Aspects of Visual Content Covered in the Audio Description of Arabic Series: A Corpus-assisted Study

AD as a Mode of AVT

Translation is an essential field of study that deals with different genres and modes. It is an activity that transfers the meaning and style from the source text into the target text. According to Newmark (88), translation is “rendering the meaning of a text into another language in the way that the author intended the text.” The three types of translation are intralingual, interlingual, and inter-semiotic. Intralingual translation renders texts within the same language (Al-Abbas and Haider; Al-Abbas et al.). In interlingual translations, texts are made understandable for audiences unfamiliar with the original language (Al-Zgoul and Al-Salman; Aldualimi and Almahasees). Jakobson states that interlingual translation or “translation proper” converts verbal signs in one language to the verbal cues of another language. The third type is inter-semiotic, in which the verbal signs are interpreted using signs of a nonverbal sign system (Torop).

Due to the widespread use of Arabic in the Middle East and North Africa and the significance of Arabic as the language of the Holy Quran, Arabic–English translation and vice versa is a crucial area of study. Literary translation, scientific and technical translation, legal translation, and commercial translation are a few examples of the numerous kinds of translation between Arabic and English. For the translation to be successful, it requires a deep understanding of both languages and their cultural nuances. It is crucial for translators to be knowledgeable about tools and software used for translation, as well as staying current with the most recent advancements in the industry.

Due to the spread of audio-visual products, such as movies, television shows, series, video games, and advertisements and commercials, there has been an increasing demand to translate and localize them into different languages. This gives rise to a new field in translation studies, namely AVT, which has different modes, including AD (Haider and Hussein; Haider et al.; Samha et al.). Mazur (228) defines AD as “audio commentary that describes the relevant visual elements of work (as well as meaningful sounds that may not be clear to the target audience) so that the work constitutes a coherent whole for the audience.” AVT mode is popular in the United States, Japan, and several European countries, gaining recent popularity in the Middle East. Hyks pointed out that translation and AD are two thoroughly related activities. It could be said that the role of AD is turning visual information into verbal information, which helps people who are blind or have impaired vision understand movies and other audio-visual displays better.

The origins of AD can be traced back to 1981 in the United States, when Margaret and Cody Pfanstiehl, the founders of the Metropolitan Washington Ear, began providing listening services for the blind, visually impaired, and physically disabled people who could not read print properly in an effort to promote inclusivity (Snyder). The Pfanstiehls collaborated with Arena Stage for American Theatre to launch an AD of Major Barbara’s English Play. As AD began to be more widely adopted by various platforms and industries, improving differently abled people’s lives became a normal part of inclusion strategies. Previously, people with visual impairment relied heavily on their able-sighted friends to whisper comments to them about what was being performed on the theatre stage. This makes it clear that AD has impacted lives throughout history. AD could additionally be used for pedagogical purposes and requirements, as it raises awareness about disability and could further develop critical thinking skills in classrooms through a guided description of visual material (Kleege and Wallin; Schmeidler and Kirchner).

The purpose of AD in filmed programs and performances, such as movies, television programs, theatre, opera, dance performances, or football matches, is to convert the main visual aspects of a performance into an informative verbal description that explains the program or event. This is related to one of AD’s most essential goals, which is enabling blind and partially sighted viewers to enjoy audio-visual products.

AD can be controversial when applied to audio-visual products like movies and series because these products can be audio-described in many ways. Overall, audio-visual products have a far-reaching influence, so every aspect is carefully considered in the pre-production, production, and post-production stages. According to Piety, the first academic record of AD is found in a dissertation written in 1975 by Frazier, in which the author discusses AD as a way to enhance blind listeners’ comprehension.

Empirical Studies

This section reviews several studies that discuss AD services and how people with vision impairments reacted to them. Fels et al. compare the reactions of sighted and blind users to first-person-described video information and conventional (third-person) descriptions in relation to the animated Canadian sitcom Odd Job Jack. The research participants watched 10 min of four video clips, and the findings showed that all of the blind participants commended the video description, preferring the first-person style to the conventional one.

