1 Introduction: background

In the past years, the scope of media accessibility (MA) has widened considerably. It now includes a varied range of emerging artefacts that aim to increase an audiovisual product’s accessibility for a variety of audiences with diverse abilities. As Greco [12] argues, such shifts have manifested in the wake of a theoretical and social revolution sparked by accessibility more generally, a revolution that was made possible by (radical) changes in human rights and in information and communication technologies. Two trends in MA illustrate the above changes and impact current access provisions and audio description (AD) in particular.

First, accessibility audiences are diversifying, blurring pre-defined barriers between user groups. Accessibility scholars have long advocated for the repurposing of accessibility services for users outside the principal target group (generally, people with sensory impairments). Subtitles for the deaf and hard of hearing, for instance, are often used by the general population in noisy surroundings or in places where screens are shared [2]. In the case of AD—the verbal rendering of salient visual elements of audiovisual products such as film, television, theatre, opera and museums—the target users are often visually impaired patrons (VIPs). Other groups such as foreign language learners, however, can also benefit from AD, as put forward by Navarrete [24] and Walczak [40]. The elderly [18] and people with autism spectrum disorder [34] have also been found to benefit from AD. A second diversification trend concerns the barriers being broken down between text types. An example is the adoption of easy-to-read principles in subtitles and AD [1, 2].

Such changes have brought about new ways of thinking about translation and the position of MA within translation studies (TS). As Greco [12] states, we are “stepping outside the safe, familiar boundaries of TS and treading upon broader ground, where translation is just one of the many factors at play, and often not the main one”. This observation resonates with Neves’ [25] concept of “achievement spaces”, which she defines as the ideal conditions in which communication can happen, and in which audiences can receive information or experience emotions based on their own abilities. These achievement spaces cater differently for persons with diverse abilities, and each person’s achievement space is unique. An accessible experience created through such achievement spaces is not limited to the provision of traditional translation and access services but focuses on creating equitable conditions for experiencing meaning in a given situation through a variety of texts and stimuli. The goal of achievement spaces is “effective communication, and that may require that (envisaged or existing) barriers be overcome through alternative or complementary communication strategies that promote multisensory engagement enabling perception, understanding and experience” [25].

Similarly, but applied more concretely to AD, is Roofthooft’s [32] concept of the accessibility “cluster”. In her PhD dissertation on theatre AD, Roofthooft [32] argues that theatres could offer VIP users a “cluster” of access services, which is not limited to the traditional audio introduction (AI)Footnote 1 that precedes a described performance, but includes, for instance, pre-recorded AIs that have a different focus than the traditional AI, or elaborate descriptions of performance photography to introduce audiences to the intricacies of the visual design of the piece. This way, audiences are free to choose which access options fit their needs best.

Fryer and Cavallo [10] have raised a similar issue, emphasising the need to make all marketing materials surrounding a performance more accessible: “In order to reach audiences with disabilities you need to recognise barriers to advertising materials in terms of sound, text or image” [10].

In short, in MA and AD alike, we can observe both the emergence of new services as well as the merging of traditional access services for people with diverse abilities with audiovisual creations aimed at mainstream audiences, breaking down barriers between disciplines, between processes, products and audiences. Numerous examples from practice illustrate these shifts in the AD environment, from integrated access [8, 10] to creative approaches to accessibility [33].

Another example, and the focus of the present article, are emerging trends in theatre accessibility and AIs. Traditionally, AIs for accessible theatre and opera performances are either voiced live as described above [11], or they are sent out to AD users via CD, particularly in the UK [9]. Increasingly, thanks to developments in digitalisation, technology and web accessibility, AIs are made available in pre-recorded formats on popular audio-hosting platforms such as SoundCloud and Spotify.Footnote 2

This trend in the domain of AIs for AD users is growing in parallel with developments in theatre podcasting. Cultural venues are investing more in the development of audiovisual content surrounding their performances. Partly under the influence of COVID-19 lockdowns which have affected the cultural sector worldwide (see, for instance, Liodaki and Velegrakis [21]), several cultural venues have started to offer audio content in the form of theatre podcasts, a format that has gained popularity in all sectors the past decade [26].

Research into such trends remains very limited. The study of AIs from within the field of MA has focused mostly on live AD and on the content and functions of such introductions for different types of audiovisual products (theatre, opera, film or television) [6, 11, 31]. Research into theatre podcast from the fields of theatre and media studies also remains scarce. The potential of podcasts for accessibility purposes has not been researched extensively either, to the best of our knowledge.

Given the aforementioned context, however, interesting questions can be raised regarding the potential role of theatre podcasts within a performance’s accessibility cluster or achievement space and about the seemingly fading borders between accessibility podcasts and mainstream podcasts. Are some of these theatre podcasts created with access in mind? Are they promoted as such to AD users? Do theatre podcasts meet basic accessibility needs? What is their added value to the creation of an accessible experience? What are the differences or similarities between traditional AIs and other theatre podcasts offered at the moment?

These questions constitute the starting point for the present exploratory study, investigating the relationship between theatre podcasts and AIs, aiming to elicit the potential of these text types as a future access tool. The objectives of the paper are to (a) describe the types of theatre podcasts cultural venues in Europe offer and their content and context, (b) contrast the traditional functions of AIs with the functions of mainstream theatre podcasts and (c) evaluate the potential of theatre podcasts as an access service.

