Skip to content

Correction Modes

OpenVoxTuner offers two distinct ways to correct pitch, selected with the Mode control:

Mode Target pitch source Best for
Auto (Live) Automatic quantization to the current Key/Scale Live singing, fast results
Graphic (Curve Editor) A pitch curve you draw with the mouse Offline mixing, note-perfect results

The rest of the processing pipeline (amount blend, retarget speed, formants, pitch shifting) is shared by both modes — only the target pitch comes from a different source.

Auto mode (scale-aware quantization)

In Auto mode, the plugin detects the pitch of the incoming signal and snaps it to the nearest note of the selected Key and Scale. This is the familiar real-time "autotune" behavior.

  1. Choose a Key (tonic 0–11) and a Scale from 14 types.
  2. Optionally enable Key/Scale Detection so the plugin determines the key itself (Auto / OpenVoxKey / Sidechain).
  3. Set Speed and Amount to shape how aggressively the pitch is pulled to the target.

Auto mode is fast and ideal for live performance, quick demos, or when you want the correction to follow a scale automatically.

Graphic mode (curve editor)

In Graphic mode, you draw the exact pitch curve the audio should follow over time. The plugin follows your drawn curve rather than snapping to the scale automatically (though snap-to-scale is available while editing).

  • Points are connected by linear interpolation.
  • The curve is time-synced to your DAW transport (via ARA2 when available).
  • Snap to scale, grid, copy/paste, undo/redo, presets, and PNG export are all supported.

See the Curve Editor page for the full editing workflow.

The core correction controls

These controls shape how the correction is applied in both modes.

Correction Mode (Modern / Transparent)

Setting Character
Modern Tight, aggressive correction — notes are locked to the scale firmly.
Transparent Gentle, subtle correction that preserves more of the natural performance.

Speed (0 – 200 ms)

Controls the retarget envelope — how quickly the pitch glides toward the target note.

Speed Result
0 ms Instant correction — the classic "T-Pain" robotic effect.
~50 ms Fast but smooth — the Antares default feel.
200 ms Slow, very natural correction with almost no audible effect.

Tip

Speed is an exponential smoothing curve: after one time constant it reaches ~63% of the target, after three ~95%, and after five ~99%. In practice, don't think in exact numbers — trust your ears.

Amount (0 – 100%)

The dry/wet blend. 100% delivers fully corrected audio; 0% is the unprocessed signal. Lower amounts are useful for natural "always-on" tuning.

Humanize (0 – 50 cents)

Adds subtle random pitch fluctuation to avoid an overly mechanical, "perfect" sound. Higher values sound more human; lower values sound tighter.

Vibrato Preserve (0 – 100%)

Retains the singer's natural vibrato through correction. At higher values the natural vibrato survives tuning; at lower values it is flattened out.

Voice Type

Constrains the pitch-detection range to a specific voice so the detector locks onto the intended register:

Voice Type Frequency range
Universal 30 – 1000 Hz
Bass 82.41 – 329.63 Hz
Baritone 110 – 440 Hz
Tenor 130.81 – 523.25 Hz
Alto 174.61 – 698.46 Hz
Soprano 261.63 – 1046.50 Hz

Latency Mode

Trades latency against processing quality:

Mode Latency
Direct Monitoring 10 ms
Low Latency (default) 12 ms
Quality 20 ms
Safe 30 ms

Quick workflow tips

  • Live: use Auto mode, Direct Monitoring latency, and a speed around 30–60 ms.
  • Ballad / slow track: raise Speed toward 150–200 ms and keep Vibrato Preserve high.
  • Pop / tight: use Modern correction with a low Speed.
  • Background vocal layers: lower Amount (60–80%) so they sit naturally behind the lead.