<img height="1" width="1" alt="" style="display:none" src="https://www.facebook.com/tr?id=311512539006924&amp;ev=NoScript">
  • Telos Alliance
  • Telos
  • Omnia
  • Axia
  • Linear Acoustics
  • twntyfive-seven
  • Minnetonka Audio

Blog Central

Sixteen Streams and What do You Get?

Posted by Kirk Harnack [TWiRT] on Mar 31, 2014 10:09:00 AM

Find me on:

TWiRT 206Doug Rowe is an engineer at Minnesota Public Radio. His title is one that we’ll see more often: Media Production Systems Manager. As radio stations, groups, and networks depend on digital media distribution, more engineers like Doug are needed; not only to keep everything working, but to develop the infrastructure for the workflows needed by talent and content creators.

Doug talks with Kirk Harnack and Chris Tarr, helping us understand the new responsibilities involved at a highly-connected, digitally-delivered network operation.




Watch the Video!

Read the Transcript!

Kirk Harnack: This Week in Radio Tech Episode 206 is brought to you by Direct Current the tech newsletter from the engineers at the Telos Alliance. Direct Current brings you technical articles, tips, and information you can use today. Direct Current is free, useful, and engineer approved. Visit Telosalliance.com/subscribe.

Doug Rowe is an engineer at Minnesota Public Radio and his title is one we'll see more often, Media Production Systems Manager. What does that all mean? Well Chris Tarr and I wanted to find out so we got Doug to tell us and you'll want to understand what he does as the radio industry will need more people with his skills.

Hey, welcome in, it's time for This Week in Radio Tech. I'm Kirk Harnack, your host. Glad that you're with us, this is our 206th show. Our show is brought to you by Axia and the Axia Element. They're my employer and I'm just very grateful that they also want to sponsor This Week in Radio Tech. This is the show where we talk about anything with broadcast engineering relate, from the microphone the announcer sits behind to the light bulb at the top of the tower, and now so many new technologies.

We're going to spend a bunch of time today talking about streaming and managing lots of streaming, and how important that streaming has become. We'll get into it with Doug Rowe our guest from Minnesota Public Radio.

Let me bring in right now, though, our co-host for the show, Christopher Tarr, the Ninjaneer, the Geek Jedi from Mukwonago, Wisconsin.

Hey, Chris, I'm glad you're here . . . and you're quiet.

[silence 01:39 to 01:43]

Is it just me?

Chris Tarr: Nope, it's me too, sorry. I was trying to be polite by muting, so that when I coughed you wouldn't hear it. I've got to remember to un-mute it when I'm done.

So, as I was saying, among those other things, I'm generally a nice guy. I'm the director of radio operations of Radio Milwaukee in Milwaukee, a Radio Guide Magazine writer, cofounder of the SocietyBroadcastEngineering.info, a bunch of kids, a bunch of radio stations. I never sleep, help me.

Kirk: Chris, you have a new set.

Chris: I'm home from work today, because I have this bronchial thing going on, so I'm actually in my living room, I'm sitting at the kitchen table.

Kirk: Ah, okay. Usually it's a little bit too rowdy around there for you to be at the kitchen table.

Chris: Right. When we were doing the show at night I had to be in my bedroom, because the kids were running around, but right now it's just me and over in the corner my dog, so we're good.

Kirk: Aww, have we seen your dog?

Chris: No, you've not seen my dog.

Kirk: Do you ever want to play show and tell?

Chris: Well here, let me see if I can lift it up here. He's actually behind me on the couch, if you look.

Kirk: Almost, almost, almost.

Chris: Aw, there he is, up on the couch behind me. See his nose there?

Right behind my chair, he's white so he kind of blends in. There you go, there's my dog sitting there pining for me because I'm not around. I'm talking and not petting him.

Kirk: I've got a couple of cats and there's no way I can get them on the show. They don't mind, it's the nature of them.

Chris: Yeah, I have a cat, too, and she's around somewhere. I don't know where she's at.

Kirk: Does yours mind what you say?

[silence] I guess that's a no.

I just wonder if Chris is still hearing us or not.

All right, let's go ahead and move to our guest. Our guest is Doug Rowe and Doug is with Minnesota Public Radio. Let's bring Doug in.

Hi, Doug, how are you doing?

Doug Rowe: Hi, good, thanks, Kirk, how are you doing?

Kirk: Good. Boy, it seems like and I met some years ago, didn't we? There at your place of business.

Doug: Yeah, so here at NPR we actually were one of the first big Axia installations, roughly 2005, early 2006. We were, for a little while, one of the largest Axia networks, I think, that Axia even had around. Now I don't know whether there's any folks who've eclipsed us. But, at this point in time we're still running a full facility with about four primary services coming out of our building as well as doing Axia installs at some of our sister stations at Southern California Public Radio and Axia installs at Figueroa Courtyard, which is Marketplace Productions out in L.A. So, lots of Axia experience, lots of gear throughout our facilities here, so it's a lot of fun to get to take care of and spend time with.

Kirk: One thing I want to make sure that folks understand we're going to talk about, it's not an Axia show, although that will play a little bit into it, Doug's title, although I'm not sure it would quite fit in the lower third. Doug, your title is Media Production Systems Manager. Media Production Systems Manager. That's a big title, but we're talking about it before the show and it seems very appropriate for the kind of work that you do. The kind of work you do is not something that most radio stations have yet, or have somebody in your position. So, let's talk a little bit about what you do at Minnesota Public Radio and American Public Media there in your building in Saint Paul.

Doug: Sure, so I started out as a more traditional broadcast engineer, transmitter sites, studio facility maintenance, the normal list of things that folks do. As a large organization we always also had a full time, or still do have a full time corporate IT team. One of our challenges throughout the years at NPR and APMG has been balancing computer based systems for broadcast versus computer based systems for general computing. Whether it be accounting, word processing, or whatever else. So, it's always kind of had this gap for us where you never know where things should fall into. Should the traditional IT team buy PCs for the studios? If so, should they be the ones to maintain the software around there? Where is that line? You say, "Oh, well it's engineering takes care of the broadcast software, IT takes care of the PC."

