We all know that voice is king. Of the six instrument categories we've explored in our “Recording the Band” series, the voice is arguably the most powerful. It's how we communicate. It is also one of the first sounds we hear as humans. It drives us to remember a song, and even where we heard it and what was going on in our lives when we heard it. Because the voice is the focus of most listeners' attention, recording and mixing it correctly is crucial to the production process.
MIKING THE VOICE
Some tried-and-true vintage mics are first choices for engineers. The Neumann U47 comes to mind as a first-call mic, as does the Telefunken 251 and AKG C12. Other choices include the SE Electronics Gemini, Audio-Technica 4050, DPA 3541, Crowley and Tripp Studio Vocalist, BLUE Bottle, Shure KSM44 and the AKG 414. However, an inexpensive dynamic mic such as the Shure SM7 or Sennheiser 421 can also be perfect for the job. Generally, a cardioid pattern is best to keep the vocalist in, and the room out of, the mix. In addition, a cardioid pattern brings the proximity effect into play, allowing you to adjust the low frequencies by moving the singer away from the mic until the desired blend is achieved.
Your recording space is a crucial factor, whether you're recording a single vocalist or a small section. It's important for the vocal to be as isolated as possible: If the room is small and reflective, then the mic will pick up slightly delayed reflections along with the source vocal, causing phase problems. If the room is large, then the echo can create an ambience that may not work in the track. In either case, it's a good idea to remove room problems by building a small “room” around the singer(s) with the dead or diffusive side of some gobos. This “mini room” can be adjusted to the size of the group and room by moving the gobos closer to or farther from the singer(s) as needed. In addition, placing an area rug under the singer(s) eliminates foot-tapping noise and provides additional absorption. The music stand is resonant and reflective, causing phase issues. Placing a piece of carpet or acoustic foam across the stand, under the music or lyric sheet, will take care of this problem. If the room has an ambience that you wish to keep for a surround mix or to add to a stereo mix, then it's best to capture it with dedicated room mics.
When miking a single vocalist, the performance and style you wish to convey will dictate how close to mike the performer. Is it a Tori Amos — like approach where every breath is in your face or a Broadway show — style vocal recording where you want to hear the singer belt it out to the bleachers? The intimate approach requires a close-mic placement — within three to six inches of the singer's mouth. This configuration may lead to stray “p” or “t” plosives that cause puffs of air to hit the capsule, making for nasty low-frequency booms or pops. This is where a pop filter can help immensely, although it is not acoustically transparent — there is some loss of high frequency. An alternative is to record without a filter and aim for a placement that eliminates the majority of the pops and leaves whatever gets recorded to be addressed later. This means you wouldn't have to add as much EQ to overcome the pop filter's effect, causing less potential phase shift.
When close miking, face the mic toward the singer's mouth, oriented from the area above the mouth in front of the nose. This position lets puffs of air go under the mic, while still capturing the voice. If recording a singer from a distance, then place the mic in front and above the vocalist, anywhere from one to three feet away from the singer's mouth and six inches to a foot above.
MIKING VOCAL GROUPS
The mic models mentioned above also work well for recording vocal groups. When recording two or more vocalists, special placement is necessary to accurately capture all the voices in the group. Place the singers in a tight semicircle around a single mic, standing nearly shoulder-to-shoulder. It's best to use one mic for five or fewer singers and let them find their own blend. This technique saves a lot of track real estate and makes the final blend easier when mixing because you have fewer faders to deal with.
Stand singers in small vocal groups shoulder-to-shoulder in a tight semicircle, facing a single mic.
photo: Chris Bailey
A great way to keep the blend consistent during a session is to put one small piece of white console tape on the music stand for each singer. Once the group has had time to run through the track a few times and establish a blend, instruct them to each take a piece of tape and put it on the floor at the tips of their shoes. This way, if they take breaks, they can come back and place themselves precisely in the same spot.
For larger groups, a riser and multiple mics are necessary to get everyone's mouth within earshot of the mic. For a group of 20 singers, for instance, an ideal configuration is a two-tiered riser with six singers on the floor and seven each on each riser. Or with a single riser, you could place them 10 and 10. The important thing is that the setup should fit the room. You don't want singers in close proximity to a wall, causing phase problems with reflections. Having “air” between the group and the walls keeps those reflections to a minimum, so keep this in mind when deciding where to place your risers. Also, leave some space in front of the group for the sound to develop so you'll have more of that space in the recording.
Each riser arrangement calls for a unique mic setup. The tighter three-tiered approach can be covered with a stereo pair, but the wider two-tiered approach will require three mics. When using pairs or groups of mics, it's best to use three of the same model to keep transient response, frequency response and sensitivity consistent. A tape measure can really help when placing mics in front of a group. Take precise measurements to be sure that the mics are equidistant from the group, then try placing the mics six to 10 feet back from the group and use your ears to determine if you're getting a good blend from all of the sections. If one side of the group is favored, then try pulling the mics back or rearranging the singers so the stronger members of the “weaker” side are closer to the mic.
