Formant Synthesis

vocal tract resonances · vowel space · F1-F2 plane

500 Hz
1500 Hz
Voiced
Physics: Vocal tract formants: resonances of the ~17cm mouth-throat tube. F1 ≈ tongue height (low vowels: high F1), F2 ≈ backness (front vowels: high F2). Vowels: /i/ F1=270,F2=2300; /a/ F1=730,F2=1090; /u/ F1=300,F2=870. LPC (Linear Predictive Coding): model vocal tract as all-pole filter with ~12 coefficients. Speech production = source (glottal pulse) × filter (vocal tract) × radiation. Vocal tract area function → formant frequencies via Webster horn equation. Articulatory synthesis models tongue, jaw, and lip positions.