Towards Reliable Prosody Detection in Teaching Practical Phonetics: A Sustainable Digital Approach

<ns5:p> This research focuses on providing quality education to specialists in the field of linguistic studies, phonology, theoretical and practical phonetics by creating an innovative automated service which can ease the linguists endeavour to transcribe and examine suprasegmental features of human speech prosody. The authors attempted to design a docker-based architecture for prosody labelling. The pre-processing stage compiled 47 manually marked recordings. The data-processing stage produced a 1.9 GB Parquet corpus of 882 denoised, labelled clips via <ns5:italic>librosa</ns5:italic> , <ns5:italic>noisereduce</ns5:italic> , and <ns5:italic>Label Studio.</ns5:italic> The model selection stage compared 4909 scikit-learn models and 120 CNNs. The strongest classical approach (Random Forest) reached 0.373 accuracy, whereas the best CNN scored 0.455. Having recognised the prototype limitations, the researchers have scheduled a roadmap for improving the tool. The suggested model applies the principles of machine learning to automatically generate prosodic analysis that extends beyond individual sound segments and reflects suprasegmental aspects of prosody including rhythm, intonation, and phrasal stress. </ns5:p>