2 Methodology

The methodological approach in this paper departs from systemic functional linguistics (SFL), commonly applied in TS and MA. Within audiovisual translation (AVT) and MA studies, discourse analysis evidently involves multimodality research. Multimodality is fast becoming a key framework within MA research [3, 35, 40] because of the complex interplay between semiotic modes that underlie such intersemiotic translation processes. Multimodality research is rooted in Social Semiotics [19] and more particularly in Halliday’s [17] SFL. From this perspective, all semiotic modes, both verbal and non-verbal, can express the same types of meaning as language, that is experiential, interpersonal and textual meaning, referring to the tripartite division of language (meta)functions proposed by Halliday [15], all within the broader concept of social semiotics [35]. Detailed accounts of SFL as an analytical framework are provided by general works on SFL such as Halliday [16] and Eggins [7]. Its translation into multimodal translation studies is discussed in more detail by Taylor [35] and Boria et al. [3]. An application to audio description in particular was done by one of the present authors [30].

Below, we provide a brief overview of key concepts, outlining their application to the analysis of AIs and theatre podcasts in particular. Our approach puts particular emphasis on describing the wider communicative context in which the discourse is used and pays less attention to the identification of recurring linguistic patterns. This approach mirrors recent podcast research as well. For instance, Lundström and Lundström [22] describe a framework coined “podcast ethnography”, which includes three phases. First, exploring the podcast by gathering metadata (number of episodes, duration and time span) and by exploring its content: internal structure, participants and relations between the speakers and the listeners. Second, engaging with the podcast through active listening, the collection of fieldnotes and the definition of analytical themes. Finally, examining the podcast by studying “the empirical material, to develop concepts, themes or typologies that reconnect with the initial research questions” [22].

The present study consists of three distinct steps: (1) description of the communicative context of theatre podcasts in Europe, (2) meta-functional analysis of a selection of theatre podcasts and (3) analysis of the communicative intention of the podcast, through expert interviews with podcast makers.Footnote 3

2.1 Description of the communicative context of theatre podcasts in Europe

Drawing from Reviers [30], the concept of “context” is key in the discourse analytical approach. Within the framework of SFL, it comprises the context of culture and the context of situation. The context of culture acknowledges how a language as a system relates to its culture, particularly when it comes to the general belief and value system. The context of situation acknowledges how specific texts relate to their immediate and specific material and social situation [16], and it is the type of context that is of interest for our research purpose.

The context of situation is further divided into field, tenor and mode. In brief, what the text is about, who is talking to whom, and through which specific channels. In the analysis section, we aim to describe the context of situation of the theatre podcast and its use in Europe, based on these three basic dimensions, by conducting a literature reviewFootnote 4 and a structured analysis of the audiovisual content main European theatres published online in 2021 (see Sect. 3.1).

2.2 Meta-functional analysis of a selection of theatre podcasts

Field, tenor and mode are expressed by three meta-functions of language: the experiential, the interpersonal and the textual function, respectively [15]. In our second analytical step, we explore how the contexts described in step 1 translate on a discourse level in a selection of podcasts in different languages and from different contexts of situation. The aim is, on the one hand, to identify differences and similarities that emerge from the different contexts. On the other hand, we want to elicit to what extent the meta-functions expressed correspond (or not) with the functions expressed in traditional AIs.

2.2.1 Experiential function

In terms of the experiential function, we studied the types of content that the podcast includes and the communicative functions they express. Previous AD and AI scholars have identified the following types of content AIs usually contain [6, 11, 31]: information about the event, information about the piece and information about the production.

For each of the assessed podcasts, we studied what type of information was included and how extensively it was discussed. Given the multimodal and multisemiotic nature of theatre, we also indicated whether the content explicitly included information about the performances’ visual modes or not.

In addition, scholars have put forward a series of communicative functions that AIs can express and that occur in different levels of relative weight [29, 31]: the informative function, the narrative function, the explanatory function for complex theatrical illusions, the expressive-aesthetic function for the aesthetic effect of theatrical techniques, the foreshadowing function to anticipate visual information that cannot be included in the AD itself and the instructive function.

For each of the assessed podcasts, we studied what functions were included, what their relative weight was (from very prevalent to prevalent, occasional, rare and absent), and how this function was expressed linguistically. This is discussed in Sect. 3.2.

2.2.2 Interpersonal function

In SFL, the interpersonal function includes the expression of mood and the conversational structure of the text. The elements of the interpersonal function are clearest in dialogue, the focus of most research on interpersonal meaning, and may be more difficult to grasp in other kinds of discourse [36] such as AD. In addition, the previous research on AD and AI has not focused on this level as extensively as on the content and function of AIs. The choices in terms of the interpersonal function are said to be inspired by the contextual dimension of tenor, i.e. the role of the participants in the communication. Thus, in our analysis, we focused on identifying the types of speakers and makers involved in the text and their roles, and the ways in which their perspective was implicitly or explicitly expressed linguistically in the podcasts. The following main questions guided our analysis: Who is the target audience? Who is the speaker? How is the relation between audience and speaker (expressed)? This is discussed in Sect. 3.3.

