Formant Extractor

Parameters

Formant Width — Bandwidth around each detected formant peak that gets included in the extraction, in Hz, 50–500. Tight widths (50–100 Hz) capture only the formant’s central peak — surgical isolation. Wider widths (200–500 Hz) include more of the surrounding spectrum — softer extraction with more vowel character preserved. Pair with mask_strength to balance how aggressively the formant region is isolated from everything else.

Mask Strength — How strongly the extraction mask is applied, 0–100%. 0% bypasses the mask (input passes through unchanged on the extracted output). 100% applies the mask fully (only formant content survives in the extracted output, everything else moves to the rejected output). Lower values produce a softer separation where some non-formant content remains in the extracted stream — useful when you want a “vocal-leaning” sound rather than a clinical isolation.

Additional controls

Selection — Which formants to extract:

• F1 — first formant (typically 200–1000 Hz, the lowest vocal resonance; carries vowel “openness”).
• F2 — second formant (typically 800–3000 Hz; carries vowel “frontness”).
• F3 — third formant (typically 1700–4000 Hz; contributes to vowel “color” and consonant articulation).
• F4 — fourth formant (typically 3000–4500 Hz; contributes to voice “presence” and timbre).
• All — extracts all four formants together. Use All for full vowel/voice extraction; individual formants for surgical control over specific vocal frequencies.

Order — Cepstral analysis order, 8–60 (visible in advanced mode). Sets the resolution of the formant detection — higher order produces more detailed envelope tracking but can pick up false formants on noisy material. 16 is balanced for most voices; 30+ for high-resolution extraction on cleanly-recorded vocals.

About Formant Extractor

Formants are the resonant frequencies of the vocal tract — the peaks in a vocal spectrum that distinguish “ah” from “ee” from “oh” without changing pitch. This node detects formants using cepstral envelope analysis and produces two outputs: the formant content (extracted) and everything else (rejected — typically the pitched harmonic content underneath). Use it for: vocal processing where you want to treat the timbre/vowel content separately from the pitch material, formant filtering for sound design (route a synth through extracted vocal formants for vocoder-style effects via cross-spectral nodes), or surgical vowel isolation. The two outputs sum back to the input — like all K2K extractors, no information is lost in the split.

Generated 2026-05-05 from K2K_Dev@96730bdc by scripts/gen_lexique.py. Edit _intros/ or _overrides/, not this file.