In the same vein, Arma adopts a corpus-assisted approach to syntactically and lexically analyze Italian and English audio-described scripts of the movie Chocolat. Using quantitative and qualitative methods to analyze the texts using AntConc software provides reliable results for corpus-driven text analysis. Examples of sentence structure, relevance, and objectivity are retrieved manually. The results show that the Italian script contains a higher number of secondary clauses when compared to English. Arma recommends further research on the hearing habits of the visually impaired community to determine the audience’s satisfaction with different approaches to AD.

Similarly, Vermeulen and Ibáñez Moreno examine AD as a tool for intercultural competence (ICC) by emphasizing AD as a culture-based translating activity and implementing it in foreign language classes. The study participants included 153 Belgian 3-year university students studying languages. The participants explored the benefits of using AD in the foreign language classroom, which promoted ICC. The study found that AD proves to be a helpful pedagogical tool to raise intercultural awareness, creating a link between “what we see” and “how we describe what we see.”

Walczak explores the impact of Polish AD on the immersion of the B/VIP consumers of this service. Two studies are conducted focusing on AD style and AD vocal delivery. The first research studied the standard and creative AD styles used by blind and visually impaired users. The second study, linked to AD vocal delivery, studied the reception of two AD voice types – human and synthetic, by the blind and visually impaired users for documentary and fiction movie genres. A questionnaire targeting a total of 72 participants was used to measure their engagement when exposed to different media. The results show that creative AD prompted significantly higher levels of presence for all participants than standard AD, and AD vocal delivery favored AD narrated by a human about a fictional story. On the other hand, presence rates for documentary clips were similar, with no statistically noteworthy differences in AD voice type: human and synthetic, which validate the hypotheses.

Likewise, Alnatsheh questions whether Arabic movies could have dialectal audio descriptions (DADs). The study includes quantitative and qualitative approaches, investigating the target audience’s preferences and determining if Qatari DAD can create an immersive environment. Two AD versions of the short Qatari movie Sh’hab are provided: one in modern standard Arabic (MSA) and the other in the Qatari dialect. The results show that the participants preferred MSA, which they found to be more eloquent, easy to follow, and described the actions in the scene more accurately.

Zengin Temırbek uulu et al. qualitatively examine the visually impaired and sighted participants’ comprehension of two versions of the same film, one with and one without supporting AD. This recent Turkish research aims to determine the extent to which AD contributes to visually impaired individuals’ comprehension of a film. When the movie is aided with AD, visually challenged participants are able to comprehend and describe the events similarly to sighted participants. These findings highlight the significance of AD assistance and its impact as a vital resource for transforming visual information into verbal information, which considerably enhances movie comprehension for the visually impaired. The study recommends AD producers to provide better support for the visually challenged.

In the context of the Persian language, Khoshsaligheh et al. focus on the quality of the Persian language AD by qualitatively analyzing three audio-described feature films produced in Iran in 2019. Intralingual AD is produced by an Iranian non-governmental organization called Sevina. The researchers’ goal is to determine whether the sampled AD transcripts provide a systematic approach based on Audio Description: Lifelong Access for the Blind project guidelines. The findings show a fairly consistent approach, even though there is room for improvement in future studies.

This study analyzes the AD of the Arabic series AlRawabi School for Girls. This might provide audio describers with further suggestions that help improve the quality of Arabic AD. In summary, the studies reviewed in this section discussed AD in several languages; however, there is limited discussion about this phenomenon in Arabic audio-visual products, a gap that this study attempts to fill.

Why AlRawabi School for Girls?

AlRawabi School for Girls is an Arabic six-part TV drama mini-series that premiered on August 2021. It sheds light on social and cultural topics infrequently discussed through Arabic-language entertainment, focusing on the topic of bullying young women. It rates 7.4/10 on IMDb, a well-known online information database related to streaming audio-visual content. Also, the maturity rating for the series is +16. The main characters include a cast of six girls named: Mariam, Noaf, Layan, Dina, Rania, and Ruqayya.

