All Articles Audio Software
From Auto-Tune to Flex Pitch: The Highs & Lows of Pitch Correction Plug-Ins in the Modern Studio
Joe Albano on Sun, September 6th 1 comments
"Auto-Tune" is ubiquitous with pitch corrected pop vocals. But pitch correction processors can be used both creatively and to correct performance errors. Joe Albano delves deep into the main players.

Ever since Auto-Tune first worked its way into the public consciousness in 1998 (on the Cher song “Do You Believe”), automatic Pitch Correction and Pitch Processing has been an integral part of modern music production. It’s primarily applied to vocal recordings, but of course it can be used on any instrument (guitar, bass) or audio signal, as long as it’s a monophonic melody (with one major exception—see below).

Pitch Processing has evolved from its humble beginnings—simply auto-correcting the intonation of slightly off-pitch notes—to incorporate much more powerful pitch editing features, including the capability to creatively edit individual notes within audio recordings in both pitch and time. And, as we all know, even the limitations of the technology have been adapted for creative use (or over-use) as a special effect. 

In fact, Pitch Processing is so ubiquitous nowadays that there are dozens of programs and plug-ins that do it, and while they all handle the most basic (auto-correction) functions easily, when it comes to the more advanced processing tasks, they bring slightly different layouts and features to the table. And a few of the most advanced processors offer their own unique capabilities. 

Fig1. A few Pitch Processors (clockwise from top): Melodyne, Auto-Tune, Revoice Pro; Logic’s Flex Pitch; Cubase’s VariAudio.

In this article, I’ll take a look at some modern Pitch Processors—both DAW built-ins and third-party plug-ins & standalone applications. I’ll start off by going through the features that (most) all of them have in common, then take a look at a few differences (in layout and implementation), and finish with some special capabilities of the most cutting-edge products. 

Common Ground

All modern Pitch Processors have the ability to auto-correct the intonation of out-of-tune notes. Auto-correction plug-ins implement this as a realtime, non-destructive operation. You simply insert the pitch correction plug-in on the audio track, make a couple of quick settings, and hit play. The grandfather of realtime auto-correction, was, of course, Auto-Tune, but it’s now been joined by many other plug-ins, including DAW built-ins like Logic’s Pitch Corrector plug-in.

As you can see below, the main controls are basically the same—there’s a tuning reference, and, most importantly, a Speed control (labelled “Response” in Logic). 

Basic (and more advanced scale) controls in Auto-Tune and Logic’s Pitch Corrector plug-in

Fig 2. Basic (and more advanced scale) controls in Auto-Tune and Logic’s Pitch Corrector plug-in.

The speed setting determines how quickly the processor reacts to the detection of an off-pitch note in the audio. This is the key to getting auto-correction to either do its job transparently, with (virtually) no noticeable side-effects, or to provide that well-known special effect. Speed settings from around 60-100 milliseconds preserve the natural imperceptible slides between notes that all singers do. If the Speed is set to its fastest setting, the pitch is fully quantized—the famous effect, something no singer is capable of. 

Most auto-correction processors also usually include options for (user-adjustable) scale settings (Fig 2), which allow the user to restrict pitch-correction to only certain notes in a musical scale. Auto-Tune has even more advanced features as well, including vibrato and formant controls (Fig 2).

Audio Example 1: A vocal line: the original; subtly auto-corrected, first with Logic’s Pitch Correction plug-in; then with Auto-Tune; then run through Auto-Tune with the Speed set to 0 (fastest) for the classic pitch-quantized Auto-Tune (“Cher”) effect:



Off The Grid

Non-realtime (offline) graphic pitch editing capability is included with some auto-correction plug-ins (Auto-Tune features both modes), and many DAWs have both realtime (plug-in) and offline (graphic) features. Graphic pitch editors are, by now, pretty familiar—they all resemble a MIDI piano roll editor, with the individual detected notes represented as bars (or “blobs”, in Melodyne), which can be dragged around freely by hand. These graphic editors also usually incorporate an auto-correct mode, but their main purpose is more specific tweaks. This is where the application of pitch processing shifts from simple correction to more creative uses. Besides the Auto-Tune plug-in’s Graphic mode, graphic Pitch Processing can be found in most DAWs these days—Logic’s Flex Pitch and Cubase’s VariAudio are perfect examples of this. And standalone applications (which can often be sync’d to the DAW) also offer it, like Synchro Arts’ Revoice Pro, which has some unique capabilities, and, most notably, Celemony’s Melodyne (Editor), which, even offers polyphonic pitch processing (see below), and is available in both plug-in and standalone form.

