OptionalaaOpen vowel as in "car". Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalCHch, j, sh: rounded, protruded lips ("chair", "join", "she"). Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalDDt, d: tongue tip on the alveolar ridge ("tip", "doll"). Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalEVowel as in "bed". Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalFFf, v: lower lip against the upper teeth ("fat", "vat"). Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalIVowel as in "tip". Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
Optionalkkk, g: tongue dorsum against the soft palate ("call", "gas"). Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
Optionalnnn, l: tongue tip up, mouth slightly open ("lot", "not"). Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalORounded vowel as in "toe". Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalPPp, b, m: lips pressed together ("put", "bat", "mat"). Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalRRr ("red"). Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalsilSilence, the closed resting mouth. Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalSSs, z: teeth nearly closed ("sir", "zeal"). Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalTHth: tongue tip between the teeth ("think", "that"). Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
OptionalUClose rounded vowel as in "book". Schema constraints: Minimum: 0; Maximum: 1; retained by Human's schema refinement.
Blend weights of the 15 standard visemes, each from 0 (absent) to 1 (full mouth shape), named as in the Meta (Oculus) Lipsync viseme reference, which follows the MPEG-4 viseme set. Omitted visemes weigh 0; weights are not required to sum to 1.
AutoMovieHumanMouthExpressionmaps each viseme onto lip units, tongue controls and a jaw opening by the articulation of its sounds (Ladefoged and Johnson, A Course in Phonetics), and sums them by weight.