An Interview With Michael "Catfish" Dosch

We cornered Michael Dosch, President of Axia, to talk about Telos' new company. 

Mike, thanks for taking time for this interview. Let’s cut right to the chase: tell us about Axia.

Axia is the newly formed studio audio division of Telos. We specialize in digital audio routing, mixing and distribution systems for broadcast and other pro-audio applications. Every Axia product is networked using Ethernet – including audio, logic, control and program associated data, or “PAD”. Devices connect together using standard Ethernet cables; audio and control routing is accomplished with off-the-shelf Ethernet switches. The foundational technology in all Axia products is Livewire, a patent-pending protocol that enables high-reliability, low-delay uncompressed digital audio over Ethernet.

Why did Telos form a new company for this?

I'd like to tell you this was the plan all along, but truthfully, we were surprised by the overwhelmingly positive client-response to the Livewire concept. We realized that in order to best serve the marketplace, we'd need to apply focused energy to this new area. And we felt the best way to do that was to spin up a new company. Axia is still part of Telos, much the same way Omnia is part of Telos… we share certain company resources, values, and technology, yet each company has its own style and expertise. We find this structure helps us remain well connected with our clients, even as the company grows. Each company does whatever is necessary to ensure the success of its own clients.

So what did you do before grabbing the top job at Axia?

For the past few years, I’ve been leading research and development for Telos and Omnia, with a particular emphasis on incubating the Livewire program from concept to product. Prior to that, I was managing director of Telos. And before joining Telos in 1999, I was an executive with PR&E (Pacific Research & Engineering), the console company. Funny thing about broadcast consoles, once you design a few, they're like an addiction...

Axia is the product of a lot of hard work from many brilliant and committed Telos (and Omnia) people. It's a great honor to have been selected to lead this new division.

Tell us about the motivation behind the development of this technology.

This has been in the lab for quite some time. Our work with high-end broadcasters often inspires great product ideas and exposes us to new opportunities. One area in particular captured our imaginations repeatedly: how could we improve broadcast operations and enable new functions by tightly integrating consoles, hybrids, codecs and PC audio systems? But one huge obstacle kept surfacing: there was no simple way to connect audio and control between studio devices. Every studio we visited had a major mess of wires and connectors in just about every imaginable format and style.

One day, Steve [Church, founder & CEO of Telos] saw a demonstration of a Cisco VoIP system and had a ‘light bulb’ moment: perhaps we could adapt the same technology that allowed low-latency voice traffic to work reliably on an Ethernet for use with PCM digital audio. If we could, a broadcast plant could be easily built around Ethernet. This was the origin of Livewire.

The ‘vision’ part was simple: Ethernet connects everything. Two devices connect with a single link of CAT-6 cable. Multiple devices connect through a switching Ethernet hub. The same wire conveys audio, logic and control messaging, program associated data (PAD), and even general IP traffic. Clean, simple, elegant. Of course, the development was a bit harder than we had imagined.

I’ll bet. Tell us about the challenges in development and how you addressed them.

Well, we could have made this a lot easier on ourselves if we had lowered our standards a bit! Audio over Ethernet is not the hard part. Big buffers can cover a variety of network-oriented problems, from lost packets to congestion. But buffers introduce delay, and we wanted a system that could deliver glitchless audio streams with as little delay as possible. In fact, our latency target was under 1 millisecond per network hop.

And that wasn't the only place we made things hard on ourselves. We also wanted to build a network without any single points of failure. And we wanted to be able to route any audio source to any or all audio destinations without limitation. Oh, and we wanted to allow normal data traffic onto the same network as the audio while guaranteeing the audio would always remain solid and instant.

That sounds like a pretty tall order.

Yes, we set some very high standards for the technology. But our team was up for the task, and the resulting architecture is robust and distributed. It’s also very low-latency to allow mic-to-headphone monitoring without discernable delay. And Livewire can carry audio and other forms of traffic all on the same net.

So how did you overcome all of the technical challenges?

You can find out about that by visiting the "Tech" portion of our website, but I will say that this work likely represents the most ambitious R&D project ever attempted within the radio equipment industry; it’s taken about 40 man-years of development so far with teams in both the US and Europe, and we're still cranking out the ideas and product variations.

How is the Axia system different from other, similar systems on the market?

First let me say that there is really nothing else like Livewire, in terms of using a non-proprietary network to convey audio and control in a broadcast plant. There are a few technologies out there that use Ethernet to carry audio, but they are much higher latency, less flexible, and not commercialized for broadcast use. The networked architecture provides Axia products with unlimited scalability and dramatically reduced cost compared with any other approach.

Regarding similar product types, Axia products consolidate and replace a wide variety of studio products, such as consoles, routing switchers, sound cards, distribution amps, monitor/source selectors, logic interfaces, et cetera. Axia products usually represent a significant cost savings over the alternatives. For example, a client recently asked us to quote on a 64x64 routing switcher configuration and was amazed to find that we were about half the price of his other quotes. By taking advantage of computer networking industry scale, we're able to dramatically reduce costs even while increasing capabilities far beyond more expensive solutions.

And if saving money on hardware isn't enough, clients should also consider the cost of connectors, cables, patch bays, punch blocks, and the labor to install it all. These costs can really add up. Our products all interconnect with standard, ubiquitous Ethernet cabling; it can be purchased off the shelf, pre-made and molded, or clients can crimp their own. Not only is it less expensive, but it's much faster than the alternatives too – a huge facility can be wired in days instead of weeks.

Where is Axia headed? How do you envision the evolution of the system over the next few years?

We have a number of very exciting ideas working their way through the lab. A modular control surface is in development now that brings the best ideas of SmartSurface into a scalable surface for smaller (and larger) applications. We are soon to announce partnerships with software developers who will have versions of their audio applications that can communicate audio and control directly to an Axia network, eliminating the need for soundcards. We have some great features planned for the DSP Studio Mix Engine software to further enhance studio operations. And our friends at Telos and Omnia are planning some cool new products to also interface directly with the network. The future is bright indeed.

Okay, one last question. How did you get the nickname Catfish?

You think I was always this respectable? When I was young and foolish and running with a biker gang, I once...

Enough. Sorry I asked.

More news and interviews...