"Well, what if it's a soundcard driver issue?"

It's always been kind of a gray area, so in an effort to better cover this type of technology a couple of years ago NPR started a media productions systems department. We're basically bridging the gap. At this point in time Tier 2 for Axia, especially when it comes to networking working closely with our full time network engineer that works in the corporate IT team side. At the same time we're the primary administrators and instigators for our Dalet digital media system software that we're in the middle of a project right now, primary administrators for a number of ENCO systems around our facility, around the network, working closely with the engineers. Then of course working closely with the IT teams for the infrastructure and administration of all of our web streams, the interconnectivity to those web streams to our stream guys and ads whizzes as we been trying to grow that side of the business. Then of course the day to day support of all the folks using Pro Tools and Dalet, and whatever else they can dig off the Internet for use in audio production and newsroom workflows. As well as the same thing for our classical and rock and roll services, making sure they have all the content that they need in the formats that they need it and are able to get it to the places they want it to go.

So, it's a very fun position.

Kirk: Some of the stuff you're describing is kind of mind blowing for me. I've been a work-a -day engineer. The biggest place I've worked at, I guess, is where we ran eight radio stations years ago out of Cleveland, Mississippi, just small stations, we ran them out of one control room. What you're describing to me is just, I mean, I've been there, it's big. You've got dozens and dozens and dozens or people with workstations for doing editing. You distribute for the different networks. You got, as you said, four program feeds that go out of your place, live, to transmitters and then you've got other distribution networks, and web streams, and so forth. It's really amazing how much stuff goes on at a place as big as Minnesota Public Radio and American Public Media there.

Boy, maybe you can just talk a little bit about that, about the scope of what you guys do as far as generating content and then delivering it.

Doug: Sure, absolutely. Well, as everything has changed with the Internet and the ability for downloading and streaming, and a variety of other devices on which you can receive content, our corporate mentality has turned from being a radio station or a network of radio stations, to a content producer and media distribution network. So obviously the Minnesota Radio FM broadcast is a big component of that, and still technically our largest component.

As far as our mentality goes it is still no more important, or I should say, the streams and other vehicles for distribution are no less important than the FM broadcast. A double digit percentage of our listeners are tuning into web streams all around the world. Daily there's hundreds, if not thousands, of downloads of our content via podcasts, via direct downloads, via third-party services like iHeartRadio or Stitcher, or whomever else. There's all kinds of folks are attempted to be affiliated, or even the Apple iTunes store.

So our goal is not just to support those things, but to make sure that there is a high level of reliability, at least as a high level of reliability as the expectation has always been for the FM transmitter sites.

As a company, Minnesota Public Radio, as I said, is the primary organization, but we're actually a parent of American Public Media. As a function of that relationship between the two organizations, we have just as much emphasis on content distribution for our own services as we do for other stations that may carry our Prairie Home Companion or other programs in the same respect and vein that National Public Radio and Public Radio International function at this point in time.

So we have tried to take the mentality that we need to approach these streams as something that's just as important and need to be just as reliable as the day to day broadcast. Both as a respect for our content production and for our listeners who expect those things to be there. Not to have that could in many cases disillusion the listener, or have somebody not want to find our content, because they don't think it's accessible.

Kirk: I guess not only desiring to have the reliability on the streams, but the interesting thing about streams, and you guys have all this ability to make content. In the terrestrial broadcast world you're limited by the number of transmitters and the number of licenses that you have and you can transmit with. With the Internet and streaming you're pretty well unlimited. If you can justify another stream with some niche of programming that people will listen to I guess you just make it, right?

Doug: Absolutely. It's interesting when I think about the history. I think the supplemental channels in HD radio for us were actually the beginning driver of that, where we started to say, "Well, what else could we put out there?" Even there we were limited to only one or two additional services per station. Well, since that point in time we've moved to a point where we've started looking at the web stream as being the initiation for many services, that sometimes even have grown then into an FM broadcast. The popularity of a web stream changing what we actually put on the radio at some point in the future.

I've seen that driven the most with The Current, which is our rock station. The Current is a kind of eclectic mix of, I don't know the best way to describe it. It's kind of a mix of styles of music with a lot of emphasis on local Minnesota artists. But, classical is definitely getting to that, and of course news content. The idea of breaking news, especially in a global oriented perspective, has become very important. We, in the twin cities here, have populations of people who are very interested as to what's happening in Egypt, very interested in what's happening in Somalia, because it hits close to home even though they may 3,000 miles away.

So for us it's both a responsibility to our listeners and then, of course, an opportunity for us to get our content out to folks, and to really push the Minnesota Public Radio, American Public Media name, and bring content that people enjoy.

Kirk: This is a show were we do talk about tech, so I just want to try to get that set up.

Doug: Sure.

Kirk: Also, I'm glad that Chris Tarr was able to join us today for being a co-host on the show. I guess Wisconsin-Minnesota you guys are a little bit rivals, but Chris also has been and is in the world of public community radio, so I figured that you two might have some things to talk about, and Chris might be able to ask you questions from a perspective that would go past me.

So, Chris, are you back with us? Is everything okay?

Chris: I am right here, yes. My Internet went stupid for a little bit there, but we're good.

Kirk: Good deal. Well, I'm sure you've got a couple of things you want to ask Doug about and certainly we will want to get into some of the technical things that Doug does. Maybe he's got some lessons to teach us, maybe there's some techniques that he can tell about that he's found particularly helpful when you're managing streaming and lots of streaming.

Chris, you got something you want to talk to Doug about?

Chris: Actually, no. But, yeah, I've been following along. It's funny, The Current is actually kind of a sister station to Radio Milwaukee, we do a lot of content sharing and things like that. We actually have a very similar set up in terms of ENCO, with streaming, and Axia, and things like that, so actually interested in hearing how he's using that.

