I was stressed leading up to the election. Stressed, not as most Americans were about the outcome, but stressed about the production of what would be a first for CitrusTV: a three-and-a-half hour live election special. This would not be the first election special the station had done, but it would be the longest and the most involved. From my perspective there were two distinct moving parts.
As we look to stay competitive with other college television stations across the country, we see who's winning awards and what they're doing. We had a one-hour special in 2014 for the mid-term elections, but that broadcast didn't receive a single nomination for any of the awards we submitted it to. It wasn't a bad broadcast, but there were other election specials from other stations that did win awards. A common theme we noticed was live hits. Pulling the coverage out of the studio into the field makes it more immediate and more representative of what's actually happening. With such an historic election, the pulse of the nation on election night was definitely an important concern. Our live infrastructure has long been a stumbling block for us. We've used Streambox for some time (and still continue to do so), but I've always been disappointed in its reliability, stability, and complicated UI. After spending many hours tweaking streaming settings, I finally got what I thought to be a usable image out of the unit. When we did a dry run the Sunday before election night, we experienced almost a 10-second delay bringing me back to the drawing board for my encoding settings. After tweaking more, we got the delay down enough to be usable. The issue then became talent cueing. Streambox does not have a built-in IFB solution unless you use their IFB server which we don't own. So, we had the talent plug a standard IFB earpiece into their cell phone which we made a phone call to and sent a mix-minus feed down the line so they could hear program. The drawback becomes no ability for the producer to cue talent, but as long as they could hear the toss and any questions from the in-studio talent, it worked just fine.
It would be all well and good if we had more than one Streambox encoder license, but we had only one and the producers wanted four live hits in the show. One each for the Democratic and Republican HQ's in Syracuse, one at a polling place on campus, and then the reporter at the polling place would go to a campus (non-partisan) watch party after the polling place closed. Fortunately, we'd made some investments in live remote infrastructure earlier in the year by setting up a dedicated computer to receive video Skype calls. I, like probably everyone, have a love/hate relationship with Skype but it worked alright for this purpose. That Skype machine has no video return, but does receive a true IFB feed (in that it is interruptible by the producer) with mix-minus program audio. Again, we ran into trouble with Skype because we only had one Blackmagic UltraStudio MiniRecorder which we used to get an HD-SDI feed from our cameras into a computer via Thunderbolt. Thus, the last remote hit would have to be done via an iPad on Skype. The issue with an iPad becomes audio (both audio from a "real" handheld microphone and IFB audio through an earpiece for the talent). IK Multimedia makes as device called the iRig which lets you plug an XLR mic and standard headphones or IFB into an iPad. It costs about $150 and because I'm cheap I figured I could build my own iRig. While it wouldn't have a pre-amp, I figured I could get enough gain out of the iPad's pre-amp itself to at least be usable. After some Googling, I figured out those old RCA camcorder cables that you could use to plug a camera into your TV actually used a TRRS connector (the same as the iPad) and that the yellow and white RCA's were audio right and left out of the iPad and the red RCA was microphone in. I just so happened to have one laying around and after trying it out with some adapters and having success, I cut the RCA's off of the cord I had and soldered on a female XLR for the mic and female unbalanced 1/4" for the audio out. It would have worked great if not for one minor mistake I didn't know about until after the special was over. When the remote reporter took the iPad out, they neglected to take the case off of the iPad. Because of the case design we have on our iPads, it doesn't let the connector seat fully meaning we got the internal microphone back on the iPad. It wasn't the end of the world (I always say, I'd rather have something that's ok than nothing at all) and we still got through it. The whole Skype thing was something of a pain because I had to keep connecting and disconnecting the two people who were on Skype -- we only have one computer set up to be a Skype endpoint, but at least we still had both Skype reporters.
Photos: Stills of the live hits from the night. Rob is on Streambox, Jamie and Nicole are on Skype calls.
Overall, the live hits performed flawlessly. No awkwardness, we could hear everyone, they could hear us or at least get a cue and one of our live reporters even got a live one-on-one with the Democratic candidate for congress from Syracuse's district. Not something that happens every day on student television.
I knew that we needed to develop a distinct graphics package for the special. Our normal graphics are newscast-specific so I elected to design a new package. We took inspiration from NBC News' election package and used our election theme from 615 Music (part of their Right Here, Right Now package). Originally I wanted to have a ticker at the bottom of the screen that would display races as they were called. My idea was to have the data for the ticker pulled from a DBLinked Excel file and have that message be persistent in the Chyron. The issue was I couldn't figure out a good macro that would keep running and keep updating the ticker while other things were going on, so I scrapped it. It was a good thing, too. Our Chyron began having issues the week before the special and we ended up running with only one channel of output for our graphics which would have made a persistent ticker rather difficult.
The most involved thing I did do was create fullscreen graphics for each state that could be used to call a race for either Trump, Clinton, or be "too close to call" if we just wanted to look at a state without calling it. The fullscreens gave a nod to our News Angles package from our News Live at 6:00 shows which I also designed and the election fullscreens came out looking pretty nice. I obviously didn't want to build 150 separate graphics for each eventuality and each state, so I got creative with macros and DBLinking. I built an Excel file that had the right data for the fullscreen in it (percents, votes, etc.) and a special field called "call." In this field, the newsroom would enter either a 0, 1, 2, or 3. They would enter 0 if the race had not been called, 1 if for Trump, 2 if for Clinton, and 3 if the race was too close to call. Whenever the message was recalled in the Chyron, an input box was used to allow the operator to select the right row from the Excel file (1 through 51 for each state and D.C.). That updated all the DBLinked fields with the right information including a "blind" field which contained the "call" field. Then, the macro looked at the blind field (so called, because it doesn't display onscreen) and had conditional statements to show or hide the three groups of the fullscreens which contained the fields and images for each eventuality.
In my testing the macro worked great and updated correctly every time. The issue was that I didn't have the fullscreen done by our tech run-through. That meant I didn't get to test in the exact way it'd be implemented on election night and a couple of things failed. First, I didn't carefully test all the fullscreens, so I had one field mapped incorrectly for one of the eventualities. That meant the wrong percentage was going to the wrong place and the wrong thing was showing up in the data. This was quickly fixed once I figured out what was going wrong, but it was still a simple error that could have been prevented. Second, the computer in the newsroom the staff was trying to open the Excel file on had an odd issue with locking/permissions. Every time they opened the data file, it thought the same user was already editing the file and tried to open as a read-only file. It also somehow corrupted the file when it opened the first time. I had to re-create it and use a different machine. The different machine was a Mac which seemed to have some issues with locking in the Finder. It didn't always let the Excel user save the file immediately (they sometimes had to wait 15-30 seconds) so this slowed production. Third, we never explicitly established the workflow for calling states. This sounds so simple, but we had called races coming into the control room where the data hadn't been entered or saved in the newsroom. Thus the control room would call up a fullscreen for a called race, and it'd be blank. This was a personnel error and could have been corrected with more forethought and planning.
Overall, things went extremely well. I watched other student-run broadcasts that were riddled with errors, technical glitches, and more. Above all, talent and content was the biggest issue. There were others that I thought had better graphics or the like, but we had so much better talent. Our moderators provided insight instead of just facts, we explored issues around the world, we explained common questions voters might have, and the broadcast was extremely well-produced as a whole. I couldn't be prouder to lead this station during this once in a lifetime experience and I'm proud of what our team was able to accomplish.