For a large group, a Decca Tree arrangement — three omnis placed in an equilateral triangle, four to six feet on each side — above the head of the conductor can work very well. If the group is especially large and placed wide on the risers, then placing a single pair of wide cardioid mics to the left and right on either side of the Decca Tree can help capture the singers on the outside of the group. (See the figure on page 40.)
The placement of the soprano, alto, tenor and bass sections is important. If the singers are used to singing in a particular setup, then it's best to keep them in their comfort zone rather than move them around for the sake of the recording. For maximum mixing flexibility, the room should be isolated by recording it to its own tracks. This will save you a lot of pain later on by allowing you to mix the room in to taste.
When miking a large vocal group, arrange a Decca Tree (three omnis placed in an equilateral triangle, four to six feet on each side) above the head of the conductor. If the group is arranged in a wide configuration, supplementing the Decca Tree with a stereo cardioid pair (one on either side) can help capture the singers on the ends.
EQ AND COMPRESSION
There are various schools of thought about the use of EQ when recording vocals. Some take a hands-off approach, letting the mic and signal chain bear the burden of accurately capturing the signal. However, sometimes you don't have the luxury of using the best of the best in your chain or room, so as a corrective measure, it might be necessary to add some top end or to roll-off some bottom if the situation calls for it. Just don't paint yourself into an EQ corner by going overboard.
EQ approaches can change when recording is in different formats. Analog tape, for instance, can lose higher frequencies overnight — as well as lose some of the luster that sounded so good during the session. If you have to add “air” later, you will also bring up the noise floor, so adding some preemptive top can help fight the laws of entropy as the tape changes over time.
Digital recordings, however, don't have this problem: If it sounds good today, it will sound very much the same tomorrow. Less is more here. Go with your ears. If you don't think it needs EQ, don't add it — you can always do it later. If, however, you need to tweak, adding a few dB at 10k with a shelving EQ can brighten the track. Filtering out the extreme low end starting at 100 Hz or so can also take some of the tubbiness out and remove low-frequency room noise.
Careful use of a good transparent compressor can make a vocal track sit down in the mix and allow every word to be heard without sounding crushed to death. In general, you want to compress lightly while recording and then add more during the mix. This approach keeps you from getting heavy-handed during the mix and makes the process more transparent because you're performing it in stages.
On a full-featured compressor, try setting the attack to 40 milliseconds and the release to 400 with a 3:1 ratio. Then set the input/threshold so you're getting about 7 to 10 dB at the maximum peak. Setting the attack to a lower number and lowering the threshold while increasing the ratio will render a more “crunched” sound, one that is very popular in pop (Kelly Clarkson, Avril Lavigne, etc.). For a very dynamic track, some engineers will use a compressor for overall gain reduction and then use a limiter set to a higher threshold to catch the odd phrase that gets through. This technique creates an overall smoothness that the compressor alone cannot provide.
Remember that every compressor has a distinct personality. A Universal Audio LA-2A will not sound like a Behringer Composer or a Waves Renaissance Compressor plug-in. Find your favorite for a specific application and add that personality to your bag of sonic tricks.
COMPING AND TUNING
The recorded performance has to be up to snuff for the mix. This is where editing and tuning come into play. A popular way to get a great vocal performance is to make a comp. This is when a singer performs a number of passes of the song onto separate tracks and then the producer and artist listen to each pass line by line, deciding which versions go into the final master track. The individual parts, which can be as short as syllables and as long as complete lines, are then recorded to a new track called the comp, which is then used as the master.
Once the master track has been established, whether it's a single track or a comp, some tuning might be in order. Automated pitch correction is a relatively new process, and both software- and hardware-based tuners are available to handle the job. Antares Auto-Tune software is widely used, especially in country and pop. Other tuners include Celemony Melodyne, Apple Logic 7 pitch utility, SoundToy PitchDoctor, Serato Pitch N' Time and T.C. Helicon Intonator. Whichever tools you choose, use with taste and in moderation. Some believe that heavy tuning sucks the life out of a vocal and makes it sound mechanical and inhuman.
Combining a tuning plug-in with automation can save valuable time in a pinch. First, find the problem spots and set the tuner's parameters until the problem is fixed. Then, automate the master bypass function. Write the plug-in to bypass for the entire mix and then only turn it on during the problem areas. If there is one particular bad note, then you can automate other parameters so the plug-in tracks better. This process saves time rather than dealing with each note individually.
