Traditionally, audio description (henceforth AD) has involved a sighted describer (or describers) writing a description to share with blind audiences once a theatre production is complete and ready to be shown to an audience. An alternative that has received only limited attention in the AD literature is "integrated AD" whereby AD is conceived from the start as an integral part of a production. 1 This article compares traditional AD with integrated approaches by exploring an initiative by Unscene Suffolk, a small community theatre company for blind adults based in Suffolk, UK. This article also incorporates material from interviews with a number of theatre practitioners carried out as part of an ongoing research project into integrated access organised by Extant, Britain's leading performing arts company of visually impaired people.

Integrated AD

Integrated AD arose in response to traditional AD in which, as Amelia Cavallo has noted, "AD is added after the piece has been created and is delivered to the blind spectators via headsets. Unless directly using the service, it is likely that those attending an audio described event will be unaware that anything different is happening", and "the describer behaves as a neutral entity that does not deliver specific interpretations of the story, characters or emotions being portrayed" (126). Integrated AD rejects the separation between the artistic team and the "neutral" interpreter that Cavallo describes. Indeed it is debatable whether neutrality in AD is possible or even desirable. 2 Cavallo argues that AD should be treated less as a neutral way of conveying the source text and more as a creative tool for "connecting both audience and performer to each other and the artistic content of a piece in a positive way" (133). When asked to define integrated AD, Jenni Elbourne, the artistic director of Unscene Suffolk said,

Integrated description means doing something which perhaps the whole audience might be aware of as opposed to just the visually impaired headset user and also potentially integrating the description into it in a way which means that it becomes a part of the artistic product as opposed to an add-on. (Interview)

It is clear from this that integrated access is currently used as an umbrella term incorporating five aspects; that it be non-neutral (creative and /or subjective); that it be collaborative so as to reflect the director's vision (auteur); that it is considered a priori; and, that it be open and inclusive, available to be heard by all. I will discuss each of these aspects, as well as other issues raised by integrated AD, such as loyalty and authenticity, with reference to my work with Unscene Suffolk.

Unscene Suffolk

In September 2015, Unscene Suffolk decided to collaborate creatively with me in my role as a professional describer. In her initial discussions with me, director Jenni Elbourne explained,

For the last 3 years we have created scripts with integrated AD, which has attracted plenty of positive feedback, however some performers find [the AD] really difficult to learn (not least because they can't see what they're describing) and a few have commented that they feel we don't actually need as much description as we are using. So, we'd like to mix things up a bit next time, and explore other ways that we might use description. ("Audio Description")

Elbourne's integrated AD project gave me the chance to join Unscene Suffolk for a number of workshops as they developed their new show, A Zimmer of Hope, a comedy set in a fictional nursing home called Withering Heights. Even having a conversation with a describer was new for some of the company members, a measure of how divorced the describer usually is from both actors and users. As one of the actors put it: "It was incredible to meet someone who actually provides that service. Loved it." There are logistical reasons for the traditional separation between company and describer(s). For a describer to work on a show that is still in development means scripting a description in the knowledge that the action you are describing and the length of time available to fit that description in between instances of dialogue will almost certainly change. Also, for a blind or partially blind cast, it helps if the cues (from dialogue or AD) are fixed, so that the cast can learn them as part of the script. Many do this from audio recordings which are not as easy to update as text-based copy. In order to write the AD into the script, at least some of the action must be fixed at the time the script is written. In this instance, Elbourne, as author and director, wrote much of the description as she wrote the script. Elbourne's engagement is an example of auteur AD and is in keeping with the importance Naraine, Whitfield and Fels place on the role played by the director in any collaborative or devised approach to AD of live events:

Directors in devising theatre act as the initial researchers and collaborators for a production, whereas in traditional theatre, directors are the original decision makers. The director's role in devising, compared to traditional directing, changes from one of interpretation to one of conceptualization, although the director is still ultimately accountable for ensuring the play is coherent and entertaining. (117)

However, as Elbourne is a director rather than an audio describer, she produced phrases that I, as a describer, would never say. For example, I work on the assumption of normality such that if I describe a continuous action, I assume my audience will understand that it continues to happen until I mention it has ceased or is replaced by some new action. Elbourne's AD contained several instances of repetition or reiteration of actions such as: "they swish their skirts…they continue to swish their skirts." This type of redundancy might explain why some of the company felt they had formerly included too much description.

