Historically, vocals have long been the most difficult element to manipulate on a recording. For keyboardists, it’s a snap to record a part via MIDI, and then change the patch the instrument is playing. But until recently, all a producer could do with a vocalist was to record lots of takes and knit them together, and then add a one or two effects (out of only a handful to choose from) onto that track.
But the technology available even to home recordists on very modest budgets gives the voice a rapidly increasing amount of flexibility. It hasn’t quite reached the “dial in another patch” plasticity of keyboardists yet, but it’s getting awfully close. So why do so many home, amateur and semi-pro recordings still have such poor lead vocals?
One reason is that they’re not fully taking advantage of the technology that’s available to them. The other is that they may not have thought about how to use that technology creatively. And that’s what this article is all about. It isn’t really a “how to record vocals” piece. We won’t be discussing how to choose a microphone or where to place it, but fortunately, there are plenty of articles and books available on that subject. What we’ll discuss is how to take full advantage of the many features that computer-based recording programs bring to even the most budget-conscious recording efforts.
With PC-based recording, the number of tracks is typically determined by the amount of RAM and hard disk space the computer has, and for many projects, running out tracks isn’t even a concern. It’s easy to record separate takes of each vocal part, often singing the entire song, or big chunks of the song continuously through. For example, Cakewalk’s Sonar can be set to cycle record, by first engaging the “Loop/Auto Shuttle” function, followed by engaging the “Store Takes In Separate Tracks” button, located under “Record Options”. Both of these windows are located under the “Transport” tab. Once both of these commands have been set, then unlimited tracks can be recorded. When the program reaches the end of a take, it cycles back to the beginning of where it should start, and automatically opens up a fresh track to record on.
Once those takes are recorded, it’s time to “comp” the vocal, which means going through the multiple takes of vocals on separate tracks and editing them down to one or two tracks of the final lead vocal. And hopefully that track will sound, at least to the average listener, like one smooth, continuous piece of singing, without punch-ins or other obvious clean-ups. This may take a fair amount of sliding material around, taking parts from one chorus to another, or doing whatever’s necessary for the vocal to both sound in tune, have the right tone, and, if the song calls for it, build in intensity as it goes on.
Incidentally, you’ve probably heard this before, but it’s always a good idea to record the vocal–even a rough guide vocal–as early possible when recording a song, even it’s purely a demo. If you’re a one-man-band home recording operation, you’ll find that your recordings will sound much more natural if your instruments have at least some interplay with the vocal, rather than trying to add a vocal as the final layer of an otherwise completed recording.
Adding The Ear Candy
Once all those takes are recorded, and comped down to one or two main tracks of vocals, and as the song gets closer to being mixed down, it’s time to add “the ear candy”. Reverb, echo, electronic vibrato and double tracking are some of the most basic vocal effects, and they date back to at least the days of the Beatles, and have lasted all these years because they work.
Years ago, George Martin observed that the weaker a voice is, the more double tracking–even triple tracking–can strengthen it. The first method is simply repeating a vocal by singing along with what you’ve already sung. The fact that you’ll never quite clone your previous vocal is both the drawback and the plus of this method. Some people swear by it, and others swear at it.
You can also try doubling the voice by singing the second track in a different tone than the first vocal. Lots of Rolling Stones songs get their character from Keith Richards’ almost yodel-like vocal timbre doubling Mick’s voice, but mixed fairly low behind it to add an unusual color or texture.
Using effects to double a vocal can work well, particularly for those vocalists who don’t have the voice, training or patience to accurately double track their own voice.
A different kind of doubling is obtained with a chorus effect, which is available as a plug-in module in numerous digital music programs. Chorus effects have long been popular on electric guitars, as they allow a six-string instrument to simulate the jangly sound of a 12-string. But they can also work well on vocals, simultaneously thickening a voice, and masking minor pitch irregularities. It’s possible to use the editing tool almost like a compressor, and seamlessly cut the beginning and end of a loud note or two, then use Sonar’s “Lower By 3db” function under the “Process” function on that section. Or raise or lower the gain of individual clips via Sonar’s envelope function. The results should be undetectable.