DE-POPPING THE TRACK
Pops are easily addressed using a DAW such as Pro Tools. Three offline techniques work very well. One easy trick is to create a quick manual volume edit of a fader move at the precise point of the pop. First, find the pop along the timeline and put your cursor in the center of it. Then, using the pull-down to the left of the wave, choose to look at the volume graph rather than the waveform itself. (This appears as a straight black horizontal line with the waveform grayed out behind it.) Then, precisely before the pop occurs, use the hand tool to enter an edit point by clicking on the horizontal line. Do the same at the end of the pop and then at the middle. Use the same tool to grab the middle point and pull it down, creating a “v.”
Be sure to zoom in quite close so that the fade is executed in a half-second or less and the volume move will be quickly executed. When listening to the move, if the edit is too apparent, make the “v” narrower. If it's early or late, then highlight all three edit points with the cursor and use the plus and minus keys to adjust all three points according to the nudge settings. The finer the nudge, the smaller the move. (A 10-sample range works very well here.)
Another method involves performing the same kind of volume edit by using the gain setting of a parametric EQ plug-in's low-frequency band. Find the offending frequency by creating a narrow band dip starting at 100 Hz. Leave the band in as you repeatedly listen to the offending pop in Loop Playback mode. Of course, this dip would never work if placed on the entire vocal track, but that's the point — only use it at the precise point of the pop. Once the frequency and Q have been determined, push the Auto button on the plug-in and add the band's gain control to the automation list. Once you do this, it will show up in the same pull-down as the Volume/Waveform view. Choose to view the EQ's band that you just added to the automation, which should look exactly like the volume wave view. Now perform the same edit mentioned above, making a “v” precisely where the pop occurs.
The last pop-reduction technique is to actually cut into the waveform itself. Once again, isolate the pop, place your cursor just before the event and zoom way in. Take a sliver of a cut out of the wave just before the pop by highlighting it with the cursor and hitting the Delete key. Then draw a very small fade with the cursor (Apple + F) at the left edge of the cut so that there is no pop across the cut. Draw a larger fade into the pop, then place the cursor before the edit and listen to see if it works. If the pop is still apparent, then use the Trim tool to lengthen the fade on the right side, making it dig deeper into the pop. If you can hear the “hole,” then you made the original cut too large. After performing a few of these maneuvers, you'll get the hang of how big or small the fades and cuts need to be.
DEFINE IT IN THE MIX
When mixing, adding a bit of top end or thinning out a boomy vocal can help define the track in a busy mix. In addition, adding some compression during the mix can help with intelligibility, allowing vocals to slug it out with distorted guitars, keyboard pads and other potentially scene-stealing elements.
If a vocal calls for reverb, then go with a high-quality 'verb. This includes plug-ins such as the Waves IR-1, Trillium Lane TL Space or Altiverb, or a hardware unit such as the TC Electronic System 6000, Lexicon 960L or Eventide Reverb 2016. The reverb is the icing on your cake and arguably the most important added element to your lead vocal track. Pre-delay creates a dry space directly after the event, before the reverb kicks in. For vocals, it effectively sets the reverb back in the mix behind the vocal, creating a 3-D relationship with the vocal. It increases intelligibility and, when set properly, focuses the listener's attention on the vocal, not the effect.
Reverb time and pre-delay settings are directly tied to the song's tempo. A slower tempo can support a longer RT60 and pre-delay because there is more space in the music to support the longer reverb time. For ballads, try setting the pre-delay at 120 ms and then play with the reverb time, starting at two seconds. If the track is swimming, you can lessen the reverb time, but first decrease the volume of the reverb by bringing down the output of your return to the console. This will mask the tail by tucking it under the track, but still leaves a lush and long effect if the track's instrumental portion drops out and the vocal is left alone to support the mix.
Timing a delay with the song's tempo can be a great way to treat a vocal in a ballad. It provides a dreamy repeat at the ends of phrases that span the gap before the start of the next phrase. The delay can be dynamically altered by automating an EQ plug-in on the return track and changing it over time or by using a modulating delay. If you don't have a delay chart, then you can grab a quick quarter- or eighth-note delay by using a runner's stopwatch with a 100th-of-a-second readout. Start the clock on beat one of any bar and then do split times at beat two of the second bar and beat three of the third bar. Those two split times will be your delay times. For a more wet delay, you can send the output of your delay to your vocal reverb.
Given that vocals are the most important part of a song, I can't stress enough the importance of taking your time with this part of the production. By using some or all of the techniques above, you can be assured that if the melody and lyric are right and the singer does his/her job well, you will end up with a recording that has the potential to reach the tip of everyone's tongue.
Kevin Becka is Mix's technical editor.