The first workshop of the project involved looking at the language of description to assess where the boundaries of describer neutrality might lie. 3 Together we watched an excerpt from the film Borat. The AD uses vulgar, colloquial expressions to describe a fight between Borat who "is stark bollock naked" and his Producer "with his sizeable arse" (Borat). Most of the company agreed that although the AD was offensive, so too was the film, making the AD appropriate and acceptable. Then the actors were split into small groups to explore descriptive language around stereotypes, such as the "chavs" Vicky Pollard and her friend Rochelle comically portrayed by Matt Lucas and David Walliams ("New Nationwide").

I was struck by the language used in the descriptions generated by the groups: many of the adjectives they chose had a tactile or embodied element. Instead of "fat" or "chubby", blind participants suggested "lumpy" and "excessively rounded." Keeping the descriptions active (a strategy favoured by describers), they gave us:

Both blokes playing women, fleshy faced, splattered with make-up, wearing shell suits or dated sportswear and white trainers, they strut bouncily. Bleached hair scraped back into a scrunchy falling down onto their ample bosom, they wear oversized gold hoop earrings and multiple gold necklaces spelling out "Once Upon A Time."

Clearly people with a visual impairment can describe and need not be restricted to feedback roles in the development of AD, as tends to be the case in traditional AD (Fryer 19). Companies such as Extant have demonstrated the impact of a partially blind actor creating and delivering their own description. Cavallo is a visually impaired actor, singer, musician, and circus aerialist. In her paper exploring the creative possibilities of AD in artistic performance, she provides a first-person description of her performance as a burlesque dancer in Show Girls (2007), commenting,

I describe parts of my body with sensual movement and music. The AD gives the piece a quality of self-voyeurism that, according to various feedback, sounds confident and comical. This would have been difficult to achieve and potentially uncomfortable to listen to in a standard AD setting. (128)

A Zimmer of Hope took a humorous look at politically correct language, illustrated by this passage taken from the unpublished script, from a scene set during a quiz show. The describer is revealed as a character and asked to step out into the open:

Although this scene is comic, it addresses important concerns over issues of subjectivity. Should the describer be given carte blanche to say what they see (Snyder) or should they couch their language to make it politically correct, bearing in mind the sensitivities of those they are describing and the audience they are describing to? Agnieszka Szarkowska suggests that auteur description lets the describer off the hook in this respect, arguing that it "incorporates the director's creative vision in the AD script … and thus gives the audio describer the artistic license to depart from the dictate of objectivism" (383).

At the next rehearsal for A Zimmer of Hope, I was able to discuss aspects of the audio introduction or pre-show notes with the actors, in particular the descriptions of each character, so the cast knew how I was describing them and I could ensure they were comfortable with it. For example, I was able to check that my estimates of ages were accurate. However, one of the cast heard my description of another ("He looks like a man who enjoys rock music and works out") and questioned it, so I revised it to "and tells me he works out." It could be argued, then, that integrated AD could be subject to a degree of censorship resulting in a more biased description than traditional 'warts and all' AD that an independent describer working detached from the company might provide.

Exposing the describer to terminology used in the production by bringing them in at the rehearsal stage has other benefits. In her discussion of the language of dance AD, Eleanor Margolies points out that in rehearsals, dance steps are not only demonstrated visually, but also communicated orally, for the company members to learn. This type of language can usefully inform the AD, helping the describer overcome any limitations to their own knowledge or vocabulary. Margolies explains, "At best, a description is a negotiation between members of the creative team, but it will always be partial, reflecting the viewpoint, experience and limitations of the describer" (18).