While it can be used an obvious effect (pioneered on Cher’s hit single, “Believe”), the Antares Auto Tune or its competitor, RBC Audio’s Voice Tweaker Pro plug-in (which has some pretty nifty patches beyond pitch-correction), can be great for correcting one or two words that are out of pitch in line, or even raising or lowering a phrase to add more melody to a vocal that’s too one-dimensional.
For correcting intonation problems in singers, plug-ins such as Auto-Tune are a useful addition to a digital arsenal. However, as Tom Brier of Cakewalk says, “if you didn’t want to spend the money for one of those plug-ins, you could use the Groove Clips feature in Sonar to actually fix pitches. So you could make a Groove Clip [Cakewalk's version of Sonic Foundry's Acid Loops] out of one phrase, or one note for that matter, and then tune it,” by raising or lowering it the appropriate degree. Simply click on Sonar’s scissors icon to edit, then cut before and after the offending the word, and then right click on the section that needs adjusting, and then click on “Stretch to Project Tempo”, and then adjust the pitch up or down a semitone or two.
Finally, while it was designed originally as a mastering program, Izotope’s Ozone plug-in has several patches that can sound great on lead vocals. They take a fair amount of processing horsepower, and may require some sliding around of the vocal’s time to compensate for the microscopic delay they induce, but they’re can an impressive “suite” of multiple effects to a single track.
Creating Lush Harmonies
It’s long been known that lots and lots of harmony vocals, carefully mixed down to a stereo track, can sound wonderfully lush. With hard disk-based recording systems, and the right singers, it’s easy to build up a large bank of background vocal tracks, and then mix them down for space purposes, if necessary.
Of course, that may not be enough. Mike Talanca says, “If we find that we’ve already tripled the background vocals, and it still isn’t big and lush enough, even after you tripled them, then we know it’s not in the vocals. Now we’ve got to add some instrumentation to make it wide, or thicken the part more.” In other words, it’s time to call in the synthesizers.
In the old days, synthesizer patches designed to simulate a choir or chorus of vocals sounded too obviously like…a synthesizer. In contrast, today’s chorus patches (not to be confused with the chorus effects we discussed above) can sound amazingly realistic. For a song that needs a lush background of “oohs” and/or “ahhs” behind a lead vocal, they may be all that’s necessary.
Call In Max Headroom!
One drawback to sampled choruses is that they can sound increasingly mechanical or repetitive each time a particular part is replayed. But it’s easy to add a human element to them by mixing in some real voices, either your own, or by using loops or SoundFonts. (Or both!)
Sony’s series of loop CDs for Acid (which are also playable in recording programs from other manufacturers) include three discs with vocals on them. These titles are the two disc Ilona/Universal Female Vocal Toolkit, and the Hip-Hop/R&B Vocals CD-ROM, as well as World Pop, which includes a few exotic-sounding vocals amongst a variety of world instruments.
The Ilona and Hip-Hop R&B Vocal CDs offer a blend of “oohs and ahhs” for choruses of background vocals, as well as a variety of scat singing and stock vocal phrases (such as “I love you”, “I need you”). For a musician working alone, these CDs can do much to open up a song, and offer a world of possibilities: choruses of background vocals, Jimmy Page/Robert Plant-style instrument/vocal call and responses, James Brown-style shouts at the climax of instrumental solos, and even whole verses sung by a different singer. (Assuming you can work the stock phrases into a meaningful sounding verse). Tracks along the lines of Pink Floyd’s “The Great Gig In the Sky”, which features verses of nothing but scat-singing by female vocalists over an instrumental bed, become child’s play–just a matter of lining up the loops.
After using them for a while, it’s easy to think of the looped vocals as just another vocalist. It’s possible to modify the melodies of the Acid Loop vocalists through a combination of careful editing and pitch changing. The resulting phrase could be used before or after its stock version, to create a vocal phrase and its response.
