Lessons in Mixing A (Historical) Documentary for Public Radio

July 30th, 2014 | by Jeff Towne

HistoricalDocScreenShot1

Intro from Jay Allison: Making historical audio documentaries is tricky. All the action is in the past, for one thing. Plus, you typically have masses of research to do, lots to organize, and, if you’re lucky, scratchy archival recordings to incorporate.

Transom has two new features to help. First, Yowei Shaw and Alex Lewis describe what they learned producing their first historical documentary for radio, and help you avoid beginner’s pitfalls.

And who better to share insights on mixing documentaries than Transom TOOLS Guy, Jeff Towne, who worked with Alex and Yowei on their piece. He includes tips on making archival audio sound beautiful, along with a general guide to making your hour-long piece broadcast-friendly for radio stations. This will be very helpful for anyone who must encounter the Clocks of Public Radio Land.

Structure Your Piece So It Will Get Aired

It’s a difficult task to edit-down many hours of tape, add narration and ambience and music, and get it to make sense and flow. But that’s not all you need to think about if you’re making your documentary for radio: you need to adhere to a few conventions of timing and structure.  If you’re only presenting the program on a website or as a podcast, or in a public performance, you can make it as long as you want, you can make it continuous or take breaks at random spots.  But if you’re hoping to be aired on Public Radio stations, you need to make it convenient for stations to air. There’s not a single correct answer to how a program needs to be formatted, but there are a few common structures that will help make your program appealing to stations.

The first convention is one of length: make your program an hour long.

Stations look at their schedule in hour blocks, and while a rare Program Director might be willing to find several shorter programs to fill an hour, they’ll be MUCH more likely to carry your program if it’s a complete hour.  If it’s longer, it should be in hour-long blocks: an even two hours, or three, or more. If the program is longer than an hour, you should structure it so that each hour stands alone:  the station may want to air each hour on different days. Or they may choose to run the hours back-to-back, so you should think of a way to write the end of each hour that supports either scenario.

An important point: in radio, an hour-long program is 59 minutes. Every station needs to have time between programs to insert a local spot or two, or do a live announce, or play a legal ID. So no matter how long your program, each hour needs to run 59 minutes.

Some stations might be willing to run your program as a continuous 59-minute block, but it’s much more attractive to them if you format the program in a few discrete chunks, with time for a cutaway or two for the station to insert local content. Options are:

  • At least one 30-second break near the half-hour point.
  • Preferably, two 30-second cutaways, placed somewhere around the 20- and 40-minute marks.

Although most stations will want to place a local spot in these 30-second holes, some might choose not to, so don’t simply leave that space silent, create 30-second sound files of music, or some other audio space-holder.

HistoricalDocScreenShot2

Many public radio stations run the NPR newscast at the top of the hour. You can hope that they’ll skip the news during the airing of your program (more likely for a one-time program than for a multi-part series) or you can make it easy for stations to insert the news into your program. The NPR newscast starts at one-minute after the top of each hour, and is exactly five minutes long. To make your program newscast-friendly:

  • Create a one-minute-long “billboard” to intro the show.
  • Create a segment that’s exactly five minutes.  Remember that any station that uses the news will NOT air that 5-minute segment, so you can’t put anything important in there. In most cases, like the 30-second cutaways, music is used to fill that space.
  • You may want to include a short announce at the start of that cutaway explaining what’s happening: “Before we get to our story…”

It’s important that you keep cutaways in mind when writing your narration: at the end of each segment, you’ll want to forward-promote the remaining portions of the program, and include a phrase such as “…when we continue.” If you’re using the optional news cutaway, be careful about your phrasing: many programs end the opening Billboard segment with “…first the news” but this sounds awkward if the station does NOT cut to news and instead uses the filler segment that you’ve provided.

If your program consists of multiple hours, at the end of each hour, include a forward-promote of the next hour, but don’t be too specific: stations may air the hours following one another, or on different days. Leave room for the local station to insert their own tag: “Coming up next, on ….” or “Tomorrow night at 8, on…” or “Wednesday at noon, here on…”

These days, most radio stations use a computer-based system to play each audio element. Each piece of audio might be triggered by a live operator, or by an unattended automation program. It’s easiest for the station to assemble all the pieces if you provide each part of your show as a discrete audio file, in “segments.”

HistoricalDoc-Clock1
In the simplest scenario, your hour-long program should have three segments:

  • Segment One is 29:30
  • Segment Two is :30 for a local station cutaway
  • Segment Three is 29:00, for a total of 59:00

HistoricalDoc-Clock2
Even better, make five segments, to create two cutaways.

  • Segment One is 19:30
  • Segment Two is :30
  • Segment Three is 19:30
  • Segment Four is :30
  • Segment Five is 19:00

HistoricalDoc-Clock3
To create a news cutaway, make seven segments.