The AlRawabi School for Girls series is written by two young female Jordanians, Tima Shomali and Shirin Kamal, and produced by Filmizion Productions (on behalf of Netflix). Tima Shomali is the founder of Filmizion Production and is also a director and comedy actress, while Shirin Kamal serves as the Creative Music Supervisor of the series. AlRawabi School for Girls is a controversial Jordanian series because it openly portrays topics that are not openly discussed. Some described the series as featuring “uncomfortable” content and projecting a “false image” of Jordanian society (see Misra). On the contrary, others have praised the show for portraying young teenage Arab girls.

The female-centric teen series crucially discusses hot topics in Arab societies, such as bullying, women’s rights, psychology, religion, revenge, and toxic relationships, to mention a few. Surprisingly, it is subtitled in 32 languages, including English, Spanish, Italian, French, German, and Turkish, and dubbed in more than nine languages. It is also available in 190 countries.

Data Collection and Corpus Compilation

A corpus of AD of AlRawabi School for Girls series was compiled to be further analyzed by the researchers. After watching the six-episode series on Netflix, each lasting about 45 min, the researchers wrote down the full audio-described version of the series manually on an Excel sheet. The AD is mainly in MSA, although the series is produced in Jordanian vernacular. Since the Netflix material is protected and is not subject to any free or open license, it is important to note that the researchers contacted the streaming platform and were informed that Netflix allows for the use of its subtitles for educational purposes, which include facilitating learning, promoting knowledge acquisition, and supporting educational goals within formal or informal educational settings and contexts.

Figure 1 shows the available options for Netflix subscribers when watching AlRawabi School for Girls. These include audio and subtitles options. In the former, subscribers can select Arabic (original), i.e., in Jordanian Vernacular or AD. They can also watch a dubbed version in British English. The subtitled versions are intralingual (Arabic) and interlingual (including English).

Figure 1

Available audio and subtitles on Netflix for AlRawabi School for Girls.

Figure 2 shows an intralingual subtitling of the series, where the actress’s words appear at the bottom of the screen and read as “هلا، المدرسة صارت كابوسي” (Lit. the school is now my nightmare).

Figure 2

A screenshot from “ALRawabi School for Girls” series with Netflix Arabic subtitles.

Figure 3 includes the MSA intralingual subtitles of AlRawabi School for Girls extracted from Netflix (the text in black). It also includes the AD (the text in red).

Figure 3

A screenshot of the MSA intralingual subtitles and AD of AlRawabi School for Girls.

Corpus Linguistics (CL)

There are different CL techniques. These include frequency, clusters, concordance, collocation, and keywords (Baker; McEnery and Wilson). In this study, two techniques are used: frequency and concordance. Frequency shows a list of all words in the corpus, and concordance investigates words in context. In this study, Wordsmith 6 (WS6) (Scott), a CL software package, is used to generate word lists and concordance analysis.

Frequency analysis enables researchers to identify the most frequent words in a corpus and compare the list with other frequent words in several multilingual and monolingual corpora (Haider). In this study, the researchers conduct frequency analysis and investigate the data accordingly. Figure 4 shows a screenshot of frequency analysis using WS6.

Figure 4

Screenshot of frequency analysis on WS6.

Concordance shows all the hits of a search term in context. In this article, the researchers generate concordance lines for some particular words and divide them into thematic categories. Figure 5 shows a screenshot of concordance analysis.

Figure 5

Screenshot of concordance on WS6.

Research Procedure

After selecting and watching the AlRawabi School for Girls Arabic series, the AD script from Netflix was transcribed by manually writing it down. After that, the original Netflix MSA intralingual subtitles of the series were extracted directly from Netflix. Subsequently, the collected data from all six episodes were converted into plain text files (.txt) to be processed and read by WS6 CL software package. After processing the data using WS6, the AD and MSA intralingual subtitles of the series were compared and contrasted. Later on, the Ads were classified thematically based on the type of information they covered. Finally, the AD parts that involve code-switching between Arabic and English were explored.

Comparing AD to MSA Subtitle Corpora

Tokens, types, and type/token ratio (TTR) (Chotlos; Templin) are important, especially when comparing corpora. For example, AD and MSA intralingual subtitle corpora are illustrated in this study. These statistics reveal information about the features of the texts examined, whether the corpora are monolingual or multilingual.

