Editor’s note: Welcome to “Ask the Engineer.” This is the first of an occasional series of technical advice columns from Rob Byers, sound engineer, loudness advocate, and technical coordinator for American Public Media. You can meet Rob — and sign up for a private, 30-minute consultation about your gear or software — at the AIR booth at the Third Coast Conference. On-site sign-up is first come, first served.
To get the column started, we asked Rob to explain an engineer’s approach to audio stories. He went a little beyond that, tapping the engineering brain trust at APM for a wider view. Thanks to Stephen Colón, Cameron Wiley, Erik Stromstad, and Johnny Vince Evans for sharing their expertise too.
The story is done. You’ve been over every word in the script, trimmed flabby copy, tightened clumsy pivots, and the tracking is complete. There’s just one thing left to do.
So you walk into the studio or send your session off into the interwebs for a final polish from an engineer.
She listens so intently, always does “that thing with the plug-in” first, maybe even frowns a little bit before claiming the mix is ready for air. You ask what she thinks about a certain actuality or a specific bit of narration, and the response is, “Oh, I’m not paying attention to the words right now.” Not listening to the words! What?
The piece comes out sounding better when she’s done — but what is she listening for? What magic is between her ears?
First of all (and I think this is the most important point), while engineers may go down some über-technical paths from time to time, we have the same set of priorities that you do
Our ultimate goal is to help you tell your story. We want to prepare your piece so that it is intelligible, has a strong sense of place, and is technically ready for distribution.
So what are we listening for? I took an informal poll of my vastly experienced engineering colleagues here at American Public Media (see further down for their specific comments). We each have a different routine, but it all comes down to timing, intelligibility, and transitions — and the theme of consistency hovers over everything.
This might be surprising because it probably feels like a producer’s concern. Since many engineers have strong ties to music, we find ourselves listening to timing as if it were the rhythm. We’re listening to the timing between acts and tracks, as wells as the pacing and placement of scene changes or music beds — it’s all about being natural.
Depending on the intent of the production, I usually listen for timing that corresponds to the pace of the conversation and the speed of the read. It’s the difference between a quick, percussive “Marketplace” edit and a reflective, conversational “On Being” edit.
Thinking in terms of breaths or beats helps (I actually find myself taking breaths out loud when making timing decisions); others might think in terms of seconds.
Intelligibility is all about answering one question: When the piece is broadcast, will the listener be able to hear and understand every second of the audio?
We want to prepare the audio for the effects of broadcast processing (the compressors and other processors that help transmit the audio) as well as the various listening environments where we find our audiences.
We listen primarily for tone (this is where equalization, or EQ, comes into play) and balance. Bassy or muddy voices aren’t intelligible against the background noise of a car speeding down the interstate, so it helps to filter out those muddy low frequencies from the audio. (I can go on and on about this problem — fodder for a future post!) Bright or sibilant voices can be harsh to listen to.
Balancing multiple elements on top of each other (speech over music, speech over ambi) can be the hardest, because what we hear in the studio does not directly translate to the broadcast environment. It takes experience — lots of critical listening in different environments — to learn how sounds translate and what fixes will be most effective.
You’ve probably experienced this problem when you mix ambi on headphones. You work on the balance of ambi to voice until it is absolutely perfect, only to find that you can’t hear the ambi once it is broadcast. Engineers have tools and experience to help you avoid that problem.
The ear is supersensitive to abrupt changes. Shifts in tone or level between the reporter’s voice and an actuality can be so distracting that they can shift the listener’s focus away from the content — reaching for the volume knob or being distracted enough to stop paying attention to the content. Poor transitions can also come in the form of scene or ambi changes that feel abrupt.
Engineers listen for these changes and try to smooth them out. There are many tools for this, such as EQ, adjusting timing, adjusting level, or masking the changes with ambi. Smoothing transitions is one of those nuanced advantages that engineers bring to the job; it’s really about the experienced ears and less about the tools at hand
Engineers will all have different approaches to these ideas. I’ve compiled some highlights here from my APM colleagues — you’ll see that these all relate, in one way or another, to the ideas of timing, intelligibility, and transitions.
• I usually do a rough pass, listening for cuts that need EQ, noise reduction, etc. At the same time, I’m listening to the ambi and music beds. Then on the second pass, I’m listening less from a technical standpoint and trying to hear the story as the average listener.
• I listen for consistency of levels (can I hear all of the words?), clarity or intelligibility (what needs EQ?), naturalness of pacing and transitions between sections or in and out of black.
• Everything needs to breathe. Quick fade-ins and -outs, or no fades, are no good!
• Speakers who start every sentence with plenty of energy, but trail off at the end, don’t come across very well in the car. Careful volume automation is the best way to address this, but it can be very time-consuming, so I sometimes use a compressor very lightly.
• If I hear a bad edit, I immediately switch to fixing and finding other poor edits.
• Abrupt ambience transitions can jar the listener out of the moment. Jarring transitions aren’t bad when they serve some purpose, but they tend to draw attention away from the narrative, either by sounding amateurish or by waking up our primordial fight-or-flight response. If your actualities don’t have clean ins and outs, you can create them by using a half-second or more of clean ambience to paste on the top and/or tail, to create fade-in/-outs. Your piece will sound a lot classier.
Also, if you’re going back to the same location of the last actuality, it might make sense to maintain that ambi under your voice track. (You did remember to record some clean ambi, right?).
• As a rule of thumb, if I think ambi is competing just a little bit too much with the voice tracks while I’m mixing, it will sound about right in the car.
Engineers want to make the “engineering” disappear from the piece and make it sound consistent. We want to hide the tools and methods applied so that the story shines through — and while we listen technically, we engineers can also be a great resource for content and production decisions (and the earlier you bring us into that process, the more we can help). Struggling with setting the scene, the pacing, or the flow of your story? Experienced engineers have heard it all, and we’ve worked with so many other producers and reporters that we have production ideas aplenty — you just have to ask.
No matter how you listen, it’s all about telling the story.
• Rob Byers (@RobByers1 on Twitter) is a technical coordinator at American Public Media, an NPR “alum,” a classically trained percussionist, a jazz enthusiast, and a member of AIR’s network. Over the years, he has been lucky enough to engineer recordings with Yo-Yo Ma, Rufus Wainwright, The Knights, Béla Fleck, Joshua Redman, Josh Ritter, Yo La Tengo, Cyrus Chestnut, and many others. He gave a recent, terrific webinar about audio levels, and he thinks everyone needs to know about #loudness.
Have a question for Ask the Engineer? Email firstname.lastname@example.org.