Another challenge of creating Integrated AD is a practical one of perspective. As a describer on stage in A Zimmer of Hope, I was not in a good place to see what was going on. I had to resort to the describer's traditional strategy of creating an AD script by working with a video of the performance. This style of description offers a view from the outside. It could be regarded as strongly ocularcentric to offer a view using an external reference frame that purports to be neutral but more likely reflects the conscious or subconscious biases of the sighted describer(s). Cavallo demonstrates the creative possibilities of a more embodied style of first-person AD that reflects the ego-centered frame of reference of the performer(s) (127-129).

One of the ideas behind A Zimmer of Hope was to make the describer more "visible" by making her a character involved in the action. As Elbourne explained,

We sort of scripted Louise as a character into the show. So there was a scene where the actors became aware of the presence of the audio describer. And that was very interesting and brought up all sorts the questions about what is the role of the describer when they come on stage? […] Are they a character? Or are they separate? Can they be spontaneous? Can it be scripted? Is it still audio description if it's scripted? (Interview)

The second workshop was in fact a rehearsal, in which I, as the audio describer, was placed in a scene that had already been blocked with the rest of the cast. Describing from within, and making the AD more transparent, had a number of advantages. It helped to raise the visibility of AD, so that the whole audience became more aware of AD as an access service. Also, anyone in a disabling situation, for example a sighted member of the audience sitting in a seat with poor sightlines, could make use of it. In addition, those reliant on the AD did not have to wear a headset, which can feel excluding. Not only the audience but also all the cast heard the AD, which is not usually the case. While this can be useful in enabling a cast with visual impairments to know what is (or should be) happening, more problematically, as I inserted description, I was effectively influencing the action. "Phyllis looks nervous," I described, after which Phyllis confided, "I'll have to practice looking nervous." In traditional AD, it is the action that leads to the description, but in this production, it was the other way around, as my descriptions ended up generating the action.

Yet AD's ability to influence the action also opens up creative possibilities. Nathan Geering is the director of Rationale, a hip-hop theatre company that has been developing Integrated AD. In an interview for Extant, Geering said,

Instead of just having audio description as an add-on, we have actually had it as a key factor in determining our choreography. So for example normally what happens is you have the choreographer creating and the AD which will be added on. For us that's not great. So what we decided to do was seek out our audio describers and we would create like an 8 count and we say to the audio describers, "OK say what you need in order to let your audio description breathe." And they might say, "Well we need you to spin on your head for another eight counts." Or, "We need you to walk another four steps because otherwise as soon as the movement's happened it's gone again and it's not enough time for us to be able to AD what we're seeing." What was really interesting is in doing that, it took our choreography in a completely different direction [ …] it became something else, just a completely different journey. It was brilliant. (qtd. in Fryer and Cavallo 36)

This two-way, creative exchange is also advantageous to the describer, who is usually constrained by a fixed soundtrack, and therefore also to the user who is less likely to suffer from the AD being squeezed into too short a gap. It highlights the opportunities for flexibility of source material which is not usually possible with traditional AD.

The second workshop allowed me to pinpoint what I felt was problematic about the pre-written description (see below). Arguably it was not description but a series of stage directions. These fulfill a different purpose, that of telling the actors/director what to do. Often, stage directions contain more information than will ultimately be needed in the AD. Unlike AD, they are not governed by the rule of synchrony which Franz Pöchhacker defines as the décalage or time lag between the original and the translator's rendition (117). Stage directions do not need to correspond with the timing of the action nor do they stick to the present tense, which is another requirement in the guidelines developed for traditional AD for screen ("Ofcom"). For example, one description written for the onstage audio describer to say at the beginning of A Zimmer of Hope reads: "Gertie who has been staring curiously at the fish tank, snaps out of her trance" (Elbourne, "Zimmer" 2). In traditional AD, this would be rendered as two separate descriptions using the present tense. The first would be placed as Gertie initiates the action, followed by a pause before Gertie looks up, at which point the second would be inserted:

• 1: Gertie stares curiously at the fish tank…
• 2: Gertie (she) looks up and snaps out of her trance.