The researchers used the wordlist tool in WS6 to generate corpus statistics for the six episodes as one unit and each episode separately. Table 1 shows the tokens, types, and TTR in the investigated AD corpus.

Token refers to the actual number of words in the corpus (running words), while type refers to the unique words in the corpus. Figure 1 shows that the second (768 words), fourth (677 words), and fifth (886 words) episodes include fewer words than the first, third, and sixth episodes with 1,005, 1,142, and 1,056 words, respectively. This might be because the important events contributing to the series’ plot occurred in these episodes, and this is why they are described more thoroughly to the target audience. This could also be attributed to the availability of portions, often referred to as “gaps” or “pauses,” between dialogues or scenes.

The Arabic language has several varieties, including Classical Arabic (CA), MSA, and Dialectal Arabic (DA) (Ferguson). MSA is the simplified form of CA and is considered the lingua franca of the Arab World (McCarus). Netflix provides MSA subtitles for most Arabic vernacular movies or series. The main reason behind using the corpus of the MSA subtitles of the series is to compare the frequencies of some words in the two versions, i.e., the AD and source script. Table 2 shows the size of the MSA subtitles for AlRawabi School for Girls. As the researchers stated before, the series is produced in Jordanian vernacular, but the researchers could not transcribe it due to time constraints, and this is why the intralingual MSA corpus is used.

Figure 5 shows that the actual size of the MSA intralingual subtitle corpus is 15,448 words, with 4,947 distinct words (types). The ratio of types to tokens is 32.06.

Table 3 summarizes the results of Tables 1 and 2.

Table 3 shows that the number of words in the subtitle corpus is larger than that of AD. This is expected because ADs are mainly added between dialogues or scenes. Such a variation is likely due to spatial and temporal constraints. AD phrases are inserted between the characters’ conversations, which is the silent part of the audio-visual scenes. This is why audio describers use phrases that deliver the message in the shortest possible way. The audio describers need to infer what the essential parts of a scene are and convey these to the blind or visually impaired audience.

Table 3 also shows that types are more common in the MSA intralingual subtitle corpus than they are in the AD corpus. The variety of the text’s vocabulary is more visible in the AD corpus (33.74) when compared to the MSA subtitles (32.06). This means that more content words are used in the AD version instead of the MSA corpus. The grammatical (function) words are used more in the MSA intralingual subtitles because there is enough space for the subtitlers to write complete sentences compared to the audio-describers. According to Al-Khalafat and Haider (136), “Type/token ratio indicates the variety of the text’s vocabulary; i.e., the highest the TTR, the more varied the text’s vocabulary.”

Frequency Analysis

The two main CL techniques used in this study are frequency and concordance. The researchers first used WS6 to generate a list of all words in the corpus with their frequencies. Table 4 shows the most frequent 20 words in the corpus.

As Table 4 shows, the most frequent words in the corpus include content words (personal names) such as مريم Mariam (mar̊yam) and ليان (līān) Layan and function words such as من (min̊) from and الى( ại̹la )to. This study focuses on content words, so an Arabic stop list that includes some words such as من from, الى to, and الذي/التي/الذين/اللواتي who is used to filter out function words. However, these words are sometimes distracting, especially when focusing on content words.

Similarly, to get accurate results and statistics, the researchers created a lemma list, which is a list of the headword with the derivatives, and used it to create a frequency list of the AD corpus under investigation, as shown in Table 5. Using the lemma list is important because Arabic is morphologically rich, as the same word may have different derivatives based on number (singular, dual, or plural), gender (feminine or masculine), and definiteness (indefinite or definite) (Haider). Additionally, prepositions and articles in Arabic are sometimes attached to words.

The researchers examined the frequency list and classified the most frequent words into thematic categories. As shown in Table 5, characters’ names were the most frequent in the corpus, with مريم (mar̊yam) Mariam being at the top of the list with 207 times. Table 5 also shows that some verbs were frequently used, such as تقترب (taq̊tarib) approaches (50 times), تبتسم (tab̊tasim) smiles (40 times), and تدخل (tadkhul) enters (34 times), to mention a few. The next section discusses the categorization of the main types of information covered in the AD of AlRawabi School for Girls series.

