TUBA CITY

In the beginning, there was the Sun. SunSITE, that is. In early 1992, Sun Microsystems put out a formal request for grant proposals from institutions interested in creating and maintaining an anonymous FTP archive on the Internet. UNC's Office of Information Technology (aka OIT, later known as ATN and now melting into ITS) submitted a proposal to not only establish an FTP archive but to enhance the archive/collection with early search technologies like WAIS and Gopher. UNC's proposal was accepted, and SunSITE was born later in 1992. Sun Microsystems provided computer hardware and funding, and UNC ran SunSITE out of offices in the basement of Phillips Hall. Later known as MetaLab and now known as ibiblio, the digital archive is one of the Internet's oldest and largest collections of free information and software.

SunSITE director Paul Jones started with a small staff consisting mainly of part-time students who were passionate about the promises and possibilities of technology and the Internet. One of SunSITE's early hires was Doug Matthews, an avid local music fan who would later start the alt.music.chapel-hill newsgroup and help create some of the earliest local-music-related websites. Several months later in the summer of 1994, SunSITE hired David McConville, a new graduate student in the School of Journalism and Mass Communication. McConville would soon become a key player on a lot of SunSITE projects, including the WXYC webcast.

David McConville: I met Paul Jones and we immediately hit it off talking about the Clipper Chip proposals. I'd finished getting a BS in Audio Engineering the year before, and I had used the Internet enough to think I knew something (I'd actually had an account on LaUNChpad - the pre-SunSITE bulletin board system). I had researched a lot of the 3D audio processing techniques as an undergrad in classes with Bob Moog and I'd done some video editing, which made me familiar enough with digital media tools that Paul made me the "multimedia researcher" or somesuch. I was really interested in online audio/video/3D distribution, and he let me do basically whatever I wanted to as long as I kept up some of the funded tasks (running an online educational archive paid for by Sun and Cisco).

McConville and Matthews quickly became co-conspirators on a wide variety of projects dealing with music and the Internet.

David McConville: I worked with Doug Matthews quite a bit early on, scheming up all kinds of experiments. We contacted Negativland and made a site for them, putting the “U2” single online in MP2 compressed audio format as an experiment to see what kind of response Island Records might have. I called up the Church of the Subgenius and worked with Ivan Stang to create the almost indefensible "SubSITE". I created an audio archive of some NC sermons from tapes and transcripts at UNC's North Carolina collection.

Simply putting digital audio files online was pretty cutting edge for the early 1990s, but Internet technology was beginning to change very quickly. Prior to 1992, there was no acceptable way to "broadcast" sound on the Internet so that it reached multiple recipients at once. Sending a digital sound file to ten different computers required ten distinct unicast transmissions of the file, and such an inefficient use of bandwidth was completely impractical for sending out real-time audio. All of this began to change with the early 1992 advent of the MBONE, a virtual overlay network built on top of the existing Internet. MBONE stood for "multicast backbone", and the multicasting protocol used on the MBONE allowed high-powered Unix workstations to send information to multiple destinations in the most efficient possible way.

The MBONE was used to audiocast a 1992 meeting of the Internet Engineering Task Force, and much of the audio content being transmitted over the network tended to be technical jargon and/or meta-discussions about the MBONE itself. One of the more notable uses of the MBONE was Carl Malamud's "Internet Talk Radio", an internet-only "radio station" that multicasted the "Geek of the Week" show at a specific time every week. Clearly this new technology held great promise for the future, but in 1993, things like Internet Talk Radio were only available to a select virtual few.

In early 1994, researchers at Cornell University released version 0.7 of CU-SeeMe, a piece of video-conferencing software that would drastically expand both the reach and the potential use of real-time multimedia on the Internet. The free downloadable software was named CU-SeeMe as a reference to both Cornell's initials and its video-conferencing capabilities. CU-SeeMe allowed users to both send and receive small grayscale video feeds (including optional audio) to any other computer that was running the software. What made CU-SeeMe such a milestone was that it could run on most any System 7 Macintosh computer that had a full (non-dialup) internet connection. All of a sudden, functionality that was once limited to those with extremely expensive hardware was now available to a lot more people.

Doug Matthews recalled his initial excitement upon the release of CU-SeeMe version 0.7.