In a stage direction, the timing of the action has to be spelt out whereas in the AD it remains implicit. Another example from the script, "As Felicity continues to speak, Wendy takes out a hip flask and swigs from it" (16), can be reduced to: "Wendy swigs from a hip flask." As the audience hears Felicity speaking, it is unnecessary to include that information in the AD. Her action of taking out the hip flask could be included if time allowed and would be necessary if the manner in which she produced it developed her character, made a noise that was hard to interpret or affected the plot, but it can otherwise be omitted without loss of meaning.

With the AD incorporated into the script, the actors of Unscene Suffolk treated it as a stage direction and used the AD as a cue. That is, they would wait to hear it before making their move, while I was waiting for them to start the action, in order to keep my description in synchrony. The combined effect was to make the performance hesitant. The other problem highlighted above is that the way the director had written the AD was not the way I would describe. As I felt uncomfortable with Elbourne's wording, I was trying to re-word it while waiting to see what the actors did, while, in turn, the cast was waiting to hear me say the scripted words. From Elbourne's point of view:

We got stuck in a slightly vicious circle of Louise needing to see what the actors did before she could describe it but them waiting for her because we had to say in our scripts Louise will speak here, here and here. We found that the actors were then waiting for her to speak before they would continue which was a bit of a problem because that meant they weren't doing anything so she didn't have anything to describe. …. And I think that the biggest lesson that we learned from it … was that for Louise as a describer with lots of experience the ability to be spontaneous and to describe things that were happening which may be slightly different on different occasions was really a kind of almost like a red line for her. She really didn't want to be so heavily scripted even if she'd written it herself. She didn't want to be tightly bound to a script which she couldn't deviate from. Prescribing the description to fix it for the cast limited the describer such that she could not reflect any spontaneous action. In fact the action ceased to be spontaneous as it became prompted by the AD, rather than reflected in the AD. (Interview)

AD can be thought of as a form of oral translation not unlike conference interpreting (Fernández et al. 72). Anne Schjoldager, who lists loyalty amongst her major criteria for judging interpreting quality, defines loyalty as faithfulness to the source text (191). It is clear from the above that, for the describer, loyalty to the source text in the AD of live events, means reflecting the spontaneous action of a live performance, those little deviations of actors' business in response to changes of circumstance from one performance to the next, such as the presence of an understudy or perhaps in response to audience reaction (Fryer 17). But for the members of Unscene Suffolk, the value of spontaneity is not so clear cut. After the show's run ended, Elbourne discussed it with the group. She reported,

They actually found it quite difficult to choose between the importance of spontaneity in a description which they do value, but also they felt that we as a group have a really great opportunity by working with describers during our rehearsal and scripting process to create scripts which mean that the description is not coming over the top of dialogue. And as a group that seems to be kind of unanimous thing that they find very difficult […] that no matter how great that description is, there's just no point in doing it if it's over dialogue and can't be heard. And they didn't have a clear consensus as to whether they felt that it was worth compromising or sort of eliminating that possibility for the sake of spontaneity. (Interview)

Loyalty is also used by scholars for interpreting quality in the sense of accuracy (Lee 168). As Elbourne's comments show, another concern of a describer doing traditional AD is to position the description so as not to mask dialogue or critical sound effects ("Ofcom"). Prescribing or pre-scripting the description makes this easier to achieve, especially if the creative team is willing or able to make changes to the source text to facilitate this.

