If smooth, realistic pitch correction is your goal, here we’ll give you the techniques to achieve transparent vocals that don’t sound like they’ve undergone any processing at all.
Modern music is swimming with pitch shifted vocal lines and out-of-this world, robotic hooks. Ever since Cher’s “Believe” introduced the world to this vocal effect en masse, and after the effect was repopularized throughout T-Pain’s catalog, it’s easy to forget the original intention behind the technology that made those sounds possible.
Vocal tuners like Waves Tune and Waves Tune Real-Time were originally intended to produce a pure, seamless and unnoticeable vocal repitching effect. It’s that effect that we’ll show you how to achieve in this article.
If instead you’re after robotic vocal effects that push the envelope, here’s a guide to How to set up Creative Vocal Tuning in Real Time, and remember to check out a free trial of our more recent plugin OVox, for a whole toolkit of instant creative vocal effects!
Why is Natural Pitch Correction Important?
As explained above, there’s a whole spectrum of pitch correction; from subtle tweaks to unrecognizable transformations. But why might we want to achieve a subtle pitch shifting that is likely to remain undetected to the average listener? More established styles of music, such as soul or rock, have evolved without the technology that is required to alter the pitch of vocals, so applying the technique in a brash manner would stick out like a sore thumb.
At the same time, as a species, us humans enjoy sounds that are in harmony with one another. Even non-musicians are able to detect melodic dissonance, and we tend not to find the sound too pleasing. This means that from time to time, we may need to take matters into our own hands and correct the odd off-note with the help of some choice production tools. Other times, you may need to shift the entire recording up or down in pitch in order to work with a new key.
How to Achieve Natural Vocal Pitch Correction
So, how can we avoid robotic, glitchy vocal pitch correction, and do it more naturally?
Start with a Good Recording
You may have heard people say about the production process, that you can’t simply “fix it in the mix”. While that is true to some extent, there are of course mixing tricks that can help to disguise or even completely fix recording or production issues. The same concept applies to applying pitch correction to vocals. In an ideal world, you wouldn’t have to apply any correction to your recordings, but as you’re reading this article, we’ll assume that’s not the case. At the very least, you should be aiming to work with vocals that need minimal pitch correction, such as this vocal take.
If something needs more extreme pitch repairs, you might get better results by going back to the vocal booth. Similarly, you should also be working with clean recordings that are free of any background noise, plosives and breaths. While this is a general rule, it’s particularly important when applying pitch correction, as your plugin may get thrown off by unwanted excess noise.
Let’s use Waves DeBreath to tackle the slight breathiness in our vocal recording:
Accurately Identify Your Target Key
We now need to identify which key our nice clean vocal recording is in so that we can correctly instruct our pitch correction plugin. You may already know the key of your vocal if the song’s been written already. If not, though, an efficient and reliable way to accomplish this is with a little help from Key Detector. Simply place Key Detector on your vocal channel and press play. The plugin will analyze the audio, and will quickly suggest its likely key. In this case, our vocal is in D minor.
Now, if we place Waves Tune Real-Time on the vocal channel and hit Transmit Key in Key Detector, Tune Real-Time will automatically apply the appropriate scale settings for our vocal.
Anchor Your Vocal’s Key
It can be difficult to detect iffy tuning when listening to a vocal in isolation. If you’re applying vocal tuning within a song’s entire project, remember to regularly listen to the vocal track alongside another melodic element. This gives you a reference point for you to confirm the pitch correction you’re applying is getting the vocal closer to its actual target. After all, the most important thing is how it sounds.
If you’re applying vocal repitching in a separate project, simply grab a piece of audio in the same key, or create some suitable MIDI. In this case, a simple arpeggiated piano in the scale of D Minor gives us a good reference point.
Adjust Correction Timing
Now Tune Real-Time is set up with the appropriate scale target, we can begin to adjust the parameters to achieve the desired effect. In its default state, the Note Transition value is set to 120ms, which triggers the pitch correction too slowly for this application due to some of the vocal delivery being relatively fast. For slower performances with lots of sustained and legato notes, this Note Transition time would be more suitable.
At the other end of the spectrum, if we adjust the Note Transition down to around 20ms we can hear audible pitch quantization as the plugin works to correct the pitch too quickly.
Increasing the Note Transition time to around 45ms gives us a happy medium, whereby the vocal’s pitch is corrected, but not so quickly that it sounds unnatural.
We can now repeat this step with the Speed control; reducing it until it sounds unnatural, then increasing it slightly.
As standard, Waves Tune Real-Time will correct any notes that deviate from the selected scale. On the whole, that’s what we’re trying to achieve here. But what about when the vocal performance contains some deliberate pitch movement that we want to retain in the recording?
Using Tune Real-Time’s Vibrato function, we can allow a certain amount of pitch modulation, whether that’s vibrato in the vocals, or glissando-style slides between notes. At 0%, the plugin will completely quantize the vocal to the chosen scale, so Vibrato mode is essentially off. At 100%, the plugin will let all pitch modulation through, and values above 100% emphasize the pitch modulation. In our example, a value of around 75% controls the pitch modulation slightly, but doesn’t lock it completely to the scale. The result is a pitch corrected but natural-sounding vocal.
‘Mult’ Your Vocal Recording
As you can probably hear, there are a couple of sections towards the end of the recording that still sound a little out of tune. Our options are to apply more extreme pitch correction to the entire recording, which then runs the risk of making other sections of the performance sound unnatural. Alternatively, we could automate some of the parameters to work harder at the end of the recording, but with many parameters to adjust, this might be a convoluted way to tackle this. Finally, we could split or ‘mult’ the recording onto multiple channels, each with their own instance of Tune Real-Time set up with their own dedicated settings.
To do this, we’ve split the vocal at a zero point in the waveform, and moved the particularly troublesome section onto its own channel, copying the first instance of Tune Real-Time and placing it on the new channel. We can now open up the second instance of the plugin and adjust accordingly. In this case, reducing the Speed and Note Transition times, thus making it faster, gets us closer to the level of pitch correction that we need. We’ve also reduced the Vibrato amount so that any movement in pitch is a little more controlled.
Fine-Tune with Cents and Time Tolerance
While the faster Speed and Note Transition values have improved the pitch correction of the second section of our recording, it has introduced some slight quantization artifacts which are preventing the vocal from sounding as natural as it can. To counter this, Tune Real-Time features two Tolerance controls; one for Cents, and one for Time.
When a target scale has been selected within the plugin, by default, it will correct the signal’s pitch when it detects audio that is at least 50 cents out of tune in either direction. We can increase this amount by adjusting the Cents in the Tolerance section. This has the effect of increasing the pitch threshold at which the pitch correction is triggered, thus softening or taming the pitch correction itself. By adjusting the additional Cents Tolerance to around 15ms, we can successfully control some of the unwanted pitch quantization that was audible before.
Not every vocal you ever work with is going to require pitch correction, and as mentioned, if a vocal needs to be completely transformed, it’s probably worth recording it again. For scenarios where some correction is required, the techniques outlined in this article will help you to do so in a subtle and natural way. If you want to see for yourself how easy it is to correct vocal pitch using Waves Tune Real-Time, you can find out more and download the free trial here.