ABSTRACT
After decades of dedicating good efforts to increase the amount of television programing accessible to the hearing impaired, awareness about the need to deliver better closed captions has grown in the last years. As a result, the focus has been moved from the quantity to the quality of the accessibility services provided.
In an attempt to explore the subtitles currently delivered on television, this paper presents the main findings of a study aimed at analyzing the quality of the live closed captions provided during one of the most relevant political events of 2016: the final presidential debate in the United States. A corpus of more than 9,400 captions was analyzed according to the quality criteria defined by the Federal Communications Commission. The results obtained shed light on the completeness, placement, synchronicity and accuracy of the captions, but they also elicit interesting questions in terms of reception.
Notes on contributor
Nazaret Fresno holds a PhD in translation and cross-cultural studies, as well as an MA in audiovisual translation and another in comparative literature and literary translation. She is an Assistant Professor of translation and interpreting at the University of Texas at Rio Grande Valley, and her research interests include audiovisual translation and accessibility to the media (subtitling for the deaf and hard of hearing and audio description for the blind and visually impaired). She is also an external member of GALMA (Galician Observatory for Media Accessibility) research group.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1. The speed of subtitles can be measured using two different units: words per minute (wpm) and characters per second (cps). There is no universal conversion of wpm into cps but an approximate correspondence may be calculated taking into account that the average English word has 5 characters (Díaz-Cintas, Citation2008).
2. The accuracy rate of a given program is calculated in Canada by dividing the number of verbatim words in the captions by the total number of words in the oral speech.
3. In some live programs, when the emission ceases for commercial breaks or to transition to the subsequent show, the transmission of captions is interrupted. However, captioners can use a command in their equipment to force the delayed subtitles to appear on screen (Zdenek, Citation2015).
4. This is the average obtained when the talk shows, sports, soap operas, sitcoms, prime time dramas, news, live performances (no music), films and documentaries in Jensema et al.’s research are considered.
5. The Internet Archive is a non-profit organization aimed at supporting researchers in the field of communication. The author contacted them after failed attempts to obtain the videos directly from the aforementioned TV broadcasters (none of them responded to the author’s inquiries).
6. Captioning speed is defined in this paper as the speed value that subtitling software (Black Box) shows for each caption when existing .srt files are loaded.
7. The average captioning speed is calculated in this paper by diving the total number of characters/words in the subtitles by the total minutes/seconds of captions included in the video.
8. To facilitate comparisons with prior research, the speed calculations in this section will be provided in cps and wpm. It is important to note that these values were not calculated using the traditional conversion rate (an English word has 5 cps). Instead, the number of characters and words included in the captions of the debate was counted separately, allowing for more precise results.
9. The reduction rate is calculated by considering the difference between the total number of words in the audio of the debate and those included in the captions.