Making the creative accessible - Using design and description to make movies more accessible for a visually impaired audience.

Ida Solvang, Trond Ausland, Magne Lunde and Miriam E. N. Begnum

MediaLT, Jerikoveien 22, N-1067 Oslo, Norway


Abstract: This paper describes the integration of universal design principles into the film-making process to improve universal access of a feature film, using audio-description and optimized sound design as strategies to improve accessibility. The study implies that combining fundamental ideas of universal design and audio-description yield a significant gain in accessibility of film, exposing the visually impaired audience more directly to the depicted events as well as providing the information necessary to follow the plot. Incremental user evaluations by blind and low vision film viewers were paralleled with the development and completion of the film product. User evaluations dispute that optimized sound design communicate the course of events, thus lessening the need for audio description. Rather, optimized sound contributes to heightened emotional involvement. The piloted approach holds a great potential for providing furthered access to a visually impaired audience, conveying through sound additional content, without inhibiting the creative process.

1. Introduction

A feature film can be described as a medium where the combination of audio and visual images compels the audience to imagine the portrayed experiences and events. As such, the film can be approached not merely in terms of the audiovisual material it contains, but the experiences it mediates. This approach was the main starting point of the study, aiming to improve access to a feature film for a visually impaired (Lunde 2009).

Traditionally, movies have been made accessible to the target group through audio-description, a voice-over describing visual elements of the film deemed necessary to follow the plot. The intent of this project was to improve accessibility by integrating fundamental ideas of universal design into the film-making process, consequently surpassing the limitations of traditional audio-description. The hypothesis was that audio-description could be reduced by introducing simultaneous sources of audio information. This is an innovative approach in making film accessible as it emphasizes design rather than adaptive services, allowing for more flexible ways of communicating the movie content to the target group.

Artistic expression and universal design have a tendency to be seen as binary oppositions to each other. For instance, there have been disputes as to whether artistic vision or universal design should be the main focus when designing public buildings. Our objective was to make the two complement rather than antagonize each other.

2. Background

When viewing a feature film, persons unable to perceive visual information will usually pay careful attention to information mediated through the audio-dialogue and sound design. When audio-describing a movie, the focus is on conveying the course of events to the visually impaired audience. A script is made describing visual information not aptly conveyed through sound or dialogue, carefully timed to fit pauses in the movie dialogue. The script is then recorded as a voice-over track to be selectively played, while the original film audio is subdued.

In Norway, audio-description was established as a service by MediaLT through the Lydtekst project (Lunde 2008). As part of this project, MediaLT trained four professional audio-describers. The knowledge gained from establishing audio-description was the starting point for our study.

Though audio-description certainly serves the purpose of greatly improving accessibility of audiovisual media for a visually impaired audience, it has clear deficiencies from a universal design theory vantage point. A universally designed product is supposed to adhere to certain principles underpinning the universal design concept. These include accessibility to the product for as many groups as possible without additional effort or adaptive services (Story 1997). In most cases, audio-description is made as a separate entity after film release. The movie product is thus not universally designed, rather the audio-description track is an adaptive service added to the product in order to make it accessible.

This, of course, has implications for the product rendered to a visually impaired audience in terms of availability and quality. First, being produced as a separate entity, the audio-description track is not available as part of the movie product, but as an assistive product. It may be released only on certain DVD editions. When a movie is telecast, the audio-description will not usually be available even if it exists. Playing the audio-description track may require special technology or additional effort.

Second, in movies like “The March of the Penguins” or “Mamma Mia!” the plot only provides a gateway to music and visual experiences. Thus, the sighted audience will be left with main impressions that have no place in an audio-description outlined by the plot. This effect on the visually impaired audience is enforced by the audio-description track offering as information what is offered as stimuli to the sighted audience, even in movies where the plot is the main focus. Being told that the character “slashes throat after throat” is of course a different experience from seeing the visual depiction. Research is currently being done on enhancing the stimulus effect of assistive services for making audiovisual media accessible (Fels 2005). Similarly, research is being done on colour enhancing closed captioning (Fels 2006; Udo 2009).

