Predict population-level f0 contours from a GCA fit

Generates population-average per-tone f0 curves from a fitted fit_gca() model. Useful for plotting predicted contours, comparing tones at a glance, or feeding into downstream Chao numeral summarisation via contour_to_chao().

Usage

predict_gca(gca_obj, n = 100)

Arguments

gca_obj: An object of class "shinytone_gca" returned by fit_gca().
n: Number of time points across [0, 1]. Default 100.

Value

A data frame with columns time, f0_predicted, tone.

Details

Internally:

Build a (time, tone) grid with n evenly-spaced points across [0, 1] for every tone level the model knows about.
Re-compute the orthogonal polynomial basis on that grid using the cached coefficients from fit_gca() (this ensures the basis matches what the model was fit with, not a fresh one).
Call stats::predict() on the lme4 model with re.form = NA, so only fixed effects contribute (random effects are integrated out to the population mean).

Note that the returned predictions are on the scale of the f0 column used to fit the model (typically semitones, if you passed f0 = "f0_st" from normalise_f0()).

References

Mirman, D. (2014). Growth Curve Analysis and Visualization Using R. Chapman and Hall/CRC.

Xu, C., & Zhang, C. (2024). A cross-linguistic review of citation tone production studies: Methodology and recommendations. The Journal of the Acoustical Society of America, 156(4), 2538–2565. doi:10.1121/10.0032356