This study analyzed the unique characteristics of the audio description that transforms video language into spoken language based on Christian Metz film semiotics theory. As a result of the analysis, the audio description which has a perfect relationship with spirits reflects Metz’ cinematic attributes that each short in the movie is incorporated into the story. The Metz five modes of expression provided appropriate concepts to explain the function and characteristics of the audio description. Metz’ theory of film, which recognizes images and sounds as equal elements, is consistent with the characteristics of the audio description, which considers images and sounds as a balance of mutuality. The audio description is focused on ‘culture code’ mainly in film semiotics theory, and the characteristic of this audio description is the point where it is combined with the narrative of the film emphasized by Metz. In addition, the audio description composed of a continuum of predicate elements faithfully reflects the narrative of the film in that it integrates auditory elements such as voice and sound to faithfully transmit the narrative of the film to the visually impaired in a limited time. In order to fulfill the function of creating discourse starting from the narrative of film, audio descriptor should keep neutrality and objectivity so that film and other discourse are not formed, and not to strengthen or weaken the discourse or ideology produced by film.