A more pervasive consequence is that the audio-describing process is separate from the film-making process. The assistive product is based on interpretation, rather than on the intents of the product designers. Audio-description is not simply neutrally referring visual information. No matter how skilled the audio-describer, he or she cannot refer all the visual elements of the film due to strict time limitations. The audio-description process largely consists of carefully selecting which details to include. This is usually defined by the plot as interpreted by the describer. Choosing the details that most poignantly facilitate the plot may mean leaving out other information that may or may not be considered important by the creative team. Being based on interpretation of a finished product, there is no way of knowing whether the intents of the creative team are being preserved in the audio-description. This becomes most visible in ambiguous situations, where the describer is forced to choose one interpretation over another. As such, audio-description crosscuts the principles of universal design.

Very little research has been published on the subject of audio-description, reflected by the references section. Virtually no research has been previously completed on optimizing sound design for visually impaired, as this is a novel approach. Some information can be found on sound design for visual viewers, however these approaches are non-relevant for our study.

3. Enhancing accessibility

The objective of our project (Lunde 2009) was to improve the accessibility of a feature film for visually impaired, by communicating as much of the movie-content as possible to the target group without posing inhibitions on the creative process. The strategy chosen was to combine optimization of the sound design with audio-description of visual elements. Sound design refers to sound in a film that is not dialogue or music. As a visually impaired audience will derive information from the sound design, optimizing sound was considered a premise for turning sound into a new main source of information. Subsequently, the hypothesis was that less audio-description would be needed.

The movie chosen for this study was non-animated, thus involving shooting at physical locations. Therefore, designing sound mainly entailed editing, re-recording and adding special effects. The sound design was optimized to allow the audience to extract as detailed and thorough information as possible. The following five designing techniques were selected to achieve this:

  1. Including organic sound: Sound from location is usually omitted in films, or edited to a greater extent, e.g. because it duplicates visual information. We however kept much of the original sound from a location. Thus, an audience using sound as a means of information in daily life may obtain information conveyed by the natural sound of the depicted events and locations. Organic background noises were kept and emphasized in order to provide location information, e.g. a kitchen fan to pinpoint a kitchen.
  2. Including dialogue: Dialogue duplicating visual information was included, referring emotions or locations, e.g. a character naming a location while a sign stating the name of the location appears. This is usually avoided because the dialogue is deemed unnecessary as long as the visual conveys sufficient information.
  3. Adding descriptive sound: In some cases, sound was added to depict an event or a location in greater detail, e.g. highlighting the sound of a closing door central to the scene. This included surround-mixing of moving objects, e.g. a train on a long bridge implying distance.
  4. Including sound in music sequences: Noises referring locations or events depicted on screen were included subtly in the background of music sequences, giving necessary cues, e.g. the sound of someone walking through a hallway and busy city streets.
  5. Removing surrealistic sound effects: A few sound effects were removed, e.g. a door slamming that was originally made surrealistic to dramatize the effect. This was done to avoid potentially disorienting a visually impaired audience.

4. Design evaluations

Incremental user evaluations by blind and low vision film viewers were paralleled with the development and completion of the film product.

4.1 Test screenings

Early access was provided to four movie scene sequences. A sound design test optimizing was next completed in these sequences, to produce test examples. Emphasize in the test optimizing phase was on including as much real sound from shoot as possible as well as adding extra sound elements aimed at providing additional information about a place or an action.

Three of the four sequences were assessed as appropriate for test screenings. For each of the three sequences, two test versions were produced; one with the original sound and traditional audio-description, and one including optimized sound as well as adjusted audio-description. The adjusted audio-description was less packed, reflecting the view that a poignant sound design would reduce the need for extensive audio-description.

Eight visually impaired participants were selected to evaluate the sequences. The Norwegian Association of the Blind and Partially Sighted (NABP) and MediaLT co-operated to find volunteer participants. The participants were divided into two equally sized groups, each presented with a different sequence view order (see Table 1). The two versions of a sequence were shown consecutively. Participants were unaware of which versions were optimized and which were traditional.

Table 1. Sequence view order
Evaluators Sequence view order
Group 1 Sequence 1 – traditional version, Sequence 1 – optimized version, Sequence 2 – optimized version, Sequence 2 – traditional version, Sequence 3 – traditional version, Sequence 3 – optimized version.
Group 2 Sequence 1 – optimized version, Sequence 1 – traditional version, Sequence 2 – traditional version, Sequence 2 – optimized version, Sequence 3 – optimized version, Sequence 3 – traditional version.

