Subtitling in a community setting such as healthcare can satisfy the needs of culturally and linguistically diverse (CALD) members of the community. It consists of a variety of stages such as transcription, spotting, translation and review. This qualitative study utilising content analysis aims to measure the quality of transcriptions by students of translation and interpreting (T&I) in Turkey as well as speech recognition tools against that of a native speaker. The data was analysed manually. The findings show that in transcribing authentic content on healthcare, student translators experienced challenges including parts of speech and textuality due to the lack of expertise in healthcare translation, poor bilingual skills as well as the speaker's pace and accent. Speech recognition tools produced better results than student translators, but occasionally had some discrepancies which can be attributed to such reasons as collocations and speaker-related issues. T&I students can be trained with the aid of speech recognition tools to make sure that transcriptions are done more effectively or post-editing skills are improved. Further studies can focus on students at varying levels of language or who have taken a course in AVT, on professional translators, or on other settings of community-based translation.