Quantitative Target Approximation Model 1.0 (Prom-on, Xu & Thipakorn 2006)


Press the top "Simulate" button to display model-generated F0 curve and the underlying pitch targets.

Make changes in any of the editable boxes and then press "Simulate" again to see the effects

Upper Panel: Control parameters

Top row -- Categorical Parameters:

Tone: H = High; R = Rising; L = Low; F = Falling; N = Neutral (unstressed)

Duration: l = long; s = short; n = normal

Strength: s = strong; w = weak; n = normal

Lower rows -- Numerical values of the categorical parameters:

  Slope: in Hz/s, rising if > 0; falling if < 0; level if = 0.
  Height: in Hz

Duration: in second

Strength: larger value = greater strength

Initial pitch: Initial F0 of the sentence in Hz

Lower Panel: Pitch targets and surface F0 contours

x-axis: time (s); y-axis: F0 (Hz); Vertical lines: syllable boundaries; Thin grey lines: hypothetical underlying pitch targets; Blue curve: simulated surface F0.


Basic Assumption: Production of tone and intonation is a process of successively approaching syllabic pitch targets at varying pitch ranges with different levels of strength.

Syllabic pitch targets are specified by: y = b + a * x, where b is height and a is slope.

Duration specifies duration of each syllable, which is the temporal domain of a target.

Strength determines the amount of strength used to approach a pitch target. Greater strength leads to faster target approximation.

Pitch range specifies the range within which pitch targets are approximated (implemented in the current version only in primitives code).

tumblr site counter