Each of the six sequences was played twice within the presented order, to ensure assessment capabilities. Immediately after each sequence was played, the four group members were individually interviewed by four interviewers. An assessment form was developed and used to conduct the interviews, focusing on the audience member’s perception of the events, locations and actions depicted. The form thus contained detail questions specific to each sequence. In addition, the participants were asked to describe their overall and sound impressions.

After all six sequences were played and interviewees individually asked to evaluate these, semi-structured focus group interviews were conducted, consisting of the four participants in each group. There, the participants were explained the specifics of sound design and audio-description in the footage of film clips. They were asked to discuss the particular sequences, sound design features and audio-description details, and further expand and continue comments given during the individual interviews.

4.2 Final evaluation

The completed feature film was shown in a regular cinema, using technical equipment for playing an additional digital audio track. Invitations to freely attend and subsequently evaluate the movie were mediated through multiple channels; e.g. the NABF magazine, MediaLT e-newsletter and project group websites. 23 visually impaired participants volunteered. In the end, one of these were unable to attend. The final audience including consisted of 22 visually impaired and accompanies, in all 39 people. Of the 22 volunteered, four were underage - visually impaired children - and was therefore excluded from participating in the evalution.

The evaluators were interviewed individually via phone in the days following the viewing. Interviews were focused on the role of sound design and audio-description in yielding information efficiency. Interviews were held via phone. The interviews were structured, with items asking the evaluators to express perceptions of sound and audio-description and assess how various audio features conveyed information.

Of the 18 volunteered evaluators, only 14 could be reached. The 14 evaluators were fairly evenly distributed between the genders; 8 were women and 6 men.

5. Results

5.1 Test screenings

Test screening evaluations gave no support to the assumption that sound optimizations and additions reduce the need for audio-description. On the contrary, the feedback revealed that low level of audio-description in sequences with extensive sound design left some confusion about actions and locations that were not explicitly pinpointed. Sound design information was not always sufficient to follow the plot and capture essential visual aspects. For example, the viewers would miss a character putting her hand to her mouth in disbelief, an action central to the plot.

Moreover, details in the sound design not concurring with the visual depiction caused confusion. For example, a dog used as a sound prop and not present on location drew much attention, and there was confusion about whether or not it was central to the scene.

However, the feedback supported the hypothesis that the use of optimized sound design enhances the movie experience and supports audio-description in conveying events. The results showed optimized sound design yielded a more detailed perception of isolated depicted events than generally availed through audio-description. This led to the awareness that though using additional sound, the audio-description could be altered, even if not reduced.

Based on this feedback the focus of accessibility design was altered; great caution should be exercised in relation to reducing audio-description, sounds should be so poignant and specific referring as possible to avoid any confusion, and the two elements should be tailored to support one another in mediating visual information.

5.2 Final evaluations

There were considerable variations in the evaluators’ knowledge of film as well as audio-description experience, and thus their abilities to accurately define and analyze the sound effects used. In addition, the degree of using sound as a primary source of orientation varied, and in particular partially sighted evaluators had difficulties with consistently being conscious of sound elements. Even so, main points were consistent.

In general, all viewers had a positive and engaging movie experience, and none had difficulties with following the plot. Audio-description was emphasized as a vital part of the movie experience, providing information that enhanced visual details and gave cues necessary to follow the course of events.

The level of information given through the audio description was well received by all viewers, though some wanted more emphasis on visual details – such as more accurately conveying the brutality of an arrest.

All viewers felt the audio-description of the characters’ body language and reactions were of particular importance, and a large contributor to the movie experience. The facilitation of optimized sound design in portraying events made including more extensive audio-descriptions of other visual elements possible, and this possibility was mainly utilized by including precisely this type of details. This proved of great value to the evaluators:

“Really good, cause you got more information, especially facial expressions and who they looked at and so on, that’s important, especially how the girls reacted when they got upset and not just their look, but that they turned away and stuff.” (Translated)

Further, test viewers expressed a large emotional involvement as a response to the sequences with extended sound design. The detailed sound design yielded a more vivid and accurate perception of the featured events:

“...for example when the girls were going to take the train in the night and just as that Julie woke Mette and threw a snowball at the window the train howled and then one figured that they were in a hurry”. (Translated)

The evaluators particularly focused on the added organic sounds in this respect. The sound of snow crunching was mentioned by many evaluators as vital to the experienced powerful auditory sensation.

