Abstract
The article focuses on determining the prosodic characteristics of texts used in educational platforms for the development of a speech synthesis system. This is обусловлено тем, что without taking into account the intonational characteristics of each sentence, the program cannot correctly determine which intoneme should be used in its pronunciation. At present, the intonational features of the Kazakh language have been comprehensively analyzed in the studies of Academician Z.M. Bazarbayeva, which identifies eight types of intonemes characteristic of the Kazakh language. In this regard, texts intended for voice-over narration were annotated with the corresponding intonemes, indicating their specific features. As research material, textbooks from educational platforms in both natural sciences and humanities were used, including Anatomy , Criminalistics , Management , and Kazakh Literature . In developing the speech synthesis system, intoneme models were employed as a reference for narrator reading. To ensure high-quality speech synthesis in the intonational (prosodic) aspect, types of fundamental frequency patterns (rising-falling, falling-rising, and level), as well as tempo parameters (fast and slow), were taken into account. Intonemes in the Kazakh language determine whether an utterance is complete or incomplete, convey its emotional coloring, and indicate the type of intonation-general or special question. Therefore, a systematic representation of intoneme models is essential. Thus, it was modeled for the speech synthesizer that the tonal level must necessarily fall at the end of a sentence, while the preservation of a rising or level tonal pattern in the middle of a sentence indicates the incompleteness of the thought; these features were incorporated into the system in the form of corresponding models.