Doug Matthews: I remember getting the email when CU-SeeMe was released and thinking, "Alright, this is gonna be really cool". We installed it right away. CU-SeeMe was basically a multimedia chat tool. It was essentially a half-duplex walkie-talkie model. You would say something, and then when you stopped talking, you could start hearing again. The original CU-SeeMe chats I remember having were almost turn-based in the same way that chat rooms are turn-based, where you can't type over someone else's chat. It was basically a "push-to-talk" interface. You pushed to talk, and no one else could talk while you were pushed down.

The entire SunSITE crew took part in CU-SeeMe chats with people from around the world.

David McConville: All of us were were experimenting quite a bit with the MBONE and audio/video conferencing using CU-SeeMe. We ran a CU-SeeMe "reflector" at SunSITE thanks to Jon Magid's technical wizardry, meaning that people would log in from all over the world to videochat. I even got to see an IRA ceasefire that was blacked out by the British media through a web cam sticking out of a guy's window in Northern Ireland. It was all so new and strange and seemed to hold so much promise. Paul just let the entire crew experiment, which of course forced me to learn much more than I was learning in classes. Paul's amazing like that.

Suddenly able to exchange audio and video feeds with people from around the world, McConville and Matthews started thinking about other ways in which the CU-SeeMe software could be used. Matthews couldn't recall exactly who came up with the idea to use CU-SeeMe to rebroadcast a radio signal, but he credits McConville for being the one to develop the idea and run with it.

Doug Matthews: A lot of stuff that we did came up in "Wouldn't it be cool?" conversations that we had. Like putting the Negativland thing online. We're sitting there wondering, "if it's illegal to distribute, is it illegal to download?" I'm pretty sure that the idea for simulcasting the radio station was similar to that. I remember one of the two of us casually saying, "I wonder if we could drag a speaker up to this microphone?" But in terms of actually turning the idea into a production affair, it was 100% David. I was just involved in the brainstorming phase.

McConville was extremely interested in media and intellectual property, and he was excited by the opportunity to meld theory with practice.

David McConville: I was thinking of writing a paper about the legalities of re-transmitting radio online. I had to take media law classes in JOMC, which was not exactly my thing, so I figured I'd focus on all of the online innovation going on at the time. Folks had been experimenting with broadcasts on the MBONE, and we'd been working with CU-SeeMe to better understand the social and copyright implications of online audio/video distribution. There was a lack of understanding about the difference between streaming and downloading at the time, and I wanted to explore the distinction from an analysis of legal history. I was interested in media history, and I wondered if early radio history might repeat itself (i.e. the takeover of the ether by commercial interests via very anti-free-market regulations created by cronies of the corporate interests). I knew that we were in a position to do more than theorize, and that Paul and the SunSITE crew would be into the challenge. Since I was as much hands-on as I was theoretical, I figured we might as well rebroadcast a station 24/7 to find out for ourselves!

McConville started investigating what type of audio frequency range could be transmitted using CU-SeeMe. The software wasn't designed to stream anything, much less audio that was richer and fuller than a simple human voice. But that didn't stop the SunSITE crew from making plans to attempt to use CU-SeeMe to continuously rebroadcast a radio station.

Doug Matthews: The way we used CU-SeeMe was a total hack, it was not what that software was supposed to be used for. Basically what we did was we said, "Well, we've got this thing that can make sound go from Point A to Point B and because it's a chat-room-type environment, lots of people can listen to it and all we have to do is be the only one talking all the time." We totally bastardized the software, because there was no streaming software at that point. None of the stuff that actually did live streams came out until probably 1995.

As the nearby college radio station, WXYC was the obvious choice for the webcasting project. Matthews had started working there as a DJ in early 1994, and McConville himself had become an avid WXYC listener.

David McConville: WXYC was by far the best radio station I'd ever heard at the time. Aphex Twin, Faust, AND Yma Sumac - it made my time in Chapel Hill most bearable.

McConville chose WXYC to be the subject of his Internet rebroadcasting project, but it was the WXYC disc jockey who worked in the office next door who did the most to coordinate efforts between SunSITE and the radio station. Hired at WXYC in 1993, Mike Shoffner worked for OIT as a part-time systems administrator.

David McConville: I knew Shoffner through Doug, and we started throwing the idea around, working with Jon (Magid) and Paul (Jones) to figure out the logistics.

While McConville and the SunSITE crew started working out some of the technical issues, Shoffner began selling WXYC on the project. WXYC managers were intrigued by the idea of the webcast, but on the whole they expressed a much more cautious and hesitant attitude than that which was exhibited by the gung-ho experimenters at SunSITE. Several WXYC higher-ups expressed valid concerns about whether the simulcast would be violating any laws or FCC regulations. Since WXYC was in the middle of plans to petition the FCC for permission to move its transmission tower, there were worries that even the slightest violation might lead the FCC to either fine WXYC or consider the station's tower-modification application in a negative light.

