Abstract
Purpose.
The article is devoted to the comparison of spoken speech and written (recognized) text in the aspect of code transformation. Traditional and modern approaches to the relationship between oral and written speech are presented.
Results.
The research material was code transformations obtained as a result of converting spoken speech into recognized text using speech recognition technology from VKontakte – ASR (Automatic Speech Recognition), working on the basis of neural networks. The subject of the study is the assessment of code transformations in the aspect of identity and difference between spoken and recognized speech. The speech given by the Minister of Foreign Affairs of the Russian Federation S.V. Lavrov at the UN Security Council meeting in New York 2023 was chosen as the speech. Performance time – 00:24:22. The number of words in the recognized text is 2847. Units of code transformations have been identified. A code transformation unit is a unit of transformations and modifications, which occupies the position of a component that distinguishes between spoken speech and recognized text. The most common distortion is units that do not coincide in grammatical form with the standard in the Russian language.
Conclusion.
The author came to the conclusion that the quantity and quality of distortions do not affect the recipient’s understanding of the text, due to the identification of modifications with the standards of the perceptual base and taking into account the context, which indicates the high quality of the recognized text.