IRIS follows one clear live pitch, guides it toward the notes you choose, and leaves room for onset, vibrato, drift, and tone. Move from transparent repair to immediate stepped tuning, then build up to four scale-aware harmony voices around the lead.
Artwork placeholders for the detector, musical target, correction behavior, formant path, and harmony engine
Choose Bass, Low Voice, Mid Voice, High Voice, or Instrument range, and IRIS follows one credible dominant pitch. When a source becomes silent, unvoiced, or unstable, the engine falls back toward the aligned dry path instead of forcing an uncertain note
Set a root and select Chromatic, major and minor variants, Dorian, Mixolydian, pentatonic, Blues, or Custom. The twelve-note strip always shows the pitch classes that correction is allowed to target
Amount and Retune set correction depth and speed, while Onset, Humanize, Vibrato, and Tolerance decide which parts of the original performance should remain untouched
Turn Retune to its zero-millisecond endpoint, remove the tolerance and expressive allowances, and the same correction path becomes a deliberate hard, stepped tuning effect
Formant compensation keeps the spectral envelope from simply following every pitch move, while an independent formant control shifts vocal or instrumental character without changing the target note
Four fixed harmony voices move by diatonic scale steps. Each voice has its own level, pan, cents offset, causal delay, natural variation, and formant character for anything from one close third to a wide stack
IRIS separates detection, musical targeting, correction, formant handling, and harmony generation into a readable path. Every audible branch is kept aligned with the plug-in's reported latency, including soft bypass and the original signal.
The main controls describe musical intent rather than exposing the spectral machinery underneath, so the first useful result takes only a few decisions and every deeper adjustment remains audible.
Insert IRIS on one vocal or monophonic instrument and select the range that best contains its fundamental before choosing a root and scale
Start with Amount and Retune, then use Onset, Humanize, Vibrato, and Tolerance to preserve consonants, note attacks, sustained drift, and intentional modulation
Enable only the voices the arrangement needs, choose diatonic intervals, then place each voice with level, pan, detune, delay, humanize, and formant controls
Use latency-preserving bypass, wet/dry Mix, undo and redo, A/B snapshots, metering, and complete host-state recall to judge the result in context
6Factory starting points
The factory library moves from Natural Vocal and Transparent correction through Modern Tight and Hard Tune to a single Third Above and a fully spread four-voice Wide Stack.
Search and filter the library, favorite factory patches, and save or import portable .irispreset files under Documents/Nodoka/IRIS