I think one thing that I would add to what Doug was saying, he hopefully can expand on that a little bit, but when you're talking about having the content available when listeners what to go to get it. I think that's a very important thing that Public Radio does, because a good part of our support comes from our listeners. Directly, we cut out the middle man, and they hand us their money directly. One of the important things about my job and Doug's job is to make sure that that content is available. That things are functioning and they're working sufficiently so that that content gets into the hands of the end users. So I'll be interested to kind of hear more about some of the steps he's taking to do that, because that is a very important part of what we do in public media, is making sure that the content is available to the people who are paying us to put it together for them.

Doug: Sure, absolutely. Sorry, go ahead.

Kirk: Go ahead, Doug. No, you go right ahead and answer that, I've got some more coming up.

Doug: Yeah. So, for us it's in the last couple of years, especially as the emphasis on these alternate means of delivery have been pushed within our organization. We've had a lot of focus on redundancy, and a lot of focus on, not so much disaster recovery, but backup systems and the ability to find alternate paths or alternate ways to maintain content to our listeners in the end. So from a streaming perspective that both is internally within the facility and externally with our contracted streaming vendor, making sure that we have high availability, redundancy, load balancing.

For us it's really become a focus of just investing in the technology and the hardware to do that. It's not that the technology hasn't existed it's just that previously I think the expense has been hard to justify. Now, for us instead, it's become something that to us has become very important, and the expectation that, just like a FM station, little or no downtime. Even though it's just a "web stream" it's no longer just a web stream. It's another service.

Kirk: Doug, earlier you mentioned that streams you're running have, you said, double digit listenership, maybe in terms of your total listenership. What math did you determine when the streams became important enough that you would make sure that they're robust and they that you would pay more attention, not just the red headed step child, but you'd actually pay attention to them?

Doug: Sure. I'll say, if somebody went and fact-checked me I'm guessing we're getting close to that. I don't know the exact number and our ratios, but if one of our folks here at NPR tunes in and says, "Wait a minute."

I'll say, I'm guestimating on it, but I know for us, part of what has helped is the Arbitron PPM aspect of things, so that's giving us some significant number readouts or the ability to compare across services or across different aspects. APMG, as a large organization, we're lucky enough that we actually have a department here that's dedicated to that type of business analysis and data analysis and is able to put that information together for our board, and for our directors, and administrative staff, to be able to actually influence those decisions about where the money should go?

I mean, if the debate was, buy a new transmitter because one is off the air or buy streaming gear. The transmitter still would probably come first. But as we then look at hardening infrastructure that's definitely something that keeps coming up over and over as a content distribution company, not just a radio station.

Chris: Now, Doug, I'm now 18 months into public media coming from the commercial world, are you finding that expenditures are really, in a good way, expenditures are really moving towards equipment and capital in your facility as well? I've noticed recently, like for example, I've been looking at doing an ENCO upgrade and it didn't take a lot, it didn't take a real hard sell to get them to make that change. But I'm finding that, and I'm seeing this at stations across the country, that all of a sudden now there's been, in a good way, a focus put on the technology to created the content. Whereas before, well you know, if you had a 15 year old console,

"Well, it's still working." But now all of a sudden it's "we need to have these newer technologies because we all this content that needs to be created" and they've become a lot less hesitant to invest in that. At least that's kind of what I'm finding. Are you seeing the same thing?

Doug: From my experience it's definitely very capital driven. That's one of the things even when I first started at NPR that I noticed, is operational expenses probably stay pretty steady, at least from the technology side very capital driven. If we're going to make a choice as to how we run a project, "Are we going to take it out of operational R and M or capital?"

Capital seems to, often times, be the first choice and leaving the R and M moneys for us for the more day to day, "Hey, this thing broke. A knob came off in my hand. Send it somewhere to fix it or internally decide to fix things." The hard thing is just making the decision, "Where do you put that money in?"

As we've gone through our facility, I mean, we went through a significant building addition in 2004 through 2006, that's when we initially added Axia into our facility. Well, we're eight years coming up on, now, almost nine years, with our Axia systems with our play to air systems, with all of the equipment that we installed at that point in time and technology has changed. I mean, for us the Axia core, for example, that we purchased eight years ago is now, If we were to buy the equivalent, it's probably overpowered and at the time under-featured for things that we actually want a system to be able to do for us. It doesn't have the redundancy that we want and at the same time it is probably a much higher level corporate core that we wouldn't spend the same money. We'd put the money in a different place or for different gear. Kind of a random example there.

Chris: Where do you look to, out to the horizon, to find out where you're going with technology and what is the appropriate technology to look at? For example, I'm imagining that a lot of this kind of starts with the germination of idea with your programming people or your production staff. I find you have to be, pretty much, right up with them in terms of, "Okay, if that's where we're going, here's where we need to look and here's what we need to look at." How do you find yourself keeping on top of technology in order to be able to satisfy those requirements?

Doug: Sure, so for us as I've watched this process, and I'm not going to say it's 100% of the time like this, but as I've watched this process, what I oftentimes see is, "mandate" isn't the right word, but a strong influence coming from the content folks in the organization. To want be able to either created something new or do something more efficiently. So those are often times the germinations of ideas. So it's not necessarily, "Hey, we want to do this show and we want it to have X, and put that together for us." But instead it's, "Hey, in our newsroom right now we're copying and pasting things now, out of our ENPS and putting them into emails to send them to the web producers, who are then producing that. But not getting updates as a story progresses though out the day, because people forget to email them. What kind of system can we put together that might help with that?" Or, "We are putting together content for a radio broadcast and it's getting dubbed on a CD and that CD's getting down the hallway to somebody who then rips that audio off of the CD into an MP3 to turn around and post it on the web. How could we make this more efficient so that we're not burning CDs, or people aren't having to walk things around?"