2.2.3 Textual function

The textual function in SFL concerns the way the text is thematically and cohesively structured. The previous research on AIs has not focused extensively on this metalevel, even though some studies have revealed a certain recurring textual structure in AIs and especially ADs. Mazur [23] proposes a functionalist model which delves into both the contextual, macro-textual (related to the content, the form and both) and the micro-textual analysis of the source text in the scope of AD. Remael and Reviers [28] adapted Mazur’s [23] model to theatre AD and AI and proposed three categories of analysis: semiotics of theatre, the role of the different stakeholders and the AD and AI delivery methods. For the purpose of this study, we focused on the analysis of the macro-textual features. On the one hand, we studied the order of information (in terms of content and function). On the other hand, we studied the textual organisation from a multimodal perspective, assessing the use of non-verbal elements such as sound and music. A particular point of attention in this respect was that the podcast in some cases functioned as a meta-text, introducing audiences to a specific performance. In that sense, the podcasts potentially interacted intertextually and multimodally with elements of the performance.

The following questions guided our textual analysis: What is the main rhetorical mode of the text? How is the text structured? How is the multimodal configuration of the text? The results are discussed in Sect. 3.3, and Appendix 1 includes the extended list of questions that guided our analysis for each of the three meta-functions.

2.3 Expert interview with podcast makers

While the text analytical approach described above reveals a lot about the overall communicative function of the podcasts, it has no explanatory power and cannot fully elicit the communicative intent of the podcast makers nor the communicative effect on podcast users. For this reason, we aimed to include a semi-structured expert interview with the podcast makers in order to obtain insight into the decision-making process of the podcast and extend our analysis in Sects. 3.1, 3.2 and 3.3 with such insights. Due to practical and time limitations, the current analysis includes one expert interview. Consult Annex 1 for a detailed overview of the interview questions.

The analysis of the communicative effect of the podcasts on users is beyond the scope of the present article, although we recognise the pertinence of a user-centred study as a follow-up to the current one.

3 Analysis and discussion

3.1 The audiovisual ecosystem in European theatres

The act of sharing different types of content through video and audio platforms has broadened the meaning-making possibilities of theatre venues. Hadley [14] describes several ways in which venues achieve this: (1) the use of social media for the creation of intermedial performances, (2) the distribution of the performance, (3) audience development—particularly in certain demographics such as younger audiences [37], (4) the task of criticism and commentary, (5) documentation and archiving and (6) the performative element inherent to the use of social media. In addition to these six, we may add a seventh possibility specific to our purpose, which is the inclusion and broader promotion of accessible materials. Podcasts in particular also become a medium of participation, according to Lundström and Lundström [22], either because the audience is directly mentioned or some of its members are invited as guests. This element is explored further in our analysis of the interpersonal function [15] of the theatre podcasts and in the interview with the creators (see Sect. 3.3).

While theatre streaming had been on the rise for several years,Footnote 5 it has experienced an extraordinary boom in the past 2 years as a result of the COVID-19 pandemic. Theatres “released online programmes that included broadcasts of past popular productions, commissions of new work made during lockdown, and open resources that showcased the work going on behind the stage” [4]. In this context, the present article aims to interrogate the audiovisual ecosystem that has been developed by theatres in response to the pandemic and beyond, and how this ecosystem interacts with or may add to the accessibility services offered by theatres.

The next part of this section aims to describe the audiovisual ecosystem of some of the main theatres in Europe. To narrow down the selection of venues to be studied, the two largest theatres and the two largest opera houses were selected for each of the regions assessed in this study, namely Flanders (Belgium), France, Germany, Spain and the UK. For this exploratory study, a pragmatic approach was taken for the selection criteria, which consisted of seat capacity and social media following. This procedure led to the following selection: Toneelhuis and NTGent in Flanders, Belgium; Teatre Nacional de Catalunya and Teatro Español in Spain; Royal Albert Hall and National Theatre in the UK; Schaubühne Berlin and Deutsches Schauspielhaus in Germany and Comédie-Française and Odéon-Théâtre in France.

We first aimed to identify trends in content creation in the audiovisual sphere of the assessed theatres. It became apparent that all of the above-mentioned venues curate their own YouTube channels, with the exception of the Odéon-Théâtre in Paris. Additionally, the Schaubühne theatre in Berlin also includes materials on the platform Vimeo in their “Mediathek” website section. On YouTube, the typical materials that can be found across theatres are season and performance trailers, interviews and colloquiums with creators and theatre staff, as well as press conferences. Some theatres include didactic materials as well, such as the Royal Albert Hall and the National Theatre in London, the Teatre Nacional de Catalunya in Barcelona and the Comédie-Française in Paris. Other types of content are only offered by specific theatre venues, such as fully recorded theatre plays (the Teatre Nacional de Catalunya has even developed its own streaming platform: TNCdigital), book readings, rehearsal clips and behind-the-scenes materials, lockdown specials and even an “inclusive discovery tour” at the Deutsches Schauspielhaus, with subtitles and AD.

In terms of audio-based materials, over half of the assessed theatres produce podcasts hosted on platforms such as SoundCloud (Deutsches Schauspielhaus, NTGent, Schaubühne Berlin), Spotify (Teatre Nacional de Catalunya) and YouTube (Comédie-Française) or a combination thereof (Toneelhuis, see Table 1). Unlike most of the video-based materials of the theatres, which largely overlap in terms of content, podcasts are more diverse. To give one example, the Schaubühne theatre hosts four different podcasts: on politics (Streit ums Politische), on social issues (Streit Raum), on book launches (Buchpremiere) and on the interactions between language and theatre (Alles außer Hochdeutsch. Dialekt und Dialektik). Interestingly for our purpose, the SoundCloud channel of NTGent features both their podcast and their AIs—intended as accessibility services—all on the same page.

