Tuning vocals is an art form that often entails a lot more than just inserting a pitch correction plug-in on a vocal track and hitting play. Depending on the intonation of the singer, the tuning process can be painstaking and require many parameter adjustments along the way.
In this excerpt from Episode 6 of “Greg Wells Start to Finish,” Greg shows us his workflow and discusses his strategies for setting Antares Auto-Tune. (Note that he’s using Auto-Tune 8, and the examples and graphics in this article use Auto-Tune Pro, the most recent version.)
GETTING IT JUST RIGHT
As Greg is working on tuning the vocal track, he sometimes listens to the voice soloed, but more often with the guitar and piano in. A good reason to leave at least one instrument on when you’re tuning is to provide a pitch reference. Not for Auto-Tune, which has its own internal tuning reference, but for your ear. You can make better tuning decisions if you have some harmonic context. It’s also always helpful to have a keyboard instrument within reach when you’re tuning—it could be a real or MIDI piano—that you can use to identify notes from the vocal or figure out the correct destination for a particular note.
Back to Greg, he says his essential philosophy on using Auto-Tune is that it shouldn’t be detectable by the listener unless you’re using it as an effect, such as on R&B vocals with “the Auto-Tune Effect.” If you set it subtly, you can tighten up the vocal intonation without anyone knowing that you did.
You can open Auto-Tune as a track insert or use it as an AudioSuite (rendered) effect. Greg does the latter in this video.
You’ll notice in the excerpt that Greg’s workflow with Auto-Tune in Pro Tools entails using it as an AudioSuite effect. He works on one phrase at a time and renders his changes to the audio file as he goes along. For those unfamiliar with Pro Tools, AudioSuite effects are not inserted on a channel but are applied destructively to selected audio regions. In Logic, the process is called “Selection-Based Processing.” Most DAWs offer something similar.
It’s always important to make a backup of the track or the project when you’re doing destructive editing. You want to have the option of going back to the unprocessed copy, if need be.
SET IT GOOD
As he’s getting his settings figured out, Greg decides that the default Retune Speed on Auto-Tune is too fast for his taste, and is making the effect too noticeable. The Retune Speed parameter, expressed in milliseconds, determines how quickly Auto-Tune corrects a note that’s out of tune. If you set the speed to its fastest setting, 0, you get the “Auto-Tune Effect,” the one used on so many R&B and pop songs, which makes the vocal glide from note to note in an intentionally obvious way.
EX 1a: Here’s a vocal line with no Auto-Tune processing.
EX 1b: Here’s the same vocal with Auto-Tune Pro’s Retune Speed set to full to create the Auto-Tune Effect.
Auto-Tune Pro setting for Example 1b
In the case of Bryce’s vocal, Greg decides to set the Retune Speed to a setting of 109, for the first phrase. That isn't particularly fast, and certainly won’t produce the Auto-Tune Effect. He jokes that setting it this way is like giving the singer, “One glass of red wine.”
Another parameter Greg adjusts is called Tracking. It governs how picky Auto-Tune will be in determining which parts of the incoming signal to tune. For the first line of the vocal that Greg is working on, he sets the Tracking to be more Choosy than Relaxed, which means it will have a higher threshold for determining that a note needs to be tuned. Referring to his Tracking setting, he says that’s like a “half a glass of red wine.”
Between the Retune Speed and Tracking settings, he’s set the plug-in to be fairly restrained in which notes it tunes and how fast it tunes them.
MAJOR DECISIONS
With Auto-Tune, you also need to set the key and scale of the vocal you’re tuning. The song is in F Major and that’s what he sets it for. Auto-Tune contains a pull-down menu with a vast number of scales, but in most cases, you’ll probably use either minor, major or chromatic. (Auto-Tune Pro also lets you edit scales, and you can even let it automatically detect the key and scale when you use it in tandem with the helper plug-in, Auto-Key.)
Setting the correct key is vital, as it lets Auto-Tune know the correct pitches to move the notes to. If you have the wrong key or scale set, the plug-in may pull the notes to the wrong destinations.
Greg says that he sometimes will just use the Chromatic setting, which tells Auto-Tune to pull each note to the nearest semitone. That works fine in most cases.
Auto-Tune Pro’s rather lengthy scale list.
For example, let’s say the singer hits an A that is sharp enough that it’s closer to the Bb above it than to the A. In that case, Auto-Tune, if set to chromatic mode, will assume the note is supposed to be a Bb and tune it accordingly. Here, Greg sets the key to F major.
He then demonstrates how he corrected the first word of the opening line of “Lucky Number, which starts with the word “House.” The first note is an A, but Bryce’s pitch wavers very slightly and hits Bb near the end of the word. Because Bb is in the F major scale, and the scale parameter is set to F major, it’s not going to know to correct that Bb to A.
Luckily, Auto-Tune has a workaround for that, the ability to exclude notes from a scale. Greg uses that feature to exclude the Bb. Now when Auto-Tune sees that note, it will correct it to A.
When he previews that correction, the word “House” sounds stronger because Bryce’s voice doesn’t waver. He then tries it with a slightly faster Retune Speed. He likes that even more, and he renders the file.
The pitch in Bryce’s vocal was good to start with, but what Greg is doing is tweaking it a little here and there to focus it a bit more. He makes the analogy to the way you can set a camera’s aperture so that the background is blurry and the subject is the only thing in focus.
BETTER LIFE THROUGH TUNING
Sometimes, you can insert a pitch-correction plug-in on a vocal track, set it once, and get the results you want. That’s certainly a faster way to work, but isn’t usually as comprehensive. If you really want to have the fullest possible control over the vocal track, it’s better to go a phrase at a time like Greg does in the excerpt.
Graph Mode offers a more “hands-on” workflow than Auto Mode.
Auto-Tune also offers Graph Mode, which entails a very different workflow. It provides a graphical representation of the audio and lets you draw in pitch changes, thus giving you more control over individual notes than you get in Automatic Mode. (This works similarly, although not identically to Celemony Melodyne, another of the top pitch correction plug-ins).
Ex 2a: This vocal needs some tuning. The first time she sings “thing,” it’s flat, the second time she sings “I” it’s sharp. The second time she sings, “thing,” she scoops into the note a little too much.
Ex 2b: Here’s the same vocal, but this time corrected in Auto-Tune Pro using Graph Mode.
The setting from Example 2b. The red is the original pitch and the blue and green lines show the corrections, which manually entered.