That, in combination with then the support and encouragement, and emphasis from higher management that we want to just make things more reliable, that we want higher percentages of uptimes, we want to take fewer outages. When you put those two things together it gives us and technology an opportunity to put together capital requests or project plans for things that really move us forward, in terms of the mission, or the goals that the programming department would like to see. In combination with making sure that we have the right mentality that the systems we put together are going to be highly available and highly redundant.

As far as figuring out what those technologies are, oftentimes the need drives the desire to go look for the technology. When we begin looking for the technology then it's your normal sources, whether it be looking at N.A.B. or going to N.A.B., looking at folks there and having conversations. Talking to other fellow public broadcasters, talking to National Public Radio, or PRI, talking to other folks in the industry. Having some connections in the industry certainly helps you hear about things. You know people who have moved from job to job and they end up with a vendor who has this product. That when you start chatting with your friend you find out that somebody has something that you could really use, that's going to solve a problem for you. So, it's a variety of ways to learn about the technology.

I will say as well, for Minnesota Public Radio, we tend to have strong partnerships with the technology vendors that we choose. Within those partnerships, we tend to pursue development for new things that those vendors may have thought about, but are not necessarily doing at the time.

A couple of examples of that, when we did our Axia project originally in 2004/2005 that was the big push for Element. We were one of the first stations to use Element regularly throughout our building. There's some development that was done for soft buttons, and scripting, and other things, a lot of automation functions.

The same thing right now, we're in the middle of a large Dalet project here at NPR. We've rolled out our newsroom onto Dalet digital media systems. We're in the process or rolling The Current and our classical music service. We have commissioned development from Dalet. In the interest of integration we asked them to develop against the IP audio driver so that we are now able to use the Axia IP audio driver on our broadcast servers and production workstations. We're working with them on a music master integration that will also for a fairly seamless ability to schedule and reconcile and keep the metadata sets between our Music Master scheduling software and Dalet play to air system, keeping information together.

Along those same lines we have an internal development team, so we're constantly looking for opportunities. For instance, we have a fairly homegrown digital media archive that's spans back 20 to 30 years. We're constantly digitizing older reels and things. We want to make that available internally for search for the newsroom or classical folks, or whomever. If Yoyo Ma came and visited 20 years ago and he's going to be coming back we need our classical folks to be able to search for that recording and pull it up.

For us, the integration point to Dalet and to the web, and ability for people to access that content both internally and when appropriate externally. It really, again, drives this idea as a content driven company.

Kirk: We've been talking about this operation at a rather high level, a lot of ideas here. I wonder, Doug, if we could get you to kind of tell us what's under the hood. I'm curious about what's hooked to what. You know, the leg bone's connected to the ankle bone. I want to hear about how you're getting things from where the content is made and then what's the process after that? Obviously it ends up at Stitcher, or TuneIn, or your website, so can you take us through that?

Doug: Absolutely. So, focusing a little more on the content production as opposed the specific radio broadcast side. Dalet digital media system is becoming, as I said, the project is in process right now. But is becoming our primary play to air, but also along those lines our primary media asset management system. So Dalet, for us, isn't just a hit the button and listen to the audio goal, but it's also an integrated audio editing tool in tandem with Pro Tools. So, depending on the project size, we'll use Pro Tools or Dalet's One Cut audio editor. But that also has become our primary newsroom and for classical our primary scripting tool as well.

So, we were previously using ENPS. We've moved away from ENPS into Dalet for scriptwriting. The ability to take these produced audio elements, drop them into scripts. Add in all the needed asset management fields, so that Dalet has the ability to customize asset manager forums. Take that, design the forms with the information that we need, depending on the user's particular needs. So, someone from The Current may see different fields that somebody in the newsroom does. Then use Dalet's tools for audio export or asset export, as far as the actual metadata component goes, into XML, then hitting a variety of systems that then have ability to parse that XML, grab that audio file, do conversions if necessary. Ideally automatically upload that to that to the web or to our website, or to the third-party folks.

So, one of our big pushes right now is to get away from the manual clicking steps. It used to be that, again, somebody copied and pasted things of ENPS and copied a file out of Dcart or NETIA, depending on which service was uploading content. Then manually copying something into to a web forum, uploading something, and hoping the two things looked well together when it got to the website. A lot of our push now is to try to automate that. To try to get this to be a one click step where and approval step says, "Yes, this is ready," and the systems take care of getting the content where it needs to be.

Our other push along with this, of course, is we do content, just trying to make the content sound as good as we can. Of course, the standards for web, because of encoding, because of bit loss as we go through that, sometimes slip a little bit. Well, for us, we've been in investing in Minnetonka's audio encoding and audio processing software. So as a part of this process we're able to export something for Dalet, hit Minnetonka audio, and do not just and encode to MP3, or MP2, or FLAC, or whichever. But also do loudness, do some EQ, depending on if that's going to improve what that's going to sound like, and allow us to really put our best foot forward, no matter the type of content we're putting on the web.

So, that's actually a part of this whole workflow process as well, is a series of watch folders or a series of drop points that are all triggered based on status changes. They're all coming out of the parent play-to-air system, which includes the metadata that's important for our users as well.

Kirk: Wow. My head's kind of spinning. Doug, this is so different than what me at my little radio stations, and what so many of my colleagues at other small groups of stations do and yet what you're talking about, is the future of where a lot of folks need to go, or some folks already are there. You guys certainly seem to be leading the way in terms of technology and then the workflow that uses that technology.

Doug: Absolutely. The challenge is, with any of these, I mean, it doesn't sound technology, but the workflow is always the thing pushes a lot of this innovation. It's getting away from having to have somebody sit in a room and double clicking a bunch of things in order to make something happen. The question always comes up, "Well, is there a way to automate this?" Our assumption is, even if we automate it, it needs to be a high level of reliability. We can't have automation breaking down, we can't have it failing. Because if there's a breaking story and the content doesn't make it up, or it makes it up well after the fact, at that point in time we've missed our audience. They've either gone somewhere else or if they haven't gone somewhere else, they're disappointed that they're not seeing the thing as it happened. We certainly hear about that from our listeners. Technology is important. But, again, it's listener driven and our listener needs are driven by the content that we produce.

