A practical guide to calculating vocal tract length and scale-invariant formant patterns

Area of Science:

• Bioacoustics
• Phonetics
• Animal Communication

Background:

• Formants (vocal tract resonances) are crucial for speech analysis and behavioral science studies.
• Estimating vocal tract length and creating scale-invariant formant patterns are key challenges.

Purpose of the Study:

• To present a theoretical framework and practical tools for acoustic analysis of formants.
• To offer open-source software solutions for formant measurement and normalization.
• To improve the accuracy of vocal tract length estimation and scale-invariant representations.

Main Methods:

• Utilizing R packages soundgen and phonTools for acoustic analysis.
• Employing formant_app for integrated formant annotation and correction.
• Implementing intrinsic (formant ratios, geometric mean) and extrinsic (averaging, mixed models) normalization methods.
• Using the estimateVTL function for vocal tract length estimation and scale-invariant vowel space calculation.

Main Results:

• Automatic formant measurement can be error-prone; manual correction is recommended.
• Intrinsic normalization methods offer simple scale-invariant representations.
• Extrinsic normalization methods provide more accurate scale factor estimates by pooling data.
• Demonstrated practical application with example datasets and R code.

Conclusions:

• Accurate formant analysis requires careful measurement and appropriate normalization techniques.
• Recommended practices include recording multiple vocalizations, measuring at least four formants, manual verification, and using suitable statistical tools.
• The provided tools and framework enhance the study of vocal tract acoustics across disciplines.