SAMPA kodavimas lietuviškų garsynų fonetinės transkripcijos užrašymui

This article describes our proposal of Lithuanian Speech Assessment Methods Phonetic Alphabet (SAMPA) for encoding transcriptions of Lithuanian speech corpora. The recommendations of how SAMPA design conventions can be adapted and extended to handle phonetic particularities of Standard Lithuanian are formulated. The codebook consisting of Lithuanian spelling symbols, corresponding IPA (International Phonetic Alphabet) symbols, and proposed SAMPA equivalents is constructed. We have taken the approach similar to that of other 24 world languages of adapting X-SAMPA to Lithuanian, paying attention to readability, unambiguity and unit independence of resulting phonetic transcriptions. We hope these properties will make Lithuanian SAMPA useful for researchers within Lithuanian language technologies community working on tasks of speech synthesis, speech recognition and automatic speech annotationVytauto Didžiojo universiteta