Shoffner refused to let WXYC's caution stop the project, however. He realized that a remarkable opportunity was knocking on WXYC's door, and in a series of September 1994 emails to the WXYC powers-that-be, he passionately pleaded with his fellow DJs to do whatever they could to answer the knock. This excerpt from a September 15th message is indicative of the colorful nature of Shoffner's fervor-filled emails:

What is being jeopardized is our chance to DO THIS FIRST, IF ONLY FOR A SHORT TIME. FIRST. FIRST. US! Not NPR (affiliates), with all their goddam money/pretensiousness. Not MTV, with all of their corporate "alternative". UUUUUUUUUUUUSSSSSSS! This is the INTERNET, folks, and no matter how many times Floyd says "the internet sucks" on the air, IT IS THE BIG SHIT. And we are ONE SOUNDCARD ($100 - i'll buy it myself if needed) away from eternal salvation and the bountiful reward of our righteousness and long suffering.

In hopes of addressing and appeasing some of the concerns, Shoffner urged the station to find out the exact whats and whys surrounding possible legal issues with the webcast. Luckily, the WXYC staff at that time contained several current and former UNC Law students, two of whom - Jeff Robins and Jay Huber - started immediately applying their legal knowledge to the situation. In a September 16th, 1994 email to various people privy to the simulcast idea, Robins attempted to clarify the general framework of the legal questions that would need to be addressed in more detail:

As far as some of the legal issues that are concerning some of us, there seem to be two types:

1. Communications Law issues (i.e. those issues related to our FCC license and anything else related to the FCC Regulations)

2. Copyright Law issues (i.e. the ASCAP/BMI stuff, which deals with the dissemination of the 'intellectual property' of others, particularly musical artists, composers, and publishers)

For the most part, these are two separate areas. As we continue this dialogue and as some of us continue to do further thinking and/or research, it might be helpful not to confuse the two types.

Robins and the other XYCers all agreed that as far as Communications Law issues were concerned, WXYC was legally in the clear. The FCC had no legal jurisidction over the Internet, and the only worry was that the FCC might potentially look at the webcast as "behavior unbecoming a licensed radio station" when considering WXYC's eventual request for permission to move its transmission tower. WXYC did not have a history of run-ins with the FCC, however, and Robins pointed out that FCC’s approval of tower modifications was largely based on engineering specifications (tower location, wattage, signal direction, other radio stations, etc.) and not on a "radio station's behavior as citizen-of-the-airwaves". Only if another radio station decided to fight WXYC's application-for-modification would there be any real trouble during the FCC approval process.

The more difficult legal issue surrounding the internet simulcast had to do with copyright and intellectual property. Like all radio stations that play music, WXYC has licensing agreements with music publishing giants ASCAP and BMI. These agreements allow the station to broadcast copyright-protected music over the radio airwaves, but did WXYC have the right to transmit that same music out into cyberspace? In order to try and answer this key question, Jay Huber sought advice from Susan Olive, a noted Triangle attorney who specialized in copyright law. After consulting with Olive, Huber and Robins began searching the Federal Copyright Act for information that might apply to the unprecedented webcast scenario. Huber soon struck gold with Section 111(a)(5), a statute about secondary transmissions that indicated that the proposed Internet simulcast would be exempt from copyright law as long as it was not done for commercial advantage. Huber summed up the statute in this email from 9/21/1994:

1) Copyright: we are (jeff and i) quite confident that sec. 111(a)(5) of the copyright code allows a nonprofit entity (SUNSITE) who wants to retransmit the broadcast of a primary transmitter (WXYC) for a noncommercial purpose to be EXEMPT from the copyright laws. no infringement will follow from such a transmission. even if some infringement could be argued, that infringement would fall on SUNSITE, not us. however, SUNSITE is still on pretty solid ground. and, as has been previously mentioned, if ASCAP/BMI got pissed, they'd send at least three letters before doing anything serious.

again though, the copyright code seems to allow us to do this with no risk of infringing on ANYTHING.

further research this evening, relatively comprehensive research, found NOTHING to contradict the above statement.
proceed with all haste.

