Formants

See AND capture the VOWEL of a voice. Wire a mono signal in (a mic) and the kernel finds its FORMANTS (the vocal-tract resonances that name a vowel) by linear prediction, plotting the live (F1, F2) point on the classic vowel chart — sing a→e→i and the dot walks the vowel space among the cardinal-vowel guides. Press Rec and sing to CAPTURE the formant trajectory (Viterbi-cleaned on stop), then Play to replay it. Wire the Formants output into a Formant Filter and a carrier through that filter SPEAKS your vowels — live as you sing, or your recorded take (saved with the project). The FILTER half of the source-filter model (the Pitch tuner is the SOURCE half). The LPC + the committed track live in the kernel (formant::detect_formants / track_formants); audio.js taps the samples it owns.

Port	Signal
audio_in	audio · audio

Port	Signal
audio_out	audio
Formants	audio · formant

Parameter	Range	Default
Rec		False
Play		False
Loop		True

Inputs

Outputs

Parameters

Used in these rigs

More analysis nodes