Title
Information on sounds and speaker identification in subtitles for the deaf and hard-of-hearing for Brazilian Netflix and DVDs
Conference name
9th International conference Media for all
City
Country
Spain
Modalities
Date
27/01/2021-29/01/2021
Abstract
Subtitling for the Deaf and Hard-of-Hearing (SDH) provides access not only to dialogue but also to the non-linguistic elements of audiovisual materials, such as noise, music, speaker identification (when it is not clear from the image who is speaking) and other prosodic features of language (whispering or shouting, for example). Despite the scarce number of studies on how these non-linguistic elements are translated in SDH, Nascimento (2013, 2018) proposed tags for the analysis of these elements so as to investigate the SDH of DVDs released in Brazil, the United States and France using a Corpus Linguistics methodology.
Making use of these tags (Nascimento, 2013, 2018), we analyzed the translation of non-linguistic elements in the SDH aired on Netflix and on DVDs in Brazil. For that, the subtitles were extracted, converted to a .txt file format, semi-manually tagged and analyzed with the linguistic analysis software WordSmith Tools 6.0. The results showed that the DVD subtitles presented technical problems related to the number of lines (some of them were 3- and even 4-lined) and their length of up to 54 characters per line. As far as subtitle speed is concerned, most Netflix and DVD subtitles were slow (less than 13,9 characters per second), or normal (between 16 and 20 characters per second). Subtitles with sound effects were present in 20% of the DVD subtitles, but only 14% of Netflix subtitles. Of these, the most frequent categories were sounds made by humans and sounds made by objects. Furthermore, background music is usually qualified on DVD subtitles, whereas almost half of the background music is not qualified on Netflix subtitles.
Research suggests that it is of the utmost importance to qualify the translation of music in SDH for a better understanding of the film (ARAÚJO; NASCIMENTO, 2011). In our sample, speaker identification occurred most frequently through the name of the characters as expected, but it also occurred through gender signaling, location (‘Man in the background’, for example) and other characteristics, such as nationality. However, it was not possible to identify a pattern on how sound, music and speaker identification are presented on both DVD and Netflix subtitles. In short, while subtitlers seem to be following most of the technical and linguistic parameters recommended by audiovisual translation literature, a pattern for the speaker identification and translation of other non-linguistic elements into subtitles do not seem to be of concern.
Making use of these tags (Nascimento, 2013, 2018), we analyzed the translation of non-linguistic elements in the SDH aired on Netflix and on DVDs in Brazil. For that, the subtitles were extracted, converted to a .txt file format, semi-manually tagged and analyzed with the linguistic analysis software WordSmith Tools 6.0. The results showed that the DVD subtitles presented technical problems related to the number of lines (some of them were 3- and even 4-lined) and their length of up to 54 characters per line. As far as subtitle speed is concerned, most Netflix and DVD subtitles were slow (less than 13,9 characters per second), or normal (between 16 and 20 characters per second). Subtitles with sound effects were present in 20% of the DVD subtitles, but only 14% of Netflix subtitles. Of these, the most frequent categories were sounds made by humans and sounds made by objects. Furthermore, background music is usually qualified on DVD subtitles, whereas almost half of the background music is not qualified on Netflix subtitles.
Research suggests that it is of the utmost importance to qualify the translation of music in SDH for a better understanding of the film (ARAÚJO; NASCIMENTO, 2011). In our sample, speaker identification occurred most frequently through the name of the characters as expected, but it also occurred through gender signaling, location (‘Man in the background’, for example) and other characteristics, such as nationality. However, it was not possible to identify a pattern on how sound, music and speaker identification are presented on both DVD and Netflix subtitles. In short, while subtitlers seem to be following most of the technical and linguistic parameters recommended by audiovisual translation literature, a pattern for the speaker identification and translation of other non-linguistic elements into subtitles do not seem to be of concern.