Multimedia documents are increasingly used to disseminate specialized scientific knowledge. They are addressed to, and accessed by, different audiences: experts, students (with differing degrees of specialization) and general audiences. This range of audiences and products can help to bridge the gap between scientific communities and the rest of the population. Multimedia documents convey verbal and non-verbal information through visual and acoustic channels. The multiplicity of codes and channels both helps the acquisition of knowledge and allows for the inclusion of different types of accessibility resources, such as audio description for the visually impaired (AD). This article focuses on audio description of dynamic images in non-fiction scientific genres, including documentaries and multimedia presentations. It discusses current research on images, scientific translation and accessibility, analyzes existing audio-described documentaries, and proposes alternatives that can improve visual accessibility to multimedia scientific texts in different formats.