Fig. 1

Visualization of a TextGrid file via Praat’s editor, including five tiers: phonemes, syllables, words, and phonetic and orthographic transcription, respectively. FalaBrasil’s G2P [10, 11] and syllabification [12] library provide phonetic and syllabic tokens, respectively, from text. Praat [7] is responsible for plotting the audio’s waveform in both time and frequency domains, while Kaldi [6] provides the temporal marks for the dark-blue vertical bars that keep tokens apart from each other