Objective. Smoothed Cepstral Peak Prominence (CPPS) is an increasingly adopted acoustic index for quantifying voice quality and tracking treatment-related change. As CPPS varies with fundamental frequency (f₀), speech materials that systematically shift pitch may bias interpretation. In tonal languages, lexical tone modulates f₀, raising the methodological question of whether tone categories should be explicitly controlled in voice-assessment protocols. This study uses Cantonese as the illustrative case, as it comprises both level tones at multiple pitch heights and contour tones. Earlier Cantonese research observed tone-related variation in CPP (non-smoothed) using a corpus of only three base syllables, which constrained segmental coverage. The present design broadens the segmental inventory to at least ten base syllables to sample a wider range of consonant–vowel contexts and reduce item-specific bias. The objective is to determine the extent to which lexical tone (level vs. contour; pitch height) influences CPPS in Cantonese after controlling for SPL and segmental context.

Method. Thirty healthy native Cantonese speakers will produce monosyllables spanning level and contour tone categories. Speech level will be monitored, and CPPS will be computed from steady-state vowel windows. Data will be analysed using repeated-measures ANOVA with tone and base syllable as within-subject factors. Planned contrasts will assess tone differences, with Bonferroni-adjusted p-values for multiple comparisons; item effects will be evaluated within the base-syllable factor.

Results. The presentation will report repeated-measures ANOVA summaries for Tone and Base syllable, including Mauchly’s test of sphericity and Greenhouse–Geisser-corrected degrees of freedom where required. For interpretability, estimated marginal means (95% CIs) of CPPS by tone category will be shown, together with partial η² effect sizes. Planned contrasts will be presented with Bonferroni-adjusted p-values. Robustness will be examined via pre-specified sensitivity analyses: (i) comparing CPPS from mid vs. late vowel windows, and (ii) using alternative CPPS parameter settings. Interpretations will be evaluated for consistency across these checks.

Conclusion. Anticipated outcomes will indicate whether lexical tone should be controlled or balanced when developing CPPS-based voice-assessment materials for tonal languages, supporting methodological standardisation and cross-linguistic comparability. This study is partially supported by the Health and Medical Research Fund (HMRF; Ref. No. 21220351).