Transcribing Without Smoothing Away a Speaker's Voice
Launch library · evergreen read
It is tempting, when transcribing a recorded interview, to quietly correct grammar, remove repeated words, and tidy up sentences that would otherwise look awkward on the page. This kind of smoothing can strip away exactly what made the speaker's voice distinctive in the first place, replacing a real person with something more generic.
A transcript that reads too cleanly often loses the particular rhythm, hesitations and phrasing that gave the original recording its genuine personality on the page. It replaces an individual voice with something closer to generic, polished prose that could, in the end, belong to almost anyone reading it aloud.
Preserving some of this rawness, even where it makes for slightly less tidy reading on the page, usually produces a more faithful and more valuable record overall. Future readers can recognise it as genuinely belonging to the specific person who spoke those exact words at that particular time.