Table 1 Context of situation of the podcasts

As for opera houses, the selection for this study was narrowed down to the Teatro Real and the Liceu opera house in Spain, the Royal Opera House (ROH) and the Glyndebourne opera in the UK, the Staatsoper Berlin and the Bayerische Staatsoper in Germany, the Ópera de Paris and the Opéra Comique in France and the Belgian opera houses Opera Ballet Vlaanderen and De Munt. Much like the theatres examined above, major opera houses in Europe are very active in terms of audiovisual content sharing. All of the opera houses publish season and performance teasers, interviews with artists, press conferences, clips from well-known arias and rehearsals. Some of them also promote didactic programmes (Teatro Real, ROH, Opéra Comique), while others upload making-ofs featuring staff from technical departments (Glyndebourne opera, Staatsoper Berlin (unter der Linden), Bayerische Staatsoper).

As per our analysis, opera houses are more invested in the curation of their “audiosphere” when compared to the theatres, at least from a quantitative standpoint. This may be due to various context-based reasons, including budget and programming differences. All of the assessed opera houses except for Glyndebourne opera host at least one podcast.

Considering both theatre and opera houses, we observed that introductions to the plays are the most common subtype of podcasts among the assessed opera houses. They mainly adopt two formats: monologic (Liceu, ROH, Staatsoper Berlin (unter der Linden), Teatro Real) or dialogic (How to Oper for the Bayerische Staatsoper). Though most of these podcast introductions are targeted either at the general public—as in the Liceu’s La Previa, Staatsoper Berlin (unter der Linden)—or at patrons with low vision, such as the ROH, we may argue that their content overlaps to a great extent. In fact, the Belgian theatre house Toneelhuis explicitly targets both the general audience (in the first part of the podcast) and VIPs (in the second part of the podcast) in one podcast.

Aside from the most common introductory podcast, however, our analysis revealed other types of formats. For instance, the Pulsations podcast from the Ópera de Paris features creators conversing about the emotions that they experience in opera and ballet. Opera Ballet Vlaanderen features a podcast called De stemmen (The Voices), which is not related to any of the venue’s performances but gathers personal life stories centring around a central operatic theme.

Based on the above analysis, we have developed the following preliminary typology of what we call “podcast audio introductions” (PAI): On one side of the spectrum, we place theatre and opera podcasts that are not devoted to introducing a given performance or piece; they function independently and are less fixed in form and content. This is the case of the Pulsions podcast from the Ópera de Paris. These types of podcasts are not included in our analysis in the next section, where we focus on the podcast types that serve as an introduction to a given performance. In the middle of the spectrum, we can find podcasts intended for the general public: mainstream podcasts which centre around one performance or piece and are more fixed in terms of form and content. Further along on the spectrum, we find hybrid podcasts, which are designed to cater for both general audiences and VIP audiences. At the other end of the spectrum, we position traditional AIs, which are intended (exclusively) for accessibility purposes and are the most fixed in terms of expected content and structure. Depending on their typology, PAIs can be found in different areas of the theatre’s website: They are either presented online as part of a larger “cluster” of communication materials (such as the leaflet, images from the production, musical excerpts, etc.), or they can be found in the accessibility section of the website only. In some cases, podcasts are distributed through several channels, both on the general website as well as through communication channels targeting only VIP audiences.

The analysis of the audiovisual cluster and audiosphere of European theatres led us to the conclusion that (1) large theatres are increasingly invested in the materials they produce to be consumed outside the physical venue, among other reasons as a response to the COVID-19 pandemic, as confirmed by the creators of Radio Toneelhuis in the expert interview, and (2) podcasts and AIs are presented together by some, while others clearly differentiate their purpose placing them in different sections of the website. As discussed below, podcasts and AIs share some features and notably key differences in terms of content, communicative and textual functions. Ultimately, they can gain from each other and even merge, as is the case in Toneelhuis. The next section delves deeper into the form and content of the different types of PAIs.

3.2 Systemic functional linguistic analysis

In this section, four podcast episodes are analysed by applying the principles described in the methodology section: La Previa de Rigoletto from the Liceu opera house in Barcelona, Zeg aan de kinderen dat we niet deugen [Tell the Kids We Are no Good] from Toneelhuis in Antwerp, Audiofeature—Die Nase [The Nose] from the Bayerische Staatsoper in Munich and Rigoletto 2022 Audio Intro from the ROH in London. Table 1 offers an overview of the context of situation of each podcast.

3.2.1 Experiential function

In terms of the experiential function, it is relevant to draw a distinction between the information regarding design modes—score and libretto, the creators’ vision, etc.—and production media—in opera, singer’s voices, orchestra sounds, gestures, stage action and sets, etc. [19]. That is, the piece itself (1), as designed by its creators vs the specific production that is staged (2) and the event per se (3), as shown in Sect. 2.2. A trend among the mainstream opera podcasts under study (La Previa and Audiofeature) is that the focus is mostly on the piece (1), with some information about the production (2) and none on the event (3). The hybrid podcast from Toneelhuis, which explicitly targets both sighted and non-sighted listeners, focuses largely on information about the piece (1) and about the production (2), but also includes limited information about the event (3) (indicating that the performance will be live described). In the Toneelhuis podcast, the production information takes up about 1/3 of the podcast and coincides entirely with the segment that is explicitly tailored to VIPs. The makers of this podcast acknowledged that they included this targeted section to cater for the needs of VIPs (production and visual design). In addition, Toneelhuis decided to explicitly mention which section was targeted at VIPs because they anticipated that this information might not be relevant for everyone. The section on the production begins as followsFootnote 6:

What follows is a visual description of Zeg aan de kinderen dat wij niet deugen. This description is intended especially for the blind and visually impaired as a preparation for the visual aspects of the performance. Rather than giving a very detailed description of what Guy Cassiers and his team put on stage in visual terms, we want to introduce you to the various technical means he uses (example from Toneelhuis).