With all of the station's legal concerns adequately addressed, WXYC station manager Jame Lathren gave his official approval for the webcasting project. And Shoffner resumed his quest for the holy soundcard, which was needed for the Macintosh computer that would convert an analog radio signal into a digital bitstream capable of being transmitted via CU-SeeMe. In addition, the simulcast crew needed to figure out a way to actually get that analog radio signal from WXYC. For the current WXYC webcast, this analog/digital conversion takes place in the radio station itself – an analog signal comes straight out of the control room board and gets encoded into bytes by a nearby Linux machine. But in 1994, the radio station didn’t have computers capable of performing such operations; the analog/digital conversion would have to take place at SunSITE. So once again, the crew got creative.

Michael Shoffner: We had no AD converter and we had no tuner to pick up the signal to feed into the converter. So David specified and I bought an external card for the Mac to do the conversion. Then I begged a "yard sale" radio off of my kid sister and we put in right beside David's workspace, tuned in the station, and plugged it in.

This cheap radio tuned into 89.3 FM would soon become the input source for a webcast that could be heard around the world. But first, the SunSITE staff had to overcome several technical hurdles. Paul Jones wanted to make sure that the bandwidth used by the experiment would not impact the UNC network in a negative way. So a trial run was scheduled for late September 1994. Network load was monitored over a 24-hour period and no major issues were discovered. Plus, since SunSITE had its own T1 connection to the rest of the Internet, connections between SunSITE and off-campus hosts throughout the rest of the world would not need to be routed through the same pipes as other UNC traffic.

A separate technical issue arose because Jones thought that the rebroadcasting should be done from an underutilized machine that would be more capable of handling multiple connections at once. The CU-SeeMe software actually allowed clients to directly connect to each other (a model now referred to as “peer to peer”), and some of the early tests of audio quality and bandwidth usage had been performed with clients connecting directly to McConville’s Macintosh. But SunSITE had a Solaris server named president that had been used to put presidential library material online. And if a new CU-SeeMe reflector could be installed on this server, the WXYC signal could be multicast from president in a more robust and scalable fashion. Unfortunately, this caused a significant delay of the project.

Michael Shoffner: The major problem was that the "reflector" wouldn't compile under Solaris, which is where it needed to go in order to scale to the public. The "reflector" was the thing that broadcast one input signal out into multiple output signals. David had it running on his Mac. The fact that we had to get it running under Solaris just about killed the schedule.

SunSITE employees Chris Colomb and Jon Magid helped McConville out with the Solaris/reflector issues, and by early November, the problems were solved. The webcast was finally ready to go. After some additional testing on November 6th, the WXYC simulcast went live on November 7th, 1994.

Traditional media did not pick up on the simulcast story immediately, but news spread pretty quickly via the Internet. Over the next several weeks, responses came in from other parts of America (Nebraska, Arizona, Idaho…) and from countries all around the world (Norway, Poland, Mexico, Northern Ireland…). The sound quality was not at all ideal (some listeners reported that the already lo-fi signal was cutting out every so often), but the mere idea of making a radio station’s signal available to the rest of the world was nothing short of revolutionary. And both McConville and SunSITE director Paul Jones were well aware that the concept itself was way ahead of the technology.

Paul Jones: We knew what we had sucked but we were willing to try anyway…we’re talking low low bandwidth in those days.

David McConville: We were actually using a little $10 radio plugged into the back of a Mac for the off-air signal. It was pretty pathetic, but it worked. It was reassuring to think that we were somehow carrying on the gonzo traditions of Tesla and Marconi with a station as incredible as XYC.

Shoffner, who described his role as “cheerleader and catalyst”, was just relieved that SunSITE and WXYC had pulled off the feat before anyone else.

Michael Shoffner: The biggest fear I had the whole time was that we would get scooped. You had to assume that somebody else had thought of it because it's almost never do you think of something and execute before anybody else in the entire *world* does. So David and I were pretty much sweating the whole time - trying to get around the problems, light a fire under everybody else, etc.

As it turned out, other stations had thought of it. Bellingham, Washington’s KUGS (another 89.3 FM!) started webcasting their signal via CU-SeeMe a couple of weeks after WXYC’s official announce date. And Kansas University’s KJHK followed in early December, also using CU-SeeMe. And Georgia Tech’s college radio station WREK deserves extra credit for actually webcasting via their own self-written audio-encoding and webcasting software. WREK successfully beta-tested their internet simulcast in early November 1994, but they did not go into official release until early 1995.

Doug Matthews: This was definitely one of those independent discovery things where a lot of people all at once said, "Hey, wait a minute - I can make noise all the time!"

It’s true. In cyberspace, everyone can hear you stream.