Fig 3. Graphic Pitch Processors (clockwise from top left): Auto-Tune’s Graphic mode, Logic’s Flex Pitch, Cubase’s VariAudio, Melodyne, Revoice Pro’s Warp Process.

Fig 3. Graphic Pitch Processors (clockwise from top left): Auto-Tune’s Graphic mode, Logic’s Flex Pitch, Cubase’s VariAudio, Melodyne, Revoice Pro’s Warp Process.

In all cases, the audio must first be analyzed, and the individual pitches detected, before editing can begin. Once this is done, the note bars can be dragged freely, or snapped to discrete pitches (and of course, in all these graphic displays, timing can also be simultaneously tweaked, though here I’m just focussing on pitch). But besides manual note adjustments, graphic editors provide the ability to tweak more than just the basic pitch of the notes. Curves within each note block display pitch variations within and between notes—(unsteady) pitch drifting, sliding in or out of notes (or from one note to another), and even vibrato (see Fig 4). These curves can also be dragged, again, for either corrective or creative purposes, or to compensate for artifacts that might result from changing pitches. Most editors even feature the ability to alter the formants of the the notes (formants are fixed resonances that give each singer his/her unique vocal tone).

Additional pitch-related parameters that can be edited in graphic displays (shown in Flex Pitch).

Fig 4. Additional pitch-related parameters that can be edited in graphic displays (shown in Flex Pitch).

Graphic pitch processing allows for customized creative edits, like the manual creation of particular harmonies, from a copy of the original melody. And since the pitches are already detected (to make the processing possible), most graphic pitch editors also include an Audio-to-MIDI conversion feature, which generates a string of MIDI notes from the audiofile, saved as a Standard MIDI File, which can then be imported back into the DAW for creative applications, like doubling a vocal line with an instrument, or replacing a less-than-steller-sounding instrument recording with a MIDI-triggered sample (Audio Ex. 2).

Audio Example 2: A bass audio track converted to MIDI, triggering a sampled bass (original audio, then MIDI conversion):





Differences and Special Features


Now, so far, I’ve been describing features that are common to Pitch Processors—so what are some differences among them? Well, I’d say differences fall into two main areas—sound quality, and special features. 

As for sound quality, most realtime auto-correction plug-ins get the job done with little significant audible difference when it comes to the most basic corrective tasks. However, they may require different settings for the best results. For example, I’ve found that Auto-Tune does its best work with Speed settings around 60 milliseconds or less, while Logic’s Pitch Correction plug-in may sound a little effect-y at that speed, typically achieving its best results with speed (Response) settings of 100 milliseconds or so.

With the graphic editors, the biggest sound-quality difference is usually how far they can pitch-shift a note away from its original pitch before it begins to sound unacceptably artificial/processed (munchkinization). Typically, with most of the processors I’ve used, this is about three or four semitones at most, and then at least some hint of artifact begins to creep in. In the past, I felt Melodyne was the clear winner here, but lately, it seems that all the Pitch Editors are quietly improving their algorithms bit-by-bit, version-by-version. Since pitch-shifting quality varies considerably depending on the particular voice/recording, it can be handy to have a couple of Pitch Processors available—one’s algorithms may simply be a better match for a particular audio file than another.

Audio Example 3: The same harmony line created by pitch-shifting with Logic’s Flex Pitch, Melodyne, and Revoice Pro:

(1) Original audio; (2) Logic harmony line; together; (3) Melodyne harmony; together; (4) Revoice harmony; together


To Each His Own


While the layout and feature set of most graphic pitch editors is pretty similar, some present a little differently. Pro Tools’ Elastic Pitch is a considerably simpler affair than Logic’s or Cubase’s—it does not offer graphic editing capability (as opposed to their Elastic Audio (time-shifting) feature, which does, and is as full-featured as any). Elastic Pitch is an offline pitch-transposition that can be applied to individual regions (via the Elastic Properties window), but it doesn’t offer auto-correction, a graphic mode, or any of the extra pitch tweaks I mentioned above. Most critically, it doesn’t have any Formant compensation, so it’s really best suitable for small pitch corrections and transpositions of no more than a semitone or so.