The traditional AI from the ROH—the only assessed podcast exclusively targeting VIPs—differs from the previous ones in that most of the information presented is about the production (2): its scenography, lighting, costumes and character identification; there is some information about the event (3) (“If you wish to continue with the AI, please return to your seats at the 10 min call”), but minimal time is devoted to the piece (1), i.e. the historical and musicological context of the play. Figure 1 visually summarises the approximate time devoted to each type of information in the PAIs.

Fig. 1
figure 1

Time devoted to each type of information in the PAIs

Regarding the communicative functions [29, 31], their distribution in the Audiofeature and La Previa podcasts is very similar, in that the informative function prevails and the narrative function comes second, with an extensive disclosure of the synopsis and the themes of the performances. Examples of the informative (1) and the narrative function (2) are illustrated below:

Rigoletto (unlike other operas from Verdi) was not an instant success. It gained appreciation in subsequent representations, because of its virtues: a well bound dramatic structure, the refined chant typical of the Italian tradition… (example from La Previa).

Had I lost an arm or a leg, it would not have been so bad; had I lost my ears, it would have been bad enough but nevertheless bearable, but without a nose a man is goodness knows what (example from Audiofeature).

In the hybrid Toneelhuis podcast, on the other hand, the informative function is not as prevalent, and the narrative function is completely missing, as the performance was based on a recital of poetry. In the ROH AI, the informative and narrative functions are rare. Instead, the traditional and hybrid AI focus on the foreshadowing function, disclosing the visual elements of the performance:

He initially wears a black velvet jacket with heavy gold embroidery and his wife wears a white dress (example from ROH).

From the moment the actor starts speaking, the light slowly changes (example from Toneelhuis).

On the contrary, the foreshadowing function is residual in Audiofeature and La Previa, corresponding with their focus on the piece instead of the production.

The podcast from Toneelhuis differs from the mainstream podcasts in that the foreshadowing function and the explanatory functions are more prevalent. The foreshadowing function clearly dominates in the second part, which targets VIPs explicitly, as it focuses on describing the scenography. Audiofeature, La Previa and Toneelhuis all introduce the explanatory function, though their explanations are not related to complex theatrical illusions (as is the function in traditional AIs). Instead, they allude to the intricate themes of the plays:

The poor jester arrives at the same time, and he is subject to a bad joke, as they make him an unsuspecting accomplice without him realising (example from La Previa).

In the performance Zeg aan de kinderen dat wij niet deugen, the poet not only measures his own fever, but also the fever of his generation and the fever of our time. More than the tensions, paradoxes and contradictions within the 'I' and between 'I' and 'you' as is the case in many of Nolens' poems, this selection of poems is mainly about the 'we (example from Toneelhuis).

In the case of the ROH, the explanatory function does refer to different theatrical illusions and inferences: “It’s almost as if the tableau is lit by a flash of lightning, with deep darkness all around”. The same type of traditional explanatory functions are also found in the Toneelhuis podcast:

The words of the poet, the body and the voice of the actor are confronted with projected images, objects and music. Cassiers wants to make the spectator's sensory experience, his poetic imagination and his receptivity to Nolens' words as intense as possible (example from Toneelhuis).

Furthermore, the instructive function in its original sense—information about how to get to the venue by public transport, for instance, or how to use the headset for the AD [31]—is only present, albeit residually, in the ROH AI and in the parts of the Toneelhuis podcast that is intended for the VIP (explaining them how to use the AI in relation to the AD that is provided during the performance). Audiofeature and La Previa further introduce two functions that have not been formally included in traditional AI thus far: (1) a didactic or pedagogical function with the aim of disseminating musicological, historical and literary knowledge and (2) a marketing function, especially considering their multimodal context on the website and social media accounts of the theatres. Interestingly, in the light of the “cluster” notion, Toneelhuis actively instructs audiences to check other materials on the site, though not explicitly with a promotional intent, since the podcasts mostly target ticket holders.

Finally, an aesthetic function—meaning the adoption of a literary style—is prevalent in the mainstream and the hybrid podcasts. In the Toneelhuis podcast specifically, there is a contrast between the first part and the second part (focusing on the production). This is not by accident. Toneelhuis makers stated that, when writing the section for the VIPs, they are conscious about being concrete and to the point, as the section serves a different function (accessibility) and audience (VIPs). It requires fewer "frills", as they said. Conversely, they are intentionally more poetic in the first part of the podcast to meet audience expectations.

To conclude, the communicative functions present in each podcast are placed on a scale ranging from very prevalent, prevalent, occasional, rare and absent in Fig. 2. The explanatory function in Audiofeature and La Previa refers to the disclosure of the intricate themes, not the clarification of theatrical illusions.

Fig. 2
figure 2

Communicative functions in each PAI

The different contents, functions, audiences and creators detailed in the analysis of the experiential function make for an interesting case for the theatre’s growing access cluster [32] and achievement spaces [25]. By focusing on the piece instead of the production or the event, mainstream podcasts shift away from the ocular-centric logic that often permeates AIs and ADs [5]. The hybrid podcast from Toneelhuis finds a middle ground by catering for both audiences. However, at the same time, it emphasises the division between audiences by splitting the podcast into two different sections. The foreshadowing function is, accordingly, less prevalent in the mainstream podcasts than in the hybrid podcast and the AI, and the explanatory function shifts in nature in the first three: The focus is no longer on the complex theatrical illusions and their subjective effect [31], but on the intricacies of the actual themes and their initial reception. The coexistence (and the hybridisation) of introductory podcasts and AIs further opens up new avenues of communicative function personalisation [27] and non-segregatory practices [13], both at the forefront of current research. The different types of podcasts, together with the other audiovisual materials offered by the venues, potentially contribute to Neves’ [25] notion of achievement spaces.

3.2.2 Interpersonal function

In terms of the interpersonal function, an analysis of the digital platforms through which the PAIs are distributed reveals who the potential audience is. In La Previa, we can assume that the audience is an opera aficionado who (1) follows the opera’s social media (where the podcast is publicised) or (2) has visited the opera’s website to check a performance or buy tickets. From the podcast itself, it is inferred that the target audience is already a fan of opera, as is implied in the empathetic use of the third person (“There are, of course, many opera aficionados who look for unsolvable moral conflicts, stories that deeply challenge our worldview”), particularly when combined when the first person in plural. This can also be inferred from the website presentation of the podcast: “This new podcast service replaces the informative session before the performance […]”. Similarly, Audiofeature listeners are understood to be opera fans/social media followers of the Bayerische Staatsoper, or members of the public that are interested in a specific performance. The tone towards the audience in Audiofeature is, however, more impersonal (there are no instances of the second person or the first person in plural), but this “impersonality” is breached through the use of literary language. Some examples are rhetorical questions, which are interwoven with texts in the third person/impersonal forms: “Was it all a dream?” “He was seen as some kind of surrealist pioneer”.

Conversely, the target audience for the ROH AI is primarily VIPs, as this material is only linked from the “Accessibility” page of the opera’s website (though, admittedly, any member of the general public can stumble upon it on SoundCloud and the number of listeners, reaching the hundreds, suggests so). The audience is addressed mostly implicitly, but there are some rare instances where they are addressed by the second person, notably when the instructive function is deployed: “If you wish to continue with the audio introduction, please return to your seats at the 10 min call”. The intended audience is directly addressed, however, in the accessibility section on ROH’s website where one can find the AI podcast:

The audio introduction is broadcast 15 minutes before the start of the performance and five minutes before the end of each interval. You can also listen before your visit on the ROH SoundCloud, where you can set alerts to notify you of new audio introductions for upcoming performances.

As we might expect from a traditional AI, the audience is only directly addressed at those brief instructive moments; otherwise, the third person and impersonal structures are the norm.

The podcast from Toneelhuis is perhaps the more novel one in this regard, as it targets two types of audiences: one type that is addressed implicitly and one that is identified explicitly. The first type of audience is general theatregoers, addressed by the podcast mostly through the third person (“the spectator”), but also by the first person in plural. The second type of audience is VIPs. The introduction to the podcast already mentions a section targeted at them: a visual description. What is very noticeable is the break in style between the largely indirect way general audiences are addressed in the first part of the podcast and the way audiences are consistently addressed in the second part. In the latter, audiences are more often addressed with the polite second person pronoun. This stylistic choice and the choice to explicitly name only the second audience type creates a clear division between audiences and their social status/roles. This “targeted” practice could be problematised, as it reiterates a “ghetto effect” [13]. The makers of the Toneelhuis podcast proved to be aware of this divide and acknowledged that the podcast is part of an experimental trajectory in which they try out several tactics to increase inclusiveness. They reported finding it challenging to strike a balance between the needs of the two different types of audiences (particularly in the descriptions for VIPs). In addition, the podcast was written by two different authors: The main dramaturg wrote and voiced the first mainstream part, and a second communications collaborator with experience in accessibility wrote the second part, with an intentional focus on not placing art over access.

In terms of the speakers, the mainstream podcasts are once again similar in that the script writers, producers and voice talents are identified, either in the podcast itself (La Previa) or in the information section of the audio streaming platforms (Audiofeature). In La Previa, the narrator that features consistently throughout the podcast is a female voice talent. Both she and the scriptwriter—who briefly intervenes for a short section providing some information on the initial uproar following the play’s original premiere—take on an institutional role. The speakers express value judgments in many different ways, showcasing their expert function, for instance, through expressions including axiological components [20]:

Rigoletto is a total loser. Unlike the protagonists of the operas of his time, he’s an anti-hero”; “[F]rom the beginning we suspect that it won’t be enough and that he is a doomed character (interpretations in the first person highlighted in italics, example from La Previa).

In Audiofeature, some of the speakers intervene directly (the scriptwriter is also one of the voice talents, and the stage director of the production also intervenes in the podcast as “himself”). The remaining female voice talent delivers the rest of the script. Their expert role can also be demonstrated by the speakers’ many value judgements (“The fact that Schostakowitsch chose the classic The Nose, written by Nikolai Gogol in 1835, for his ‘opera of the future’ was no accident”), and the amount of factual knowledge they provide in the podcast (“Schostakowitsch knew what he wanted: to compose new, wild music, also for opera”). Finally, in terms of direct interaction, the audience may leave comments on the social media posts promoting the podcasts, but there is no more interaction beyond that.

Conversely, in the podcast from Toneelhuis, the authors and speakers are not clearly identified by name or function. The website only mentions the institutional author (Radio Toneelhuis). The speakers can only be identified by their institutional role as “people/artists working for the institution”, which can be inferred by the content and language use in the podcast itself. This illustrates the relative status between speaker and listener as that of an expert talking to a non-expert. Sometimes “we” is used, putting speaker and listener on the same level. This is mainly done when talking from the perspective of how the director wants audiences to perceive something: “we as audiences”. At other times, “we” refers to “we makers of Toneelhuis”. Overall, three people intervene in the podcast, and this is the most notable resemblance with La Previa and Audiofeature: The main actor recites a poem, a female voice introduces and ends the podcast (instructive function), and a male voice presents all the other (more expert) information. According to the makers, having the actor recite a poem was a deliberate access strategy to make sure that the VIPs were acquainted with his voice, even if the recital was included in the general audience section. While this was inspired by access needs, it turned out to be an aesthetic addition catering to all audiences, which can be seen as an example of “blindness gain” [5]. The interview with the makers from Toneelhuis showed that not revealing the creators' and voice talents’ identities was not a deliberate strategy, they simply had not considered it. The choice of the voice is, however, deliberate: It is the voice of Toneelhuis' main dramaturg who is closely involved with most performances at Toneelhuis and knows them well. He is an expert. He is partly responsible for writing the podcast (excluding the section intended for VIPs, undertaken by a communications collaborator who followed AD training as well). The interview also highlighted that Toneelhuis learnt from experience that the best results are achieved when the author also voices the podcast. This was, however, not possible for the section targeting VIPs only, which was voiced by the dramaturg as well. This observation shows, in any case, that the choice of author and voice and their expert knowledge of their target audience is a crucial issue in PAI creation. Training and awareness building regarding both accessibility as well as theatre studies for all people involved in podcast creation is a crucial point to consider.

As for the ROH, though we do know—from the accessibility webpage where the AI is linked—that the company producing the AI is Sightlines,Footnote 7 the name of the audio describer (or female voice talent) is not disclosed, and their literary/musicological/historical expertise does not permeate the text. Indeed, some degree of objectivity is fostered. One exception occurs when the ROH AI mentions the name of an art piece used at the back of the set: “It’s a recreation of Titian’s Venus of Urbino, and features a naked woman reclining on a bed, completely uncovered and turning towards us”. There is another exception when it comes to the constant “distance” between the audio describer and the audience in the sporadic use of the pronoun “we”. One way this podcast counterbalances the distance between the audio describer and the audience is through the use of rhetorical figures such as simile and metaphor and some sporadic direct quotes from the libretto.

When comparing the four podcasts, and following Halliday and Matthiessen’s [17] terminology for the interpersonal function in mainstream podcasts, we can summarise that: (1) The institutional role lies heavily with the producing theatre. The (2) status role is somewhat unbalanced, as the podcast makers represent the expert input, and there is no sign of power-sharing. The (3) contact—familiarity, from strangers to intimates—and (4) sociometric affect, positively and negatively charged—roles are diametrically different from what we may expect from a traditional AD or AI [30]: The distance between the podcasters and the audience is narrowed through the first-person language, overt subjectivity and charged affect. Indeed, the podcast does not pretend to be objective, nor is it implied that it is an access service: Its stated that aim is to “complement and enrich” the play. This is opposed to the general expectations of traditional AI and AD, which “predominantly use third person pronouns, as they reflect the voice of an omniscient narrator” [29], and the fact that traditional AD and AI “urges describers to adopt a neutral approach to word choice and delivery” [8].

To conclude, within the interpersonal function, the podcast format as compared to traditional AI introduces more and different speakers and voices, with authorship acquiring a more central role. As experts introduce their input, the role of the neutral mediator/translator disappears or at least shifts substantially, as is suggested in proposals such as director-centric AD [38] and integrated AD [8, 10]. The danger in this regard is that accessibility needs are to some degree pushed to the background in mainstream podcasts. In other words, it is fair to question how mainstream podcasts actually reflect the needs of VIPs and other audiences. The same concerns were echoed by the Toneelhuis podcast makers. While the analysis shows the potential of opening up new types of information to VIPs that were previously absent from AI and AD, mainstream podcasts cannot fully replace the AI and would arguably only work as part of an “access cluster” in combination with other access services. In Toneelhuis, for instance, the hybrid podcast does not replace the traditional live AI that is offered along with the AD. It is a conscious addition to the traditional offer to enrich and broaden users’ experience.

3.2.3 Textual function

In terms of the textual function, our focus lays on the podcasts macro-structure. The order of the information presented in La Previa goes as follows: Rigoletto’s general themes, synopsis, historical and musicological information, original premiere and information on current production (cast and brief overview of the scenography). Similarly, Audiofeature scatters the contents of the plot in between the introduction to Die Nase’s themes, musicological information about the composer, information about the original author of the novella and main takeaways from the story. To put it simply, the structure in which the information is presented is less fixed. This contrasts with the clear “dual” structure of the ROH’s AI: First, there are visual descriptions of the stage, the lighting and the characters (their wardrobe, appearance and makeup). Then, each scene of the first act is explained, mostly in visual terms, but also with some plot snippets. The second part of the AI focuses mostly on the combination of visual and plot description. Lastly, in Toneelhuis, the order of information is very much guided by the dual audience that is addressed: Part one is about the piece and the themes and content of the performance, while part two is about the visual scenography. Nonetheless, the makers of the podcast did remark that the content, style and order/structure of the information within each section is never fixed, as it is tailored to each piece. For other pieces, the structure sometimes follows the narrative or focuses more on the visuals than is the case here. In short, they adapt the PAI to the style and logic of the piece itself.

Furthermore, all of the podcasts are monologic, though with some nuances. The ROH’s AI is strictly monologic. In the podcasts per se, there are more resemblances with a storytelling session, for instance, in Audiofeature, the voice talents and creators of the performance take turns to discuss the plot/themes/context in a more artistic way. In between this main “storyline”, the stage director adds his own impressions towards the themes. The interaction is not real, it is “enacted” in the post-production stage.

In terms of the multimodal configuration of the podcasts per se (excluding ROH’s AI), it is worth considering their co-occurrence with other semiotic information in the performance website: from the podcast to a downloadable programme, information on the cast, audiovisual materials (video clips and interviews), images from the production and an easy-to-read programme in the case of the Liceu opera house. In all of the studied podcasts, the focus is evidently on the audio-verbal semiotic mode. There is music, as well as some soundbites, in the mainstream podcasts (La Previa and Audiofeature) and in the hybrid podcast from Toneelhuis, but not in the ROH’s AI. The music is at times in the background (with an aesthetic purpose or in order to create an atmosphere) and other times functional (to indicate division between paragraphs, to present a particular well-known aria, etc.) at the beginning/end. In some instances, the music can also adhere to a foreshadowing function [31] as it reproduces music or voices from the performance, linked to certain developments of the plot.

To conclude, in terms of the textual function, the mainstream podcasts (and the hybrid podcast from Toneelhuis) prove to be richer in terms of the inclusion of music, sound effects as well as more than one voice to complement the verbal component. Within this function, there is room for cross-fertilisation between the different approaches. One direct example could be the inclusion of background music and excerpts from interviews with the creators or from the play itself [13]. In the scheme of the meaning-making possibilities, AIs could gain from the non-verbal practices from the podcast, both in functional and aesthetic terms.

4 Concluding remarks

The proliferation of audiovisual text and audio materials available on theatres’ websites had not yet been targeted from an MA approach. This article has opened up a novel research approach to podcasts as an (explicit or unintentional) avenue for accessibility research, particularly when it comes to the “introductory” genre to theatre and opera plays. The mainstream podcasts can be seen in different ways in the experiential, the interpersonal and the textual functions, as the opposite parameter of AI. The argument presented in this paper is that the coexistence of traditional AI and podcasts introducing the piece (vs the production of a given play), as well as the merging of the two, i.e. addressing the needs of both VIPs and the general audience, could be a step forward in terms of personalisation in accessibility [27].

The analysis in this paper has highlighted a clear trend towards more inclusive approaches, but these are still varied and experimental. It raises the question of what an inclusive approach entails or should entail. There seems to be a continuum emerging in inclusive approaches: At one end of the spectrum, there is an inclusive achievement space or cluster [25, 32], consisting of a collection of texts that each has a different focus in terms of content, function and audience. On the other end of the spectrum, there is inclusive design, i.e. one product that tries to cater for all. Both have benefits and pitfalls. For instance, the cluster notion presupposes a great deal of user autonomy (which could be a problem in terms of the digital divide, age, etc.). It could risk losing sight of particular access needs. It risks putting "art over access" [8] and could cause information overload.

For instance, the proposal of a hybrid genre enjoyable for all, as in the case of Toneelhuis, is not without its risks, just like integrated AI [8]. From an accessibility perspective, traditional AIs focus on having all relevant information in one place (because the programme leaflet and website were traditionally not accessible). Drawing from the notion of “access clusters” [32], different—complementary—texts focus on different types of information and functions. The theatre access cluster does broaden when podcasts are added to the mix. No longer is everything squeezed into one place or format, but audiences can personalise their “preparation for the theatre” to their own needs and preferences. It remains to be seen if those accustomed to traditional AIs are at all interested in the podcasts. While the interview with Toneelhuis demonstrated a great awareness of accessibility and a willingness to offer more than only the traditional approach in order to cater for more diverse audiences, it also became clear that the current offer is still in the experimentation phase, and the collection of user feedback will be crucial in determining the relevance and usefulness of such initiatives.

This brings us to some of the limitations of the study and future avenues of research. A number of crucial questions remain unanswered, particularly if we apply a user-centred approach to this issue: Do AD users find mainstream podcasts to be a valuable asset in terms of accessibility? How should AIs be modified in order to cater for broader audiences? Should these types of audio-based programmes remain separate, or would it be best to aim for hybridisation? In order to answer these questions, a preference study is the next logical step. As this article is largely product- and maker-centred, we suggest shifting towards a user-centred direction [12] in prospective studies on the audiosphere.

Lastly, there is an inherent paradox ongoing in this study. We have referred to podcasts as an incidental access service contributing to the overall accessibility cluster, but there is at least one collective that is excluded from this practice: Of all the assessed theatres, none provide transcriptions of the podcasts, thus excluding people who cannot access the audio. In terms of successfully reaching a true achievement space for all, research as well as practice needs to carefully consider the diversity of audiences, abilities and stimuli to include in the developed access cluster.