Meet producer/mixer Rob Chiarelli (Christina Aguilera, Janet Jackson, Pink) and violinist/arranger Christine Wu (American Idol, Justin Timberlake, Celine Dion). Their names consistently appear on albums and film scores for a reason—they understand that music is supposed to be, well, musical.
“A string instrument is a piece of wood—sometimes, hundreds of years old—that resonates,” says Wu. “It’s a very natural and beautiful sound, and you must preserve that sound as much as possible.”
Once you have all your string tracks recorded, then what?
Wu: One of the most important things about mixing the strings is that they should be there where they need to be, but they should not take attention away from anything else.
Chiarelli: I listen to each track individually so I know what I have to work with. Once I get a picture of what the parts are, and how they work within the song, I draw some conclusions. The first consideration is the sound, or the balance of the orchestra or string section. If the dynamic range is too wide for a pop record, then I’m going to have to compress it somewhat. With strings I try to use gentle compression. I don’t want to be too heavy-handed.
Specifically, what do you mean by “gentle” compression?
Chiarelli: With strings, I tend to like a real low ratio—1.5:1, 2:1, or maybe 3:1. I don’t think a ratio of 10:1 sounds musical. If I need more gain reduction, rather than asking one compressor to do a ton of work with a high compression setting, I might ask two compressors to do a little bit of work—each with lighter settings. That way, they’re just gently riding the gain. If your attack and your release are set properly, the compression should sound pretty transparent.
So how do you determine your attack and release settings?
Chiarelli: The string envelope is pretty slow, so you don’t want to have too quick of an attack. Depending upon whether it’s an individual instrument or an orchestra, your attack is going to be in the range of 10ms to 30ms, and the release will probably be somewhere between 80ms and 160ms. However, you may have a staccato section that requires a faster release, and if a section has a ton of long notes, you may even need a release time as long as a second.
Which compressor plug-in do you like to use?
Chiarelli: I think the Waves C4 is probably the most musical compressor.
Do you find yourself using EQ?
Chiarelli: If something is recorded correctly, you don’t usually need a whole heck of a lot of EQ.
Wu: From time to time, Rob will ask me to use a different microphone to get a sound that will pop in the track more. We try to anticipate mixing problems when we record, so we don’t have to use EQ later on.
Chiarelli: However, we have found that strings recorded in a small room tend to produce undesirable frequencies if the track is heavily layered. So, on a pop record, I might use a high-pass filter to roll off an 18dB slope starting somewhere between 60Hz and 80Hz. That’s the area where you’re not really affecting the fundamental sound of strings, you’re just cleaning up some bottom end that might conflict with, say, an electric bass.
What are your go-to reverb plug-ins, and how do you apply them?
Chiarelli: I like Digidesign Revibe and Waves Renaissance. I usually begin with their out-of-the-box standard hall. I don’t make a ton of changes to the preset, but I make sure the balance between the samples, the real strings, and the reverb feels rich and real. I start by listening to the strings without any reverb to get a sense of where that balance might be. Then, when I start blending in the reverb, there’s a point where it begins to feel about right. It’s an instinctive thing, but it also comes from experience. Depending upon the size of the string section, I usually start with a pre-delay between 24ms and 40ms. A good rule of thumb for early reflections is 24ms—which essentially translates to 24 feet from a wall. The reverb time—depending upon the section—is between two and three seconds.
What do you do if, no matter how much you process a sample, it just doesn’t sound like it fits in the mix?
Wu: If a sample isn’t working, or if it needs something more, I’ll add a couple of passes of real violin and/or cello to give the whole section a texture that makes it sound more realistic.
Chiarelli: Her articulation and phrasing combined with the synth or sampled strings is night and day. The combination gives you a huge sound.
What’s the biggest mistake one can make when mixing in a string section?
Chiarelli: I don’t like it when I hear something that sounds as if it was recorded in 20 different locations—a vocal with a super-bright reverb, an orchestra with a super-dark reverb, a guitar with a super-boingy reverb on one side and some bizarre reverb on the other side. Don’t just start turning knobs—that’s why records don’t sound like records. If you want the vocal in front of the mix, the way to make it more present is to have a little bit less reverb on it, and to have a different early reflection time. Naturally, the orchestra would be behind the singer, so the delay time should be slightly different, and not as loud. The listener is standing in front of the audience, and the diva is out front singing. Therefore, the shortest delay is between the vocalist and the listener, and the largest delay is between the percussion—or whoever is in the back of the orchestra—and the listener. At the end of the day, this thing has to sound cohesive and musical.
Wu: That’s why you keep your settings simple. If you get too fancy, you’re going to screw up the rest of the mix.
Chiarelli: That’s right—it’s a game of inches, and you have to make sure the song wins in the end. The string arrangement exists to enhance the melody, so put the most musical surroundings you can around that melody, and make it pleasing to the listener.