Along those same lines, of course, I've mentioned Axia a number of times. Integration is one of our big things that we're really pushing. So, the ability to integrate Axia into Dalet. The ability to not just have Dalet be our dropping a bunch of files in places, but also being our primary playback tool. Integrating directly with Axia, of course, there's cost savings for us, not having to wire out GPIO nodes, not having to invest in sound cards and higher level PCs, when we can do things in the server room.

All those things, again, fall in to the same idea of being able to deliver a high quality product to our audiences and make sure it's highly reliable. But also make sure that we're not reinventing wheels or having to take extra steps to get that content to the end places.

The ability on the Axia side to be able to multicast is huge. The fact that we can have one room hitting multiple nodes across a variety of devices and different audio processing for each of those devices, with just a couple of clicks of a button, is also a huge thing for us.

We have, as I said, four primary services, but with our HD supplementals we have probably another seven on-air services. If I remember correctly. But we have, I think, about 16 web streams, and in the works, ideas for probably another ten. Many of those automated, so many of those coming out of Dalet, but always wanting to make sure we the appropriate audio processing and are making sure that it's not just slapping a Winamp playlist on the radio and saying it's okay. But instead making sure it's high quality and going to be there, and reliable when we need it.

Kirk: Wait a minute. Isn't that how you do it? You slap a Winamp playlist on the stream.

Doug: I would probably get a dirty look from my boss, but I will admit we probably have a couple of those playing somewhere, but there's also detailed plans as to how to replace that and make it better.

Kirk: Yeah, exactly.

Hey, you're watching This Week in Radio Tech Episode Number 206.

I'm Kirk Harnack along with Chris Tarr, who's in Mukwonago, Wisconsin. We're talking to Doug Rowe. Doug is the Media Productions Systems Manager at Minnesota Public Radio and their daughter company American Public Media. He does a lot of things with them including helping them set up streaming and studios in other locations. Minnesota Public Radio apparently isn't just limited to the confines of the great state of Minnesota, so we'll be talking about that in a few minutes.

I want to hear more specifics about, how do you work out what audio processing sounds right for a stream? Then I'm also interested in hearing about specifically, what kind of things go be a podcast that you can download versus the live streams and kind of how that workflow works.

But in the meantime, our show is brought to you in part by, my employer, the folks at Telos Alliance. I want to tell you something that's new at Telos Alliance, it's a newsletter, "Okay, newsletter, ho hum." But this is cool. I look forward to this thing showing up in my email box. It's called Direct Current.

There you go, Direct Current, and here's the one that just came out a few minutes ago. In fact, during the show here I just received this one. Here's an article, "What you need to know about digital audio delay." This is kind of cool, "The first audio delays were tube based, but they were garden hose, that kind of tube." Did you know that? "A speaker at one end and microphone at the other. Today's radio broadcasters can take advantage of technologies that are quite a bit more sophisticated for audio delays as well as for shrinking and stretching program material." There's more information on that article.

Go down a little bit more and, this kind of has been making news. That is that the Telos Alliance has doubled, in some cases, or increased the warranty for all products to five years. So it was Axia products that were five years, now the division that I run, Telos, all of our products are now guaranteed, warranted for five years. So our Omnia, 25-Seven, and Linear Acoustics. So a five year warranty, brand new. Something that we're going to really let people know about at the N.A.B. show, but we're letting you know about it right now. If you register your product that is retroactive to January 1st of this year, so if you've already purchased something this year, January 1st, you can go register your product online and the warranty is retroactively effective for five years.

A little farther down an interesting article about, "Found in the Attic: a Studer A710 cassette deck." A little fun that we have there.

And there's often a little blurb about what we did last week on This Week in Radio Tech in case you want to have a reminder to catch up that way.

"Direct Current" is the name of the newsletter. You can sign up for it by simply going to the website Telosalliance.com and if you scroll on down toward the bottom of the page you'll see a link for "subscribe to Direct Current". Way down near the bottom of the page. Just sign up for it and we'd be glad to count you as somebody who receives, and perhaps even reads, Direct Current. I think it comes out about every week or so. If it's too often for you, you can easily unsubscribe, we've got all that working the way you're supposed to. But I've found it really enjoyable. The guys who put that together just do a great job. Check it out, Direct Current from the Telos Alliance.

That's all we'll talk about is that newsletter. Boy, if we could just get everybody to subscribe to that. I think you're going to find it quite enjoyable and informative, too.

Hey, how else would you know about the five year warranty? Unless you read it there, or maybe you saw it on Facebook.

All right, we're talking to Doug Rowe on this Episode Number 206 of This Week in Radio Tech. I'm Kirk Harnack along with Chris Tarr.

Chris, you still with us?

Chris: I sure am.

Kirk: All right. Now, I don't know about you, but I'm dying to know what Doug and his colleagues what their criteria are for deciding on processing. Not necessarily the brand of processor, it would be great if they used ours from time to time-but how do you decide how you're going to do processing on a classical stream versus a talk, instructional programming, that kind of thing?

Doug, can you speak to that?

Doug: Sure. We have been investing. We have a group of Omnias and some Orban OPTIMODs as well. The lower bit rates, like the Omnia 1 and the Orban 6200s or 6500s, I can't remember which one, but the idea there is they're optimized for lower bit rate streams, both for streams and HD radio.

As far as the actual content processing, it's a combination of our program directors, one of two of our more golden ear hosts, and then we're lucky to have an entire ops team here at Minnesota Public Radio that does concert recordings. We record the Saint Paul Chamber Orchestra, Minnesota Orchestra, all kinds of groups. Getting those folks into a listening room, into a good listening environment and really, honestly, just tweaking things until it is what they think it should sound like. Obviously, opinions on that run the gambit and often time it is mostly just that, an opinion.