Logic’s Flex Pitch and Cubase’s VariAudio both have the full set of graphic editing features I described, as does Melodyne, and the Warp Process feature in Revoice Pro, though there are slight differences in implementation here & there. With Audio-to-MIDI conversion, both Cubase and Melodyne convert not only the notes, but velocities as well (from the level of the audio wave for each note), while Logic currently does not, setting all the converted MIDI notes to a fixed velocity. While all the editors have the same note(block)-dragging capabilities, the way the extra performance gestures (slides, vibrato) are addressed can vary slightly. For example, Logic and Cubase feature draggable curves on top of the note blocks, while Melodyne employs different tools for these various editing functions (Fig 7).

Editing features in Logic, Cubase, and Melodyne

Fig 5. Editing features in Logic, Cubase, and Melodyne.




Special Sauce


Two Pitch Processors that have particular and unique extra features are Revoice Pro and Melodyne.

Revoice Pro is a specialized application—its primary purpose is to synchronize dubbed dialog to location recordings—that is, to match the timing of one take precisely to another. But it also handles pitch, and can precisely match the pitch, spoken or sung, of one take to that of another with a single stroke, a unique capability that can come in very handy. Revoice Pro also has a full-fledged graphic time & pitch editor (Warp Process) like the others here, which provides note blocks and pitch curves that can be dragged for more specific editing. The latest version, 3.1, adds a new pitch editing feature for tweaking pitch curves in the Warp display—the Pencil Tool. Most pitch editors let you grab a pitch curve and drag to shift it, or to compress/expand the degree of pitch variation—Revoice has this too. But the new Pencil Tool lets you draw your own new curve from scratch over the original (performed) one—this can be a powerful and convenient feature, on those occasions when it’s needed.

Revoice Pro 3.1’s Warp Process has a new Pencil Tool for pitch curve editing

Fig 6. Revoice Pro 3.1’s Warp Process has a new Pencil Tool for pitch curve editing.


Polyphonic Power


Probably the most impressive pitch-processing feature is Melodyne’s DNA—Direct Note Access. While all the other pitch processors restrict their magic to monophonic (one note at a time) melodies, Melodyne’s DNA is alone in its ability to process polyphonic material, with full control over simultaneous notes within chords. DNA has been around for a while now, but this capability is still unique to Melodyne. This means you could go into, say, a polyphonic recording of acoustic guitar chords, and manually fix or alter individual notes within the chords—you could even perform scale-based edits, like turning major chords into minor, for example, and there are presets for doing just that. 

Melodyne’s DNA—Direct Note Access—polyphonic pitch processing feature in action (the red “blobs”)

Fig 7. Melodyne’s DNA—Direct Note Access—polyphonic pitch processing feature in action (the red “blobs”)

Audio Example 4: Major-key acoustic guitar chords pitch-corrected and changed to a minor key with Melodyne’s DNA:

The DNA feature represents the current cutting edge of pitch processing technology—it’s only limitation is that it can’t process polyphonic audio from different instruments that are already mixed together. But who knows—there are already software packages on the horizon that promise to be able to do that, so we’ll just have to wait and see. 

But for now, the various tools that are currently available provide a wealth of editing possibilities when it comes to tweaking pitch in our audio recordings. Whichever DAW, plug-in, or application you choose, you’ll have no shortage of features for everything from quick & efficient realtime pitch correction to in-depth manual fine-tuning and creative tweakery. Most of the editors I’ve mentioned here have demo versions, so if you need to get up to speed on the current state-of-the-art in pitch processing, you should be able to find just the tool to suit your needs.


Watch Video Courses on Melodyne HERE.

Watch Video Course on ReVoice Pro 3 HERE.

Watch Video Course on Logic Pro X's FLex Pitch HERE.

Watch Video Course on Cubase HERE.


Related Videos
Comments (1)

You must be logged in to comment.

  • lucas pottersky
    really cool article! thanks for the great summary on current state of the art pitch correction tools! =) Now, I didnt exactly understand Melodyne's DNA. On their website, they say the basic (essential) version has: "Universal – Complex, polyphonic material without DNA Direct Note Access™". So, looks like DNA isn't just about being able to detect Polyphony?
    • 4 years ago
    • By: lucas pottersky
Advanced Melodyne
Melodyne 201
Dream It. Do It.
Do you want to learn Advanced Melodyne?
Yes, I want to learn!
No Thanks, I just want to read the article.
Course Advisor
Don't Know Where To Start?
Ask A Course Advisor
Ask Us!
Copy the link below and paste it into an email, forum, or Facebook to share this with your friends.
Make money when you share our links
Become a Affiliate!
The current affiliate rate is: 50%
Classes Start Next Week!
Live 8-week Online Certification Classes for: