Phonetic-Phonological Fundamentals of Synthesis of Kazakh Speech

Abstract

At present, information and digital technologies are widely used in the humanities. In this regard, it is necessary to conduct various studies to improve the status of the Kazakh language and expand its functional use. In the era of digital technologies, experimental phonetic methods are increasingly used in the field of linguistics. A computer program is used to study segmental and supersegmental units. Experimental research can identify and describe the features of oral discourse. The phenomenon of digitization allows for the analysis of voluminous texts. It should be divided into syntagmas, phonetic words, then syllables and the smallest phonetic unit, and their acoustic-articulatory characteristics should be determined. The smallest linguistic units — phonemes — are taken as the source material for the development of speech synthesis. Currently, research on the synthesis and recognition of the Kazakh language is particularly relevant. Speech synthesis involves the presence of an automatic procedure for forming the current melody contour, pitch, phoneme duration and pause duration based on the analysis of certain properties of the input text and its prosodic marking. Digitization processes are widely used in linguistic research. Computer programs are used in corpus linguistics, text automation, and speech synthesis development. The article is devoted to determining the phonetic-phonological foundations of speech synthesis. Based on experimental phonetic research, the selected text is divided into sound units — phonemes and, through a special program, converted into sound text. Phonetic units in the stream of speech undergo various sound changes. Digital sound data entered into the computer's memory is converted into a speech signal. Modifications of phonemes, the weakening of some phonemes, their acquisition of properties of other phonemes, the reasons for the processes of reduction and assimilation are analyzed. Variants and variations of phonemes in weak position are identified with a description of their prosodic properties. Each phoneme in a word is reflected in its allophones. The sound fund of word synthesis is based on acoustic-articulatory characteristics. When studying the sound fund, identifying linguistic and extralinguistic causes of phonetic changes in the flow of speech, and determining the typological patterns of sound transformations, it is possible to synthesize texts of various styles. Thus, the pronunciation of sound units in various shades, phonetic changes in sounds, identifying the reasons for the modification of phonemes and allophones, and the general typology of sound changes are necessary for applied linguistics, in particular, for creating a speech synthesis platform. The basic principle of speech synthesis is the creation of a phonetic-phonological database, the main elements of which are allophones of phonemes found in a given language. Changes in invariant phonemes in speech, their implementation on the syntagmatic axis are studied using a special program.