A third aspect to loyalty concerns where the describer's primary loyalty lies. It might be to the source text, to the AD user or to the artistic team. This would seem to be at the heart of any potential conflict between spontaneity and auteur AD, as the describer aims to help the audience enjoy the particular performance, while the artistic team might prefer the describer to stick to the AD script, ignoring accidental mishaps and describing a potentially non-existent 'ideal' version of the play. The training for describers in the early 1990s pressed home the importance of the describer steering clear of the artistic team in order to retain their objectivity. It was considered essential that the describer reflect the view of the (sighted) audience, saying what could be seen, rather than what the director might want the audience to see. One argument for this is that of inclusivity in recognition that AD users want to be no different from the sighted audience, and especially want to be able "to laugh at the same time as everyone else" (qtd. in Fryer 175). A description that, for example, neglected to mention that when he walked across the stage an actor's trousers fell down, due to the unexpected loss of a button, would exclude the AD users from conversations in the bar at the interval.

Another question that arises from Integrated AD is whether the description is more authentic if it is generated collaboratively with the artistic team. In her discussion of the translation of Native American poetry, Susan Hegeman argues that "authenticity has always been a category of value in [Western] culture, opposed to the copy" (268) but she points out:

What an extremely problematic standard authenticity is…. First, it assumes that a few people can represent the interests, beliefs, knowledge, or agendas of a large group, which is seldom the case. Secondly, it assumes that we can correctly place, evaluate, and understand "authoritative" cultural knowledge…. (271)

Levi Strauss, when talking of myth, suggests,

There is never any original: every myth is by its very nature a translation, and derives from another myth belonging to a neighbouring, but foreign, community…that some listener tries to plagiarize by translating it in his fashion into his personal or tribal language, sometimes to appropriate it and sometimes to refute it, and therefore invariably distorting it. (qtd. in Hegeman 272)

If "performance" is substituted for "myth", surely the same doubts remain about AD, integrated or otherwise. As a describer of live theatre, I see a play many times and each performance varies. Sometimes it varies only marginally, sometimes in more major ways as when a cue is missed or an actor introduces new stage business or the set breaks down or one actor is substituted for another. Which then is the authentic performance? If I say "He holds out his hand in greeting" and he doesn't, does that make my AD inauthentic? When a described action fails to materialize, a describer might mutter to their co-describer "well he did yesterday." In that case, it could be argued that even if inaccurate for the performance in question, the AD might be more authentic in the sense of being more representative of the production as a whole than the anomalous performance seen on one occasion by the sighted audience.

To summarise, my experience with Unscene Suffolk has demonstrated for me that there are pros and cons to both traditional and integrated AD. Because traditional AD is developed by one or more describers working in relative isolation it may be regarded as strongly ocularcentric, offering a view using an external reference frame. Traditional AD is not available to everyone, and for live events it is usually only provided at selected performances. In its favour, traditional AD is highly responsive, as the describer can improvise to reflect the demands of live performance, although this risks the AD intruding on dialogue. By contrast, integrated AD is conceived during the production process with the describer working in collaboration with members of the artistic team. This ensures it will be included in every performance so that anyone might make use of it. In addition, collaboration with the artistic team at the devising stage means that the source material may be manipulated to accommodate the AD, creating more space where necessary. Integrated AD is also more likely to present the embodied "viewpoint" of a performer. This ego-centred frame of reference is less likely to be neutral but any subjectivity will be that of the artistic team, which could be seen as more authentic. If it is created too far in advance, for example at the same time as the script, Integrated AD risks sounding like stage directions without reflecting the full visual nuances of the final production. It is important that there is a two-way exchange between the describer(s) and the creative team.

By being conceived from the start with the sanction and full participation of the originating artists, integrated AD demonstrates a commitment to access on the part of the creator(s). The advantage for the users is that AD is part of the weft of the piece and will be available at every performance. It is unlikely, for example, to be dropped if budgets shrink. It has the added advantage of raising the profile of AD with the wider audience and the performers. The extent of integration may vary, but at the very least an integrated approach carries with it a seal of approval in the sense that the AD users know that the director, and possibly the actors too, have sanctioned what the describer says. Perhaps it is time for this form of accessible theatre-making to become the primary approach to AD, so that AD it is thought about from the start, and is an integral part of every production.