This paper discusses the complexities and implications of interview transcription in qualitative research. While often seen as a behind-the-scenes task, transcription is argued to be a powerful act of representation that can significantly impact the understanding of participants' experiences and the outcomes of the research. The authors explore two dominant modes of transcription: naturalism, which aims to capture every utterance in detail, and denaturalism, which focuses on the substance of the interview rather than the mechanics of speech. They highlight the constraints and opportunities of different transcription styles and advocate for a reflective approach where researchers interrogate their transcription decisions to ensure they align with research objectives and participant confidentiality. The paper is grounded in the authors' experience conducting a highly sensitive study on HIV-positive men who have sex with men (MSM) and their disclosure practices. They discuss the challenges of transcribing sensitive data, such as the need to balance detail and confidentiality, and the importance of reflecting on transcription choices to create trustworthy qualitative data. The authors recommend a balanced approach that combines elements of both naturalism and denaturalism, tailored to the specific research questions and goals. They also suggest the use of a codebook to standardize transcription processes, especially in large studies with multiple transcribers.This paper discusses the complexities and implications of interview transcription in qualitative research. While often seen as a behind-the-scenes task, transcription is argued to be a powerful act of representation that can significantly impact the understanding of participants' experiences and the outcomes of the research. The authors explore two dominant modes of transcription: naturalism, which aims to capture every utterance in detail, and denaturalism, which focuses on the substance of the interview rather than the mechanics of speech. They highlight the constraints and opportunities of different transcription styles and advocate for a reflective approach where researchers interrogate their transcription decisions to ensure they align with research objectives and participant confidentiality. The paper is grounded in the authors' experience conducting a highly sensitive study on HIV-positive men who have sex with men (MSM) and their disclosure practices. They discuss the challenges of transcribing sensitive data, such as the need to balance detail and confidentiality, and the importance of reflecting on transcription choices to create trustworthy qualitative data. The authors recommend a balanced approach that combines elements of both naturalism and denaturalism, tailored to the specific research questions and goals. They also suggest the use of a codebook to standardize transcription processes, especially in large studies with multiple transcribers.