I'll say at one point in time I had dialed a couple of those in and I know that when one of our hosts who wants absolutely zero processing in his show and the other host who wants it to sound like a local R&B station. Well, they both were really angry at me, because they thought it sounded terrible and I thought I was pretty close. We actually haven't moved it too far since then.

As far as what that ideal sound is? For a classical station it's, I think, from an operations standpoint or somebody who has a good set of ears, it's what sounds the most natural. But taking into account the fact that many of our listeners are listening in a car or are listening in non-ideal environments, whether it's ear buds in a subway, or an FM radio, or in their car, or in a house full of kids with a mono speaker web radio, sitting up on their counter top, or something like that. We have to take all those things into account.

Along those same lines you also have to in your market take into account what your competition is doing. So for FM broadcast we did using a Golden Eagle, what is it? Aztec, Audemat Aztec Golden Eagle, we did audio recordings and analyzed that. Both as far as loudness went, and as far as where we saw folks modulating. Then had folks listen to that and made adjustments to our rock station to, not match, but to what we felt was competitive with the local stations.

Along those same lines for the news and information side of things. Of course, it starts in the studio. So we were careful about mic selection, throughout the entire air-chain what processing we need. Then it's based, again, on what our competitors are doing as well, in addition to just what our operational folks sound good. So just because someone chooses ultra-bass for their talk radio show, doesn't mean that we'll necessarily do that. But at the same time we want to be sure that folks in the car, or folks listening in a non-ideal environment don't hear us disappear when they tune to our station or dial in our web stream.

Chris: Do you have sort of guidelines for producers and content creators in terms of audio quality, recording set up, things like that, because obviously garbage in-mostly garbage out, so you really need a lot of quality control on the front end as this content goes in. Do you have guidelines set up for producers like that?

Doug: Yes. One thing that helps us here at NPR is the majority of our content is produced by staff in-house. So as opposed to having contractors or folks submitting thing to us, that does help give us a certain degree of control, that if it was purely freelancers sending in things. You may be run into that a little bit more where you don't have as much control over the actual recording process.

Of course our concert recording and things, are professionally done from our operations teams. So in that sense we know we'll be getting a high quality file.

Along those lines, though, still; the majority of our content is recorded out in the field on Lowrance recorders. Typically mono just to save disk space, although most everything is stereo, once it gets in house. We'll do L plus R as far as the production goes. But, they'll do it using a Lowrance recorder, usually WAV files 44.1 kilohertz, 16 bit recordings. They'll bring those in. If they need to submit something remotely, if it has to go over a 4G or 3G network and bandwidth is a concern, typically they'll use a standalone Dalet editor. Or an editor of their choice and compress those files usually as FLAC files. 6 to 1, I think, is usually the ratio they choose to use. They'll upload that via FTP or a Windows drop folder if they're able to get a solid VPN connection.

Once it gets back into our system, Dalet does have the ability to convert on the fly, so you can mix sample rates, you can mix bit depths, you can mix types even within one edit session. But just to try to keep things consistent we'll actually still re-up convert that to 24 bit, 48 kilohertz, which is our house standard, both into let [sounds like 42:43].

We chose that house standard because we knew it was integrating with Axia, 24/48 was going to give us the cleanest transition as Dalet played out onto the Axia network, without having to add in an extra layer of sample rate conversion.

Once in Dalet they'll edit at 24/48. They'll put together their package. When the package is complete, at that point in time it'll stay 24/48 all the way throughout play out until it gets to the actual audio encoder, if it's going to the web as a stream, or audio encoding, MP3 or whatever it might be for a podcast and that sort of thing.

Chris: [inaudible 43:28]

Doug: I'm sorry, go ahead.

Chris: Oh, I was going to ask, with your rock station, with The Current, I know you play a lot of local music and I'm sure some of the submissions there are a little questionable in terms of audio quality. We have the same issue here. Is there anything you can do to fix those up or do you have guidelines even for people who submit the music?

Doug: Typically, for us, we will actually push folks to submit a higher quality audio file, if at all possible. So, we're not playing too many MP3s or things that somebody just dropped off. Not to say that it never happens. But our folks who are actually on the audio engineering side they have good ears. They will typically give everything a listen before it goes to the radio. We rarely, if ever, have folks, and I'm not going to say we've played something off of somebody's iPod, but it is very much the exception, not the norm.

Same thing, we have really pushed not to do YouTube videos or grab something off the Internet and play it directly. Both, because it doesn't necessarily sound good, but also you're always running that risk that if something goes into buffering in the middle of the tune and now you've destroyed the listener experience, because you're not providing good content anymore. Just like a CD player dying or something, but much more likely to happen.

Chris: But I know for a lot of times now for music the big thing is SoundCloud, trying to rip something from SoundCloud or something like that. That's a challenge that I run into handling the audio at my station, is setting up those guidelines and trying to get them to understand that it's not just because I like to make rules. But, those sorts of things, while they sound good on their own in the computer speakers, once you put it through an audio chain that actually has some high end audio processing, boy, it's hard to fix that after a while. But I can see it as being one of the challenges of playing local music especially.

Doug: Certainly. No, and typically if we do have something that's a new release that' only available in MP3 or something, if they feel they just have to get it onto the air, "We have to play this today. We can't wait for anything." They will put that MP3 on the air. They'll do a couple of cuts from something, but once a higher resolution copy becomes available it's part of their workflow to go back and update that file in the system. To make sure that at least for the long term we're playing something that's better, I guess, in that respect.

The other half of that, of course, is that we're lucky enough to have, not that we always have this opportunity. But we have an in-house recording studio. A lot of our local content that we play isn't Joe Schmoe's band sending us an MP3, but it's using going to one of their concerts and actually recording something, or inviting them in for an in-studio and actually having complete artistic control.