Recording Monks in Surround
Four years ago, I had the opportunity to record a CD for the New Camaldoli Hermitage (www.contemplation.com) in Big Sur, Calif. Along with Glen O'Hara, a colleague from the Conservatory of Recording Arts, I ventured up the coast of California for five days to capture the monks in their “native habitat,” a beautiful 800-acre piece of land overlooking the Pacific Ocean. The room in which we recorded was the monks' chapel, a rectangular room attached to an octagonal dome that rose 30 feet. The ceiling was tongue-in-groove wood and the floor was an unpolished granite. This combination made for a lush recording space and the most interesting acoustical environment I've encountered. The RT60 was more than four seconds, and because of the shape of the room and the “upside-down wedding cake” diffuser built into the ceiling of the dome, the reverb was very balanced, with no unevenly balanced frequencies — the perfect place to record the chants of monks.
The monks preparing for a vocal take in the rectangular chapel
The gear for the project was donated by various manufacturers, including a 16×2 tube mixer with mic preamps from Manley, four CM6 microphones from Schoeps (two cardioid, two hypercardioid) and a Genex GX8000 magneto optical 8-track recorder. We used Canare Star Quad mic cable and good old tape-based 20-bit ADATs as backups.
The plan was to record three smaller groups of monks for three consecutive days, ending up with 16 tracks that I could mix both in stereo and surround. (A short surround recording can be downloaded from www.mixonline.com.) I originally wanted to keep the recording on the Genex at 88.2 kHz and 24-bit right up until mastering, but a problem with the recorder on the second day made the use of backups a disappointing necessity. For the recording, a group of six to eight monks stood at the far end of the rectangle facing directly into the domed area. Two hypercardioid mics were set up in an X/Y pattern to capture them directly, with the two cardioids set up in an ORTF pattern pointing into the dome about six feet from the floor.
For the overdubs, the plan was not to use headphones, as we wanted to keep the experience as familiar as possible for the monks. Instead, two speakers were placed precisely at the off-axis side at the back of the mics pointing at the monks and flipped out of polarity so there was minimal bleed into the mics. The other mics were behind the speakers and pointing away, so they were not a problem. We ran some level tests to be sure that whatever reflected bleed from the room was not obtrusive and then did all the overdubs that way. At the end, the combination of the main feed and room mics worked perfectly for both surround and stereo mixes.
— Kevin Becka
THE STORY BEHIND THE SOUND
Capturing the Vocal Harmonies of the Little River Band
Producer John Boylan has produced a string of hits for vocal-heavy bands and artists such as Linda Ronstadt, Pure Prairie League, Boston and The Dillards. Boylan was initially impressed by the Little River Band's vocal strengths. “When they first approached me, the thing that struck me was that their sound was an interesting twist on that California country rock three-part harmony,” he says. “They had that strident middle guy who gave it more of a Hollies thing, which made for a nice combination of those two ideas.” Tracks for the band's 1977 record, Diamantina Cocktail, the first of four albums Boylan would produce for the band, were recorded at Armstrong's Studio (later named AAV) in South Melbourne, Australia, with staff engineer Ross Cockle. Boylan found that his recording technique evolved past those records. “When I first got down there, I used what I had used in L.A., which was an amalgam of techniques I refined over the years working with different background singers including Venetta Fields, Clydie King and Mary Clayton,” says Boylan. “They always sing around one mic, taking one ear off the headphones so they could blend themselves in the room.” He decided that placing the singers in a semicircle in front of a cardioid-pattern mic provided the best opportunity for the singers to balance themselves.
By the third LRB album, the recording budget expanded, as did the number of tracks available, so the tracking technique changed. “We refined it to individual mics (and tracks) for each singer, as well as custom self-mixing monitor systems,” says Boylan. The strength of this approach was that a different mic could be used on each individual singer. The mics were placed in a triangle and each singer sat in front of their mic so they could see the other singers. This approach worked very well, allowing singers to mix their own headphone feed and hear exactly what they needed to nail the part. However, Boylan feels that the original single-mic technique also has merit. “There's something magic about the blend in the room — moving around until the phasing and everything is just right,” says Boylan. “Great singers do it instinctively, although it is harder on the singers because one bad note can ruin a long take.” Boylan found that the technique on the later records changed the way he and the mix engineer worked in the booth. “The individual approach is harder during the mix because you're getting the blend yourself in the control room,” he says. However they recorded it, once the track was completed, the singers would then double the track, resulting in the thick vocal sound that launched such hits as “Reminiscing,” “Lady,” “Lonesome Loser” and “Cool Change.”
Click on links below to download .WAV files to hear the New Camaldoli project, recorded by Kevin Becka and Glen O'Hara, in surround.
These four .WAV files (labeled accordingly) should be panned discretely to LF, RF, LS and RS speakers with the levels set to nominal for all four channels.
Finding the right mic is like finding the right lover, says New York-based producer/engineer/studio owner Steve Rosenthal. Why? Click here to find out!
Recording vocals on a DAW?Click here to find out how engineers Marc DeSisto, Sylvia Massy Shivy and Dave Ogilvie do it.