Since the early 1980s, Ken Burns’ name has practically become synonymous with the historical documentary genre. He is best known for the epic multipart miniseries he’s shepherded in his various capacities as executive producer, producer and/or director: The Civil War (1990), Baseball (1994, adding two more episodes in 2010), Jazz (2001), The War (2007), National Parks: America’s Best Idea (2009) and last year’s Prohibition. His latest, which recently aired on PBS (his home for decades), was a two-parter called The Dust Bowl.
About to be engulfed in a gigantic dust cloud is a peaceful little ranch in Boise City, Okla., where the top soil is being dried and blown away. April 15, 1935. Photo courtesy Associated Press
To get a glimpse into the sound world of Ken Burns’ documentaries, we interviewed re-recording mixer Dominick Tavella, who has been the mixer on all of his films since Unforgivable Blackness in 2005.
Tell me about how a show such as The Dust Bowl comes to you and what your stages of development are in terms of mixing the different elements.
It’s the Midwest in the late ’20s and early ’30s; everything dried up and blew away. They go into the historical reasons for what happened and lay out very clearly the result of what happened. The Dust Bowl is fairly dramatic because the dust storms are so huge, and we tried really hard to get the sense of a mile-high and 30-miles-wide wall of dust coming at 80 miles an hour, totally covering everything, blacking out everything in its path. People can’t breathe; they have to crawl on their hands and knees to go from house to house because it is so thick.
Is the narration recorded first?
They usually record the narration over a period of time because there is a script, of course, but the interviews change over time as the film is being edited, all the way up to the last minute. So sometimes the narration is affected by that. Of course, [Burns] is very organized, so the “last minute” is months before it airs. But at any rate, it takes a while for everything to cohere to its final form, and the final form is a mixture of narration, voiceover, the experiential voices—people reading quotes—and on-screen interviews.
Farmer and sons in dust storm, Oklahoma, 1936
By the time a film gets to me, it’s virtually solid. Of course, they’ll have piles of voice tracks, and there are usually quite a few voices that recur from episode to episode. I’m very careful to keep track of all the settings for each voice as a starting point, and then I’m careful to mold it so it feels good with what’s behind it, and it will consistently match throughout the course of the film.
The way I usually work is I’ll do a first pass doing the voiceover and the on-camera interviews and music. With documentaries, what you want is a clean consistency to the voices. And you want to control the impact of the voiceover. Then I’ll do a second pass where we’ll do the effects, and for the effects there are premixes. Then I’ll do a third pass where we combine everything into the final track. But because of the nature of how we’re doing it and because of the flexibility of the console [an AMS Neve DFC], I do it in such a way that at any point, no matter how far along in the process we are, I can go all the way back to step one without having to reload anything into the console. This is great for Ken because when he comes in for his final playback and he says, “This voice isn’t right, we have to move it a few frames back,” I don’t have to rip the board down, put up the board for the dialogue, put in the dialogue tracks, reset the board for the final. All I do is press a couple of buttons, move it a few frames, press a couple more buttons and we’re back. So the whole process has become very self-contained.
Is everything arriving to you in Pro Tools sessions at this point?
So you’re primarily working with a number of stems supplied to you?
Yes. The dialogue stem would be left-center-right; then, depending on how thick the voices are, I’ll have two or three tracks for on-camera interview, and then I’ll have two or three tracks for actual footage of FDR or an interview with Woody Guthrie, or whatever.
Re-recording mixer Dominick Tavella
The music stem is usually a 5.1. The effects stem is usually a 5.1, and then I’ll have two mono tracks, which are often Foley, right in the center. I like to keep those separate from the backgrounds because when you put them all together, those are usually the things you adjust the levels of the most. Particularly with Foley for historical footage, you have to be very, very careful. You don’t want to make it feel too current. You don’t want the audience to feel for a second that this sound they’re hearing is possibly manufactured by you.
In The Dust Bowl there are those sequences of the wall of dust coming toward you, so we tried very hard to make it feel like it was coming up and passing over and really surrounding you. You do that by going through all the sounds and figuring out which sounds you want to move where, and how to move them. It’s difficult to talk about—I do so much of it instinctually, without analyzing it. I’ll have four winds and this one is sort of deep and throaty and that’s not going to move a lot—that’s going to be like the undercurrent—and here’s a wind that has a little more high-end blow to it, and this is the one we’re going to feel going over our heads. That kind of thing.
How long did you work on The Dust Bowl?
At this point, the rough average is about two weeks per two-hour episode. Sometimes it takes longer, sometimes it goes a little more quickly, and the time has expanded through the years and through projects.
There’s more material because we’re doing more with the sound. In The Dust Bowl we would sometimes have eight or 10 different winds weaving in and out of each other over the course of a one-minute sequence. You can still open the faders and have something that’s fine for basic cable. But all of Ken’s stuff is very carefully structured, almost like feature film in terms of quality and depth of sound.
As a black blizzard rolls in to Ulysses, Kan., two women and a girl pose for a photograph before taking shelter. Photo courtesy Historic Adobe Museum
At what point does Ken get involved in the sound?
He’s involved in the spotting of the sound as they’re editing it. Then we sit down with Ken and the sound editors and we go through a couple of the episodes shot by shot, scene by scene. He’ll say, “In this scene, I want to feel this, I want to feel the car coming around here. We’ll get quiet here and we’ll have some winds...”
So he’s pretty perceptive in that way.
Oh yeah, he’s very perceptive. At the same time, though, he’s not a micro-manager, which is terrific. And the reason for that is he knows he can trust his whole crew. We know what he wants and we have a really good sense of how it will come together, and he knows what our notions and ideas are like. It’s at a point now where he basically will come for a screening for playback of the “finished” track and he’ll give us some notes and then we’ll fix them and that’s that. Basically that’s done in a day. These playbacks are getting shorter and shorter because we’re nailing stuff. He’ll come in and he’ll say, “It’s good!” Everybody here is incredibly proud of the work we do.