Thematic Categorization

In this section, the researchers analyze the thematic categories derived from the most frequent content words in the AD corpus. In this part, the researchers use the CL technique of concordance (KWIC), which shows the node word or the search term in context. Concordance investigates a particular linguistic item in its context by considering the surrounding words that might range from one word to the left or right of that item to the whole text if needed. Table 6 shows how the first scene of the series is AD.

Table 6 shows the first scene of AlRawabi School for Girls, which is taken from the first episode entitled “School was my Happy Place.” There are 18 utterances in total, 11 delivered by the audio describer and the remaining 7 by the characters. In utterances 1 and 2, the audio describer concisely gives a detailed account of what is presented on the screen, even before the actual beginning of the series, where the Netflix logo is described. Such an outset can appear promising, as minor details are described from the very beginning. The number of utterances in the AD is higher than that of the script. This could be because the first scene, in specific, can create a good first impression for the visually impaired audience, who are the main target audience of AD. In the series, such detailed descriptions might also be challenging due to spatial and temporal constraints. Therefore, audio describers are always advised not to pay that much attention to the events that do not contribute much to the series plot. As for the dialogue, according to Netflix AD guidelines, sound effects, music, and intentional silence are necessary items and should be described.

Examining the frequency list of the AD corpus, six main thematic categories are found. These include description of characters, description of actions, interpersonal interactions, description of settings, emotional states, and on-screen texts. This categorization would be helpful for audio describers, AVT students, and trainers as it shows the type of information to be included in the AD. Table 7 shows the six main thematic categories, their sub-categories, and a short description of each theme.

The key information included in the ADs are names, locations, context, details, and descriptors. These aspects often answer the who, what, when, where, and how questions (Braun). The following subsections discuss in detail the categories outlined in Table 7.

Code-Switching

Code-switching is defined as the practice of altering between two or more languages or varieties of language (Nilep). Similarly, code-mixing refers to “all cases where lexical items and grammatical features from two languages appear in one sentence” (Muysken 1). In this study, the researchers spotted seven examples of code-switching in the AD version and many others in the original dialogue, as shown in Table 18.

Table 18 shows seven examples of code-switching. Jordan is among the countries that use English frequently because it is taught in schools and universities. Some English loans are frequently used in Arabic conversation instead of their Arabic equivalents, such as iPad, tablet, and mask. This phenomenon is more common among formally educated people, such as academics, as they have the linguistic proficiency and sociolinguistic awareness necessary to move smoothly between these two languages. Most of the English words used in the AD are familiar to the receivers, and some of them, such as mobile, iPad, tablet, mask, and selfie, are sometimes used even more frequently than their Arabic counterparts.

It is crucial to remember that code-mixing, code-switching, and code-shifting are all common in bilingual and multilingual cultures, so when speakers employ two or more languages in their communication, these linguistic phenomena take place. Such phenomena should not be interpreted as a lack of linguistic ability or aptitude, though. They may occur when speakers include phrases or words from one language into their speech in another language, as in examples 1, 2, and 3. Another strategy is using loanwords, where speakers may utilize loanwords from English, as in examples 4, 5, 6, and 7. Code-switching may occur in the middle of a sentence, where speakers start with Arabic, then switch to English, and then use Arabic again, as examples 4, 6, and 7 show.

These phenomena may also occur when speakers use different Arabic varieties in their communication. Arabic has many dialects spoken across different regions and countries. For instance, some of the major Arabic varieties, including MSA, are Gulf Arabic, Levantine Arabic, Egyptian Arabic, Maghrebi Arabic, and Sudanese Arabic. In the investigated corpus, the audio describer mainly chose to use MSA and sometimes Jordanian vernacular. Again, they are used by speakers to accommodate different linguistic situations and to express themselves more effectively.

The use of code-switching in the AD of this series would also be helpful in creating a similar context to that of the original script, as the characters in the series are students in an international private school and belong to high social classes, so they maintain the use of English all the time.