Human Vocal Tract Length
Description
This dataset comprises recordings collected from nine native American English speakers, including five males (M1, M2, M3, M4, M5) and four females (F1, F3, F4, F5), aged between 20 and 46 years. All participants reported no hearing or speaking impairments. Vocal Tract Length (VTL) Measurement 1st column: The lowest resonance frequency, Φ=c/4L, was calculated for a lossless uniform tube of length L, where c=35300cm/s (speed of sound). Here, L represents the Vocal Tract Length (VTL), which was measured using real-time magnetic resonance imaging (MRI) videos. Acoustic Features 2nd column: Fundamental Frequency (F0) in Hz. 3rd to 6th columns: Scaled formant values: f1/1: first formant scaled by 1 f2/3: second formant scaled by 3 f3/5: third formant scaled by 5 f4/7: fourth formant scaled by 7 12 Mel-Frequency Cepstral Coefficients (MFCCs). The 0th-order MFCC coefficient was excluded. For additional information: P. Vasquez-Serrano, J. Reyes-Moreno, Rodrigo Capobianco Guido, Alexander Sepúlveda-Sepúlveda, MFCC Parameters of the Speech Signal: An Alternative to Formant-Based Instantaneous Vocal Tract Length Estimation, Journal of Voice, 2023, ISSN 0892-1997, https://doi.org/10.1016/j.jvoice.2023.05.012