In addition to our Axia system in our studio we've got a full, I think it's a 110 fader, Neve console, a very, very acoustically isolated room that's basically well set up for live opportunities. The great thing about that, of course, is it's not just a recording of somebody playing, but it's also an opportunity for hosts to be in the room, to do interviews, to have a conversation. Bring more than just a performance, but bring an actual experience with the actual artist.

So that's something that, as a large organization, we're lucky to have those opportunities.

Kirk: Okay, I'm really curious from this technical point of view, you care about your audio quality, you have a responsive audience that, obviously you're talking to the audience back and forth. Which streaming audio formats have you guys decided to standardize on at this point in technology. Is it any different for podcasts that people might download from your sites?

Doug: Sure. So, we're streaming MP3 and AAC. We have dropped Windows Media. We were doing Windows Media for a long time. That was actually one of our projects. Take a step back. Within StreamsGuys we've expanded our infrastructure and both on what we pay them for as far as load balancing and things, and gone through a couple of conversions. As we've done that we've had the opportunity to drop Windows Media, but to able to divert customers over to our MP3 or AAC streams.

As far as podcasts go, I believe we're also MP3 and AAC. I think if you go to our website, if you dig back into our archives, you can still find some Real audio, but very rarely anymore. We've been going through a process of trying to convert that to actually be a more modern codec, I guess, for lack of a better word.

Then along those same lines, I don't remember the bit rates right off the top of my head, I believe our MP3 streams are 64/128. AAC, I think, is 64. I don't remember if we have a 128 option for that or not, as well. So we were very deliberate about that though. We did a lot of listening testing, spent a lot of time with folks, with golden ears folks within our own teams, to make sure that the things we're putting up we're encoding at the levels that we wanted them to and sounded good.

Kirk: I'm sorry, is that the same for real time streaming versus podcasts?

Like if I go download some audio from one of your websites?

Doug: That is correct. The MP3 for sure, I'm about 98% sure that our podcasts also have AAC options most of the time, and for sure the web streams are MP3 and AAC.

Kirk: Okay. Earlier you had mentioned FLAC, do you have use for FLAC in your facility?

Doug: Primarily we're not broadcasting or making directly public any FLAC content at this point. But we use it primarily for some of our remote recorders. We have bureaus throughout the state where we have reporters, I guess, "embedded", for lack of a better word. We have a state ranging MPLS network that we get through CenturyLink. That MPLS network becomes our backbone then for a lot of our content pushing, but often times we're bandwidth limited to maybe only a couple of megs worth of bandwidth for pushing files. If they're in a hurry a WAV file will crush that network connection pretty quickly for a little while. A FLAC file is often times the most efficient way to get something across the network to us here.

Go ahead?

Kirk: What I was going to ask next is... Okay, so inside your facility you've got an audio over IP network at high speed, you're using Axia and LiveWire, so this is all linear uncompressed, low latency. That's all the stuff in your two dozen or more studios plus all the work stations that you've got. Then your public facing streams tend to be these lower bit rate, of course, as you mentioned. 64 kilobit and maybe 128 kilobit streams that we'd all be fairly accustomed to. So then you have great audio for the contribution audio and you knock it down to a low bit rate for distribution. So, I'm curious about your other infrastructure. What about going from your Saint Paul offices to the various transmitter sites, is any of that audio over IP? If it is, is any of it linear or is any of it compressed?

Doug: Sure. Well, we have a combination of things of course. Again, looking at redundancy we've implemented things in multiple iterations or in a variety of manners just to make sure that, if a system experiences a failure that we can continue to provide content to our sites, whichever. So, for our primary, we have 37 radio stations around the state as well as a couple in a couple of other states.

I think we have one in Idaho, a couple in Iowa, and South Dakota, and Michigan if I remember right, off the top of my head. We using actually Pico Digital's XDS system, satellite delivery system, DVB, in order to actually distribute our audio to those sites, so we have satellite downlink dishes. I believe we're encoding at, I think it's, 256 MPEG2 if I remember correctly for that system, as a primary delivery.

In addition to that, Pico Digital's XDS system has the ability to stream, so many of our sites or on MPLS, the other sites that aren't, have standalone Internet connections with CISCO VPN back to our primary facility here. Pico Digital can send a stream over those connections as a backup audio option with a switcher at the site to choose.

In addition to that some of our sites, and as we're doing for more and more of remotes, we're using a lot of Comrex gear. So, we're using Comrex Axis' and or Comrex BRICs depending on the appropriate application. In order to provide an actual link between different sites, whether it's a temporary thing for a news conference. Or, as we're doing for instance, for our Sun Valley site, a pair of Comrex BRIC units that are the full time STL to that site, as we weren't able to put in a satellite dish. So, it's a variety of things for that as well. For most of those devices the tradeoff is, of course, audio quality versus bandwidth and then latency is a factor in there.

Typically we're using AAC codec for that and usually, I think, it's between 128 and 256. If it can be bumped higher, we go higher. But it's very dependent on usually the unique situation for each individual site.

Kirk: Wow.

Chris, we're going to be out of time here pretty soon. You got any follow up questions here at the end?

Chris: No, I was just intrigued by listening about the content distribution, and using the satellite versus the Internet, and those sorts of things. But, what I'm taking away here is that it's a very non-traditional sort of setup in terms of how radio stations go with technology.

How has your experience been integrating all of this equipment from different vendors? For example, Axia, Comrex, Dalet, ENCO, those sorts of things? I know that fortunately they sort of talk the same long language, but has that been a challenge trying to get all those things to work together?

