The voice of terror in the silence of the image.

The concept of intersemiotic translation enhances audio description for horror films.

While the beautiful young woman refreshes herself under a shower, the silhouette of someone entering the bathroom can be seen. Suddenly, the figure of an old woman forcefully opens the shower curtain. The young woman turns to the old woman, startled, and lets out a scream, which is muffled by the knife blows delivered by the assassin. In another scene, a little boy is riding his tricycle through the carpeted corridors of a hotel when he is surprised by two mysterious twin girls, wearing light blue dresses, who stare at him intently, inviting him to “play with us… forever.”

The scenes described are from classic suspense and horror films: Psycho (1960) by Alfred Hitchcock, and The Shining (1980) by Stanley Kubrick. Certainly, the chilling sensation that the memory of these scenes can evoke in the reader is not solely due to the textual description. Much of the terror built by Hitchcock is due to the framing and editing, the use of light and shadow, and the unmistakable soundtrack. Kubrick, on the other hand, builds a surge of tension by filming the boy Danny in a single take from behind his tricycle, making the viewer follow the child through the long corridors. The difference between the sounds produced by the wheels in contact with the wooden and carpeted floors also contributes to raising the tension until the encounter with the twins.

Every element in cinema serves a function of producing emotions and meanings. Therefore, the mere description of scenes can provide an idea of ​​the film's plot, but it is insufficient to evoke the same emotions mediated by sight and hearing, especially in horror films. It was with this in mind that a doctoral research project in Applied Linguistics, conducted at the Institute of Language Studies (IEL) at Unicamp, focused on the work of audio description, a resource that ensures the accessibility of audiovisual works to blind or visually impaired people. Starting from the concept of intersemiotic translation, which views audio description as a translation process between different sign systems – from images to words – Marcella Wiffler Stefanini proposed innovations for the audio description of horror films. The thesis was supervised by Professor Érica Luciene Alves de Lima.

Translating fear

Translation goes far beyond simply converting a message from one language to another. According to the Russian linguist Roman Jakobson, there are three types of translation: intralingual, when we seek synonyms or paraphrase ideas within our own language; interlingual, when we exchange one language for another; and intersemiotic, when there is an exchange between different sign systems. "There is a consensus in research that audio description is an intersemiotic translation. However, professionals come from very different areas, so there is a lack of deeper theoretical understanding of why it is a translation," assesses the researcher.

In the case of audio description, the work involves converting visual signs into verbal elements, which are vocalized for visually impaired people. The challenge is to translate the nuances of cinematic language into verbal signs. For example, how to describe a zoom on a character's face, or an editing of shots that generates tension in the viewer? This requires a degree of subjectivity in the textual construction from audio describers. The problem is that this is discouraged by Brazilian audio description manuals, which recommend the greatest possible objectivity. "I see something in the images, you see other things. I interpret the images in one way, you may interpret them in another way," says Stefanini. Therefore, she proposes a form of audio description that can account for these aspects through adjectives, adverbs, and lexical choices, but also through intonation in the narration of the scripts.

The choice of horror films is justified by the ease with which viewers can recognize the emotions evoked by that genre. To test the possibilities of an audio description that involved subjectivity, Stefanini used two excerpts from the film. Caged (1997), a Brazilian medium-length film directed by Kleber Mendonça Filho. Initially, she created two audio description scripts: one following conventional protocols, and another with indications of subjectivity and expressive narration. After an initial test with five blind viewers, a single script was chosen, with variations in the audio describer's narration. This stage was supported by a blind consultant.
After the script was written and the two narrations recorded, the excerpts were shown to a group of 17 other blind viewers. All listened to excerpts with both formal and expressive narration, and then answered a questionnaire describing the emotions felt in both modalities. In the end, the conclusion was that 19 of the 22 participants in one of the two stages of the study preferred a more expressive narration, stating that this style seemed more aligned with the type of narrative.

Part of the process

According to the researcher, the research experience impacts how audio description work can be planned. “Everything depends on the film's genre and the associated emotions, but I advocate that the audio describer look at the cinematographic aspects, the framing, and take the soundtrack into account. This is so that the description doesn't anticipate what the other elements will bring to the scenes,” she points out.

This strategy directly influences the planning of audio description. According to the research advisor, there is a problem when it is not seen as part of the film production process. “Sometimes, because of a second of screen time, we can't insert a word of the audio description. If it's present from the beginning, several problems can be avoided and several choices can be made to favor this work,” argues Lima, who advocates for a broadening of the concept of translation as taught in universities. “Audio description is not limited to cinema. There's television, streaming, museum exhibitions. Therefore, I advocate that translation courses also include intersemiotic translation, as is the case with audio description.”