ABSTRACT This article investigates the use of pivot templates in audio description (AD), focusing on AD template translators' performance, strategies, and attitudes. Data was collected in a study where six participants, five audiovisual translators, and one describer, were asked to translate audio descriptions for five Spanish clips from a pivot English template into Polish. Following common subtitling practice, the templates were time-coded, and culture specific references (CSR) were not localized but explained in annotations. Three data types were collected: screen recordings, scripts translated into Polish, and post-task interviews. Quantitative analysis of the scripts revealed that the translated scripts retained most CSRs and, on average, transferred 94% of the CSRs from the pivot template and that translators rendered them via tactics known from subtitling. Thematic analysis of the interviews revealed that when deciding about a given tactic, translators took their own and the audience's frame of reference into account. They also considered how helpful the tactic might be in visualizing culture specific references. While the participants found CSRs to be the most challenging part of the task, they considered the task feasible mainly because of the identification and annotations, which they thought to be the main advantage of pivot templates.