Serial’s wild success was the result of two basic things. First, the podcast told an amazing story that got more interesting, mysterious and addictive the more you listened to it. Second, it had Sarah Koenig’s amazing writing leading us through that story.
As the first season began, Koenig and producers Julie Synder and Dana Chivvis were still extremely busy uncovering the details of the story and finishing the writing process. So they reached out to me for help with the scoring and the mixing. I had the pleasure of scoring and mixing episodes 2 through 12, and I like to think that by helping with the presentation of this amazing story, I made the show a tiny bit more enjoyable to listen to.
Working on Serial was a dream job for me, aligning perfectly with my background. I’ve always been a musician, but in 2005 started working in public radio as a producer (and occasional reporter) for shows like On the Media, RadioLab, Soundcheck, Studio 360 and The Brian Lehrer Show. In 2011 I left public radio and started working full-time as a sound designer, re-recording mixer and composer on films and commercials.
Helping out on Serial seemed like a rare chance to combine the three fields in which I work: music composition, radio production and audio engineering.
I’m a freelancer who works out of my home studio, not a member of the staff of Serial or This American Life. But over the course of the season, executive producer Julie Synder and I developed a great workflow that turned the scoring and mixing into a quick and efficient process.
Each week Synder would send over a rough assembly of the episode at some point on Tuesday. This was a Pro Tools session that had a quick cut of Koenig’s narration and all the tape for the episode.
From there I would edit the audio to fine tune the edits and adjust the pacing, write music or use music written for previous episodes, restore archival tape and then finally mix the episode. I would send an MP3 to Synder at some point on Wednesday and she would get back to me with a list of notes. Sometimes they were as simple as, “Add an extra beat after this piece of tape,” while at other times a note would call for an entirely new piece of music.
I’d make those changes and then she would listen to my Pro Tools session in real time through a program called Source-Live. I put the Source-Live plug-in on my master bus and then it would stream my session in real time via a web link. Julie listened though her browser while we talked on the phone, allowing us to make fine-tune edits without being in the same room.
We’d finish the episode at some point Wednesday night so it would be ready to post first thing Thursday morning.
A Film Mentality
From the beginning I thought it would be interesting to approach Serial from a musical and audio production standpoint, more like a film project than a radio show.
Since it was one story told over 12 episodes, it seemed like we could create a musical palette and a high production standard that could link all the episodes together. With the music my approach was to have it sound more like a film score and less like songs from a library used as buttons or stingers.
I wrote pieces for specific moments in each episode; often these were quite simple arrangements. I tried to use a similar palette of instruments (piano, synthesizer, percussion) to create a unified tone for the scoring. The goal was not to create a piece of music that sounded good on its own but to create a score that helped the story.
They say the best score is one that isn’t noticed at all by the audience; I think if we had used more traditional songs instead of score, the music would have been more noticeable and could have detracted from the episodes. (We also used a few pieces by Nick Thorburn, who wrote the iconic Serial theme song, as scoring throughout the season.)
From the mixing side, I also tried to incorporate much I had learned from working in film.
This included mixing and restoration techniques, but mostly just sometimes taking extra time to make something sound better. In film projects I often spend hours on just a small section of tape—isolating dialogue, removing hiss and crackles, or transforming poor quality music. Because you know the audience will be listening in a theater at a loud volume, there is a very high bar for the sound quality; you often have to go the extra mile to get it there. Obviously a podcast is different—people will be listening on earbuds on subway, running or wherever else—but with Serial my philosophy was that if it will help the story, fix it!
This started with the treatment of Sarah Koenig’s narration.
Since it made up such a large portion of the podcast, I thought it should sound as good as possible to help keep the listener engaged. Using EQ (Pro Tools EQ III and Universal Audio’s Neve 88RS) and compression (Universal Audio’s LA-2A and Precision Limiter), I sought to make the voiceover tracks sound tight, under control, but still warm. The specific plug-ins I used happened to work best with her voice and mic; I recommend trying an array of EQ and compressors to find a good fit for your material.
An artifact of the compression and her mic placement was that her breaths were very loud; so as a last step to each episode I manually reduced the gain (using Pro Tools clip gain feature) of her breaths. It was a tedious process but I thought that having her voice sound as good as possible ultimately was worth it.
I also spent significant time restoring all the archival tape. Using an array of plug-ins that I picked up through working in film, I was able to make some very poor tape sound much more listenable. There were many characters whom we never hear in present day or high-fidelity. A main character, Jay, is only heard on poorly recorded police interview tapes from 1999; but so much of the series rests on whether you believe him or not. For that reason, I felt it was worth it to spend significant time restoring the tape with plug-ins like iZotope’s RX and an array of EQ and enhancer plug-ins. It was extremely time-consuming, but once you have a decent-sounding recording the listener can hear more than just his words: They can start to try to read his tone, the way he’s saying what he’s saying, and the tiny quivers in his voice that might be fear, nervousness, or whatever else.
The same was true with Adnan Syed (the central character of the story). We only hear him on a very poor phone line, so I spent a lot of time getting rid of clipping and using plugins to make him sound as good as possible. I didn’t want the audience to get tired of hearing his voice because it was constantly clipping or too noisy.
Finally, I thought the overall mix should sound as good as possible. I used Pro Tools HD’s advanced automation features to EQ, de-ess or add noise reduction any time I came to a piece of tape that needed fixing. I also did a great deal of volume automation rather than adding compression to the overall mix but tried to keep the dynamic range on the smaller side.
Again, people listen to podcasts in all sorts of situations—with car noise, while running on the treadmill, on the subway—and with too much dynamic range they’d have to constantly change the volume of their player.
Unfortunately there is no accepted loudness standard for podcasts, but I tried to have Serial sound “nice and loud” (a very scientific term, I know) on iPhone earbuds with the volume at 75 percent. This translated to a loudness of about –16 LUFS, and I threw a loudness meter on my session to make sure I was always hovering around that point. Finally I placed a brickwall limiter on the master bus set to –3 dB to prevent clipping that might result from digital compressor or all the various ways people could listen to the show.
I hope these details about the production process on Serial can help others looking to make high-quality programming. My job was very easy on Serial. All I could do was make an amazing program just a little bit more polished-sounding. But I believe good production quality and music can make it easier for an audience stay engaged with a program.
Mark Phillips composes much of the music for “Serial” and mixes the show. He is a composer and sound designer and has worked on films such as the Oscar-nominated “Cutie and the Boxer.” When not scoring and mixing films, he releases music under the name Sono Oto.