 
• extendable, it is as performant as you need – no other AVoIP scheme is able to support such a wide range of user needs, such a wide range of network bandwidths.   
• interoperable, brands can stream to each other – including when HDCP is used. Integrators can design systems across manufacturers – choosing ‘best in class’ or some other criteria to select the component hardware parts right for the specific needs of the client. 
• adaptable, the IPMX conduit supports gateways to and from other approaches. IPMX to HDBaseT,   to SDI, to HDMI, to NDI? to DanteAV? to others? 
• controllable, the NMOS API is common to all brands which means that manufacturer A’s software can control manufacturer B’s hardware, and vice versa.  
• flexible, whether the network is tightly timed using Precision Timing Protocol (PTP) or is a low-cost asynchronous network, IPMX devices are able to function. 
• available: the others are not useful without this one! The future starts now.