Doug: Sure. Everything is an opportunity and everything has the opportunity for a challenge. Typically what we've found if we can work with the vendor and if the vendor sees that it is an opportunity for them, for other potential customers. They are interested in partnering on this technology. Usually there are some growing pains. I'll say that our Axia-Dalet integration, the integration between the two pieces of software actually was very solid from the beginning. But we went through a couple of iterations of hardware, as an example, in order to make sure we were getting-, it wasn't so much the integration that was a point. But it was having a high CPU demand piece of software that's taking a lot of system resources on the Dalet side, in combination with an IP audio that is trying to do low latency, full resolution, 24 bit, 48 kilohertz audio streaming, for lack of a better word, all off one box.

So our challenges have oftentimes not been getting companies to do the integration, but it's figuring out exactly how to host that on the hardware properly. Then, again, coming back to the redundancy aspect of things, making sure that when we put any of those systems together, that it's not in a way in which a failure on a piece of hardware or a failure on a network segment can cause an outage. For this it's a constant growing process.

An example for that, right now we're in the middle of a large Axia core replacement project where we're moving away from our CISCO 6500 that was installed six years ago, or seven or eight years ago now. To a new Nexus 7 dual core platform that allows us to dual end everything. It gives us complete redundancy. We can lose half of our core and keep, theoretically, every switch in every part of the building still running. But, the challenge there is getting that to function as far dual-homing on PCs, dual-homing on servers, and making sure that even connecting that Axia network to our business network. Which is actually sometimes frowned upon, but is actually a primary core part of our infrastructure. Making sure all of that can always stay up and running and no one segment doesn't take down the rest of the system and cause a problem.

Kirk: Wow, you guys have got a great plant there. I look forward to coming back and seeing this after you've done this core switch replacement, because you guys put a good amount of effort into getting your Axia system, which is normally on its own physical network. To also be present for all of your business needs, including all those workstations that were out there. So, I'm eager to see how that's going to go, replacing those core switches with something new.

Doug: April 29th is our cut over date. We've promised our bosses that we won't take any audio outage, and my fingers are crossed. But, it's a fun project. Actually I'm excited, I know you're wrapping it up here in a second, but I'm excited that we're getting to go down to CISCO labs down in Omaha. We're actually going to completely mock up everything with a copy of our existing core, the new core, and the other gear. Put everything together as far as the code goes, step by step, and figure out exactly how to make this cut over happen without taking any outages. So, definitely another way to honor our commitment to our listeners and try to make sure that the systems we put together to deliver the content is close to 100% of the time as it can.

Kirk: Doug, I have enjoyed this immensely even though probably half the time you're talking over my head. Your environment there, with doing so much content, is just amazing. It's like having just dozens and dozens of radio stations and lots of professional talent. It's hard to imagine for a guy who for years was a work-a-day engineer at a small group, so I'm glad that you have broadened my horizons a bit on this. I can start to see this really big picture.

Chris, any last comments, parting?

Chris: No, like I said, I'm glad we were able to get you on as a guest. A lot of what you're doing is, I always kind of joke about that. In fact, one of our guys used to work at The Current and I always kind of joked that I'm about two or three years behind you guys. So, I watch real intently with what you guys are doing as kind of a template, because I'm sure that's the direction that we're going to be going. So it was real interesting to hear what you're doing.

I just have to say, from my point of view, that's one of the great things about working in public media, especially. Is that we're given a lot of opportunity to work with things like this, because it's become a whole lot more than just about being on the radio. With these alternate streams of content, and all these different types of ways to create new content. It's an interesting opportunity. It's something that I think is really neat. I think as an engineer or technology guy keeps things interesting because you're always forced outside of your comfort zone and pushed to go up to that next level. I think that's always a lot of fun. So, it was great to hear this past hour some of the things that you're doing.

Doug: Oh, absolutely. I've always thought the fun part of this job, and especially working in public radio, for me, is it's not necessarily putting out fires. That's not to say it's never stressful, but it's often times the challenges come in trying to figure out new things. I think our funding model helps a lot with that as well as just a good attitude as far as our management. There's opportunities to make things better, it's not just break fix. So, it definitely keeps the job interesting and it was fun to have the opportunity to come on your show and chat about this for a little while.

Kirk: We're going to have one of Doug's other colleagues on a future show coming up. That's Tom Nelson. Tom does a ton a engineering there at Minnesota Public Radio, so we're excited to have him on, probably some time after N.A.B.

We've got to go. Doug Rowe at Minnesota Public Radio the Media Productions Systems Manager has been our guest.

Chris Tarr was able to join us today, another co-host.

Chris is with...Tell us about your website, Chris, where you have some forums? What's that?

Chris: Broadcastengineering.info is the name of the site and that's a great place to exchange ideas, questions, thoughts, whatever you want about the discipline of broadcast engineering. We have a bunch of great guys on there. As a matter of fact, there's a story there. The site's been around for, I think, seven years now, there is a radio station, I won't go into a whole lot of detail, but was conceived, planned, built, and put on the air entirely using help from that forum. So, a guy came on several years ago and said, "Hey, I'm looking to get a radio station." We had guys help him find the allocation, help him walk through the process of getting it, helped him pick out gear, answer questions about building it. So from start to finish, was able to put a radio station on the air. So some neat stuff over there.

Doug: How cool.

Kirk: That's great. That's awesome.

All right, thanks for joining us. Our show's been brought to you by the Telos Alliance and the Telos Direct Current newsletter. I'm going to encourage you to go Telosalliance.com, scroll all the way to the bottom, and over on the left-hand side there's a little button "Subscribe to Direct Current." It's funny, pithy, informative, and you'll find out some cool things like you'll find out about the new five year warranty on every Telos Alliance product. So check it out if you would.

We're going to be back next week with another show, just before N.A.B. We'll also do a show live from the N.A.B. floor in two weeks from today, so I hope you'll join us for that. We're going to have a good time there and we'll see who we can grab from the floor and talk with.

I appreciate Andrew Zarian being our producer in New York City and we're going to see you next week on This Week in Radio Tech.

Bye, bye, everybody.

Topics: Radio Engineering