Users of SoundFonts aren’t left out in the cold either, when it comes to vocals. Sonic Implants’ Female Jazz Vocals patch is based around the scat and harmony singing of Patrice Williamson, an award-winning jazz vocalist. Sonic Implants recorded her singing sustained tones, as well as some nifty sounding scat phrases, and choral pads. These samples can easily be manipulated via a MIDI keyboard and/or a sequencer such as Propellerhead’s Reason.
The ultimate extention of these types of sampled voices is Yamaha’s Vocaloid program. At the moment, it’s chief drawback is that it’s relatively complicated to program. As I wrote back in March for Tech Central Station:
In 2003, Yamaha created a product that they call “Vocaloid”. Its programming can be input by any musician; its basic graphical user interface is much like any software-based synthesizer. But after that, words are typed onto the screen. And then the user enters phonemes if the words aren’t in Vocaloid’s vocabulary. Phonemes are the phonetic unit of language that conveys distinction in meaning, such as the “mmm” of mat and the “buh” of bat. Finally, performance characteristics such as the tone of the singer’s tone, breathiness and dynamics are programmed in.
When all the programming is complete, press play, and it sings. No, not computer generated singing — but it sings with a human voice that has been digitally sampled, deconstructed, and re-assembled according to your input. And Vocaloid can do its own harmonies and back-up vocals.
Currently, Yamaha, through Zero-G, their UK distributor, markets Vocaloid with three different voices each available separately, two female, and one male. They’re named Leon, Lola, and Miriam, after the actual humans whose voices have been sampled to create the synthetic singers.
The initial programming typically does sound like a synthesized android’s voice; it takes a lot of programming to make Vocaloid sound like a real voice, and not a synthesizer.
But synthesized instruments have gotten increasingly realistic sounding over the past twenty years, and there’s no reason to believe that within a much shorter period of time, Vocaloid, and its inevitable competitors will eventually become indistinguishable to most listeners from the human voice.
There’s no reason why the vocal loops, SoundFonts and synthesized voices such Vocaloid can’t be run through flangers, phasers, or even the Antares Auto-Tune, to give them their own effects. (Incidentally, you may find that unless the Acid loop is first copied to a new audio loop, or its loop feature disabled, it won’t track properly through the Auto-Tune patch.)
Of course, you can always “roll your own” in addition to using pre-sampled sources. It can be helpful to record a variety of grunts, screams, yelps, and other lines of scat singing to be able to drop into songs during solos, drum breaks and other key parts. If you employ a friend for your vocals or backup vocals have him or her go crazy singing–and screaming–some of this material. You’ll never know when a nice scream fits the bill perfectly after a drum break, but before a verse starts.
The Ronco Voice-A-Matic
In 2002, Line6 debuted their GuitarPort product, an interface through which an electric guitarist can plug his trusty Strat or Les Paul into the USB port of his computer. As part of its software, it allows guitarists to download patches that are often excellent recreations of the tones used by Jeff Beck, Eric Clapton, Jimmy Page, and other guitar heroes.
I have no doubt that within a few years, a similar product will be available for budding vocalists. Who do you want to sound like? Al Green? Paul Rodgers? Barbra Streisand? Aretha Franklin? Simply dial in a patch and sing into the Ronco Voice-A-Matic!
In the meantime, TC-Helicon’s VoiceOne product may be the ultimate–or at least certainly the most technologically advanced–solution to getting great vocals. The patches built into the VoiceOne allow the user to edit and change not just pitch and timing, but alter the vibrato, breathiness, tonal characteristics, even the throat and neck sizes, and/or add a Louis Armstrong-like growl to a recorded voice. It can be controlled manually or in real-time, via MIDI. The VoiceOne can also correct the pitch a vocal whilst simultaneously altering its timbre, as well as pitch change a vocal. RBC Audio’s Voice Tweaker Pro
For the vocalist who can't truly sing, lessons and lots of practice are good first steps. But there are certainly plenty of tools--and lots of techniques, ranging from the sublime to the ridiculously advanced--to make an OK vocalist sound (as a famous baritone once said) “Impressive…most impressive.”