Vocal synth models vocal tract as branching acoustic tubes, not spliced phonemes
flngr · x · 2026-11-09
Developer jamesonhodge shared a self-built vocal synth that models the vocal tract as a network of branching acoustic tubes, unlike Vocaloid's approach of stitching and pitch-shifting phoneme recordings. He is still tuning naturalness against recordings of real singers but is already impressed by the result — shared alongside a meme showing a glowing vocal-tract diagram labeled "Doo Doo Bah Doo."
More from Fun
- Gary Marcus digs up 7-year-old tweet: Tesla FSD still not fully autonomous — GaryMarcus · 2026-11-09
- AI meme: Jensengate shipped before Qwen4, Fable5.5, or even GTA6 — bclavie · 2026-11-09
- Wolf meme: this critic turns hostile every time OpenAI announces Voice updates — Angaisb_ · 2026-11-09
- Miyazaki: Young anime artists who only study anime are 'spinning wheels' — threepointone · 2026-11-09
- "Lori Huang is low-key the most powerful person in the world right now" — AIandDesign · 2026-11-09
- AI circle meme jokes Anthropic will drop fictional 'Fable 5.5' before IPO — haider1 · 2026-11-09