Package index • shinytone

Launch the app

Run the bundled Shiny UI locally — same interface as the hosted version, no upload limits, recordings stay on your machine.

run_app(): Launch the Shinytone app locally

Normalisation

By-speaker semitone or z-score f0 normalisation, plus landmark-based time normalisation for multisyllabic words.

normalise_f0(): Normalise f0 by speaker
normalise_time_landmarks(): Landmark-normalised time columns
normalise_time_token(): Whole-token 0-1 time normalisation

TextGrid landmarks

Read interval tiers from Praat TextGrids and attach per-frame segment labels and boundaries (e.g. syllables) to long-format f0 data.

tg_interval_tiers(): Interval-tier names across a set of TextGrids
assign_tier_landmarks(): Assign each time to its interval in a TextGrid interval tier
attach_landmarks(): Attach TextGrid landmark columns to a long-format f0 data frame

Outlier and artefact inspection

Token-level outlier detection by speaker z-score and by speaker x tone level, plus sample-level pitch-tracking artefacts and a low-intensity check (Sundberg 1973, Steffman & Cole 2022).

inspect_f0(): Inspect f0 data for token-level outliers and sample-level jumps
flag_outliers(): Flag per-token f0 outliers using by-speaker z-scores
flag_level_outliers(): Flag tokens whose overall f0 level is unusual for their speaker and tone
flag_low_intensity(): Flag low-intensity f0 samples within tokens
flag_pitch_jumps(): Flag sample-to-sample f0 jumps within tokens

Contour clustering

Unsupervised grouping of tokens by f0-contour shape to discover candidate tone categories when the number of tones is unknown (Kaland 2023), with feature extraction and a choice of how many groups.

cluster_f0(): Cluster f0 contours into candidate tone categories
cluster_features(): Build per-token feature vectors for f0-contour clustering
choose_k_f0(): Diagnostics for choosing the number of clusters (candidate tones)
cluster_mdl(): Recompute MDL information cost for stored clusterings at a given bending
cluster_agreement(): Agreement between clusters and known tone labels

Sonification

Render an f0 contour as an audible waveform (pure tone, complex tone, or a source-filter synthesised vowel) so prototypical tones can be heard, not only seen.

sonify_f0(): Sonify an f0 contour

Curate tone labels

Re-label tone-category variants or exclude tokens, without overwriting the original labels.

apply_relabels(): Apply tone re-labels and exclusions to a long-format f0 table

Contour modelling

Token-level polynomial fits, mixed-effects growth-curve analysis, and GAMMs over tone contours, with AR1 correction and model-checking diagnostics.

fit_polynomial(): Fit Legendre polynomials to f0 contours, token by token
fit_gca(): Fit a Growth Curve Analysis (GCA) model to f0 contours
predict_gca(): Predict population-level f0 contours from a GCA fit
fit_gamm(): Fit a Generalised Additive Mixed Model (GAMM) to f0 contours
predict_gamm(): Predict population-level f0 smooth curves from a GAMM fit
diagnose_gamm(): Diagnose a fitted GAMM

Chao tone numerals

Convert mean or predicted contours into Chao tone numerals (reference-line, interval-based, robust FOR methods).

compute_mean_contour(): Compute the per-tone mean f0 contour from long-format data
contour_to_chao(): Convert per-tone contours to Chao tone numerals
classify_contour(): Classify a Chao tone numeral string as a shape

Bundled data

A small citation-tone corpus that ships with the package for examples, vignettes, and the Shiny app’s “Try with our sample data” button.

sample_f0: Sample f0 contour dataset

Package overview

shinytone shinytone-package: shinytone: A Citation Tone Research Hub