  • Segment One is :60
  • Segment Two is 5:00
  • Segment Three is 13:30
  • Segment Four is :30, Segment Five is 19:30
  • Segment Six is :30, Segment Seven is 19:00

Here’s an interesting example of an even more complex program clock, from 99% Invisible.

It’s crucial that all segments add up to 59 minutes, exactly, but the lengths of the main segments can vary. It’s best to locate the local cutaways somewhere close to the 20-minute and 40-minute marks, but it’s fine if they occur a few minutes earlier or later. Because most station automation systems simply play a sequence of audio segments one after the other, the segments can be random, irregular lengths, as long as they add up to 59:00. But it’s good form to make the segment lengths a round number (whole minutes, or 30-second increments) if possible. That makes it easier for stations if they want to do live announcements during the cutaways, and it makes it easier for you to make sure that your segment lengths add up correctly.

GoingBlack-BB-pad

Whatever the length of the segments, each one should be formatted in a specific way: the sound files should be exact lengths as described above, AND: the actual audio should not start right at the head of the file, nor should it extend all the way to the end. Leave a half-second of silence at the head of each segment, and leave a second of silence at the end.  This allows for more natural pacing when segments are played-out in succession.

GoingBlack-SegmentStart

Make your archival audio sound beautiful.

When working with archival recordings from a variety of sources, you’re bound to run into some noisy audio. To a certain degree, crackly, scratchy, muffled, even buzzy audio is OK: it signifies that it’s old, that it was recorded on old media, that it has a history. But if the noise is distracting, or makes it hard to understand the content, it’s worth trying some clean-up.

Here are a few possible issues and some simple solutions:

  • If the sound file is in stereo, determine whether the source audio is stereo or mono.
  • Old recordings, especially those made on tape, might have different kinds of noise or degradation on the left and right channels of a stereo track, creating a twitchy, unstable, or unbalanced stereo sound.
  • If the source audio is actually in mono, such as from a single microphone, it may sound better to use only the left or right side of the stereo sound file.
  • Most digital audio editing programs have a function to split a stereo track into two mono files. We were using Hindenburg, and that function can be accomplished by highlighting a stereo clip, then going to the Tools menu and choosing Split Stereo.
  • Listen to each of the resulting mono files and determine whether one sounds better than the other. If so, keep that one and delete the other clip, panning the remaining clip to the center.

GoingBlack-HighPassEQSome audio clips can be improved significantly by simple EQ. The original recordings may be flawed, or a subsequent transfer to another medium may have been done poorly, resulting in muffled sound, grounding hums, or other problems. Some of these issues can be mitigated by using an equalization plug-in.  In our project, several old clips were improved by simply reducing low frequencies with a high-pass filter, perhaps adding a slight boost to high frequencies with a high shelving filter.

For the most troublesome audio, broadband noise-reduction software is the best tool. There are several programs that work similarly: the Waves X-Noise and Z-Noise plug-ins, Sonic Solutions No-Noise, Sound Soap, Izotope Audio Rx, and others. Good de-noising tools are built into the Adobe Audition and Audacity editing programs.  For this project, we used Izotope Audio Rx 3 Advanced to reduce unwanted hums, buzzes and rumbles.

GoingBlack-AudioRx2In most cases, finding some noise in the clear and allowing the de-noising software to “learn” the noise gave the best results, but occasionally, Izotope Audio Rx3 Advanced’s “Adaptive” algorithm worked better. Although some of the de-noise programs can run as real-time plug-ins in your audio workstation, the sound quality is usually much better when run off-line as a non-real-time process.

GoingBlack-AudioRx1Even the best de-noising programs can create audio artifacts, usually a swirly, watery or metallic sound, so this processing should be used carefully. Sometimes the original noise is preferable to the cleaned-up version.  But when used judiciously, broadband de-noising can improve the intelligibility of your audio clips, and the smoothness of your overall mix.

With careful use of noise-reduction and EQ, old audio clips can fit more smoothly into your documentary. The goal isn’t to make those archival recordings sound pristine or perfect, but to make the content more intelligible, and eliminate, or mitigate, extraneous noise that may be distracting or annoying.

These techniques for cleaning-up audio, and for formatting the timings of your program’s sections, can apply to any audio production, not just historical documentaries. When preparing any audio for broadcast on the radio, it’s important to consider what would be most convenient for stations to use. If you can format your program so that it sounds smooth and professional on the air, with easy-to-use cutaways for local breaks, and mix it so that there are no annoying hums or buzzes, even on archival audio, you make your productions more attractive to stations.


One Comment on “Lessons in Mixing A (Historical) Documentary for Public Radio”

Links to “Lessons in Mixing A (Historical) Documentary for Public Radio”

Leave a Comment