How to Autotune Your Voice — Natural Vocals and Full Songs
You have finished recording a cover. The timing and emotion feel right, but a sustained note in the chorus is a little flat, and the end of another phrase drifts sharp. Pitch correction tools such as Auto-Tune can help you adjust a recording like this. You can then compare it with the original and decide whether to keep the changes.
The result depends on how much of the original performance you want to preserve. Lighter correction can bring unstable pitches closer to the intended notes. Faster, stronger correction makes the transitions between notes more noticeable, creating an electronic vocal effect.
If you already have a recording, the <a href="/autotune-online">Lacuna online autotune tool</a> accepts both isolated vocals and full songs with accompaniment. This guide starts with choosing your audio, then walks through processing it, listening to the result, and adjusting the settings.
Choose your audio: isolated vocals or a full song
Pitch correction needs to identify the notes being sung. A separate vocal track is easier to process. When the voice is already mixed with drums, bass, and chords, it needs to be separated first.
When you open the tool, choose the input type that matches your file:
- Isolated vocals: Choose dry vocal mode. This is suitable for a vocal recording without accompaniment or a vocal track you have already exported.
- A song with accompaniment: Choose full song mode. Lacuna separates the vocals and accompaniment, corrects the vocal pitch, and mixes the vocals back with the accompaniment for the output.
For example, if the only file on your phone is a cover already mixed with its backing track, you can choose full song mode without separating the tracks yourself. If you have both the isolated vocals and the complete song, start with the isolated vocals to skip the separation step.
The source audio also affects the result in full song mode. Several people singing at once, dense accompaniment, or heavy reverb can leave other sounds in the separated vocals. After correction, listen to them together with the accompaniment.
Decide between natural correction and an electronic effect
Natural pitch correction aims to adjust notes that are out of tune while preserving pitch slides, vibrato, and changes at the ends of phrases. A pronounced electronic effect moves the pitch to its target note more quickly, making that transition part of the sound.
Pitch correction can achieve either result. The Antares <a href="https://www.antarestech.com/blog/getting-started-with-auto-tune-and-soundtrap" target="_blank" rel="noopener noreferrer">official getting-started guide</a> also covers both natural correction and more obvious effects, and recommends adjusting the correction speed to suit the recording.
In Lacuna, choose a starting point based on the result you want:
| Desired result | Suggested starting preset | What to listen for |
|---|---|---|
| The original is already close; only a few unstable notes need adjusting | Subtle polish | Whether the original pitch slides and phrase endings remain |
| Steadier vocals that retain the movement of the performance | Natural | Whether sustained notes are steadier and transitions sound smooth |
| More noticeable correction that brings the voice closer to the notes | Pop vocal | Whether fast phrases remain clear and the amount of correction feels right |
| An obvious electronic vocal sound with abrupt pitch transitions | A hard-tuning preset such as T-Pain | Whether the note changes create the rhythm and texture you want |
For your first pass on a cover or an a cappella recording, try Natural or Subtle polish. A preset is a starting point. The version you keep should depend on how it sounds in the song.
The key determines which notes your voice can be corrected to
After choosing a preset, check the song's key. This determines which notes the tool can correct the voice to. C major and C minor, for example, do not contain exactly the same notes.
If you know the key, select the corresponding root note and scale. If you do not, start with automatic detection, then check whether the corrected melody sounds as expected. If the backing track has been transposed up or down, use its current key.
The correct key does not mean the tool knows which note you intended to sing in every phrase. A scale contains several possible notes. A note sung far enough off pitch can still be pulled toward a different note within that scale. If a phrase starts to sound strange, check the key first, then consider reducing the correction strength.
There is no need to try every scale just because more options are available. For a first result, confirming the song's current key makes problems easier to assess than changing many settings at once.
For natural results, listen closely to these three settings
Presets already combine the settings for you. If you hear something that needs adjusting, start with these three controls.
Correction strength determines how far the original pitch is moved toward the target pitch. Higher strength applies more correction. If the original needs only a little polish, reduce the strength to preserve more of its subtle variations.
Correction speed determines how quickly the pitch reaches the target note. The interface uses milliseconds: the lower the value, the faster the correction. When correction is too fast, a slide into the next note can become an abrupt jump. The Antares <a href="https://help.antarestech.com/hc/en-us/articles/41076455881364-How-can-I-get-that-classic-Auto-Tune-effect" target="_blank" rel="noopener noreferrer">guide to the classic Auto-Tune effect</a> also identifies very fast correction as a key setting for a pronounced electronic sound.
Vibrato preservation determines how much of the singer's vibrato is retained. Regular pitch movement in a sustained note can be part of the performance. For natural correction, listen for whether that movement has been flattened too much.
Three of Lacuna's current presets show how these settings work together:
| Preset | Correction strength | Correction speed | Vibrato preservation |
|---|---|---|---|
| Subtle polish | 45% | 40 ms | 100% |
| Natural | 70% | 25 ms | 90% |
| Pop vocal | 85% | 15 ms | 60% |
These are the tool's preset values. They illustrate the progression from subtle polish to more noticeable correction, rather than a fixed answer for every recording. When you need to make a change, adjust one setting at a time and listen to the same phrase again.
Compare the same phrase to hear whether correction helps
Once you have chosen the input type, preset, and key, select an output format and start processing. When it finishes, you can preview the original and corrected versions in the results and switch between them at the same playback position.
For the comparison, listen to these three parts in turn:
- The phrase that was least in tune: Have the sharp or flat notes improved? Has any part of the melody been pulled to a different note?
- Pitch slides and sustained phrase endings: Does the voice move smoothly from one note to the next? Are the vibrato and expression at the ends of phrases still there?
- A passage in the context of the full song: Do the vocals fit the accompaniment? If a full song was separated and mixed back together, has any obvious noise or unnatural change appeared?
The pitch curve can help you see where adjustments occurred, but listening should guide the final decision. If a section looks straighter yet has lost details you wanted to keep, reduce the strength, slow down the correction, or switch to Subtle polish.
The right version also depends on the music. A ballad cover may benefit from keeping pitch slides and breathiness. For electronic vocals, the obvious jumps between notes may be exactly the effect you want.
A few practical questions before you start
Can I use audio recorded on my phone?
You can upload MP3, WAV, FLAC, AAC/M4A, or OGG files, or record with your microphone on the page. Uploaded source files currently support a maximum size of 50 MB and a maximum duration of 10 minutes. Pitch correction processes the recording after you submit it; the recording feature does not let you hear corrected vocals while you sing.
Do I need to pay for processing?
You need to sign in to upload and process audio. Uploading does not use credits; credits are deducted when you start processing. Each processing run currently costs 15 credits for isolated vocals or 30 credits for a full song. Free-tier accounts can use MP3 320 kbps output. WAV 16-bit and 24-bit output require BASIC or above. WAV output also costs twice the credits. Refer to the page for the current details.
Which recording problems can pitch correction fix?
Its main job is to adjust pitch. Singing too fast or too slowly, or not pronouncing a word clearly, requires separate attention. Those problems do not automatically disappear when the pitch is corrected. Identifying the problem while listening makes it easier to decide what to change next.
Start with a recording you know well. Choose Natural or Subtle polish in the <a href="/autotune-online">Lacuna online autotune tool</a>, process it, and compare the same phrase with the original. First decide whether the adjustment helped, then consider whether you need stronger correction.