In addition, the surround sound mixing to facilitate location orientation received attention:

“The sound recording was very nice, especially when they sledged and the sound moved from the one channel to the other as they sledged.” (Translated)

The viewers thus found both sources of information vital to the movie experience, and valued the tailored collaboration of the two sources of information in efficiently mediating details.

6. Discussion

The initial hypothesis was that including fundamental ideas of universal design in the film-making process would reduce the need for audio-description. Our findings however indicate that audio-description and sound design seem to serve different purposes in providing access. The results from the sequential and final screenings were coherent. Also, similar feedback was given regardless of the audience members’ skills in watching movies.

Results indicate that an optimized sound design greatly enhances a movie experience. It mediates specific details about locations and actions. However, information given through sound design falls short in giving an overall understanding of a sequence, a plot and a chain of events. This is perhaps due to the information nexus in a feature film, which is both clearer cut and more compact than in a real life situation. In life, a visually impaired person is usually aware of the current location, people present etc. In a movie, locations and events change rapidly, and visual details like gestures and costumes are typically more central in facilitating the plot. This could of course depend on how visually concentrated the movie-content is.

Although a transparent and referring sound design also makes the movie more accessible in a setting where audio-description is unavailable, it is not considered a sufficient information provider on its own. A normal level of audio-description was evaluated as vital for grasping purely visual elements and for following the plot.

Including universal design in the creative work process of the film played a significant role in making both the movie itself and the assistive audio-description service more suited to the target group’s needs. It minimized the lack of consistency between the visions and intents of the product designers and the adapted product. In this case, the cooperation between the creative team and the audio-describer allowed for supervisions and revisions of the audio-description script. Thus, it was certified that the information given was accurate, to the point and that no details deemed essential were left out.

Equally important, the film content was communicated in greater detail than is usual in a film without a transparent sound design. Having an extensive and poignantly referring sound design relieved the audio-description of mediating some information. This reduced the time constraints on the audio-description, allowing more focus on conveying additional visual information. Hence, transparent sound design and audio-description supplement each other as key elements in improving accessibility of the film content. The accessibility made possible through a combination of the two strategies yielded a more flexible and holistic way of making movie content accessible to the target group.

The study shows that universal design does not need to entail harder work or higher costs. The main technique used for improving accessibility was to include sound from location usually omitted because it is considered unnecessary. The key aspect of this approach is that audio is not merely a supplement to the visual information in a film, but a provider of information and stimuli in its own right. As such, universal design is not a straightjacket for the creative process. Rather, universal design is an instigator for a dynamic way of purposely comprising the different components of a movie into a whole.

Our experiences imply that seeing universal design as compromising the creative process, frequently a debate regarding public buildings, may not be true in film production. The techniques used to optimize the sound design may seem invasive. However, they appeared very subtle even though their functionality for a visually impaired audience was great. A greater consideration for the stimuli given through sound and dialogue, and how the elements comprising a film may be used to convey information, implies a purposed communication of the movie content to the general audience as well as to the visually impaired audience. Thus, we suggest that in movie making, universal design principles may increase the quality of the creative end-product also for the general audience.

Our approach entails a new realm of movie design opportunities that are as of yet little explored. A distinction can be made between the kind of information provided through audio-description and through an informative sound design. Of course, both will provide information and stimuli. However, realizing that the two, to some extent, serve differing purposes has implications for future work, allowing greater accuracy and assertiveness in providing accessibility to movie content. Further research into the areas of sound and information design of film audio to mediate movie content holds a great potential for increasing the amount and aptitude of audiovisual media entertainment available to the target group.

7. Conclusion

Universal design principles were integrated into the film-making process of a feature film, aiming to improve universal access for a visually impaired audience. A combination of an informative sound design and audio-description was used. The hypothesis was that the need for audio-description would be reduced when information was conveyed more efficiently through the sound design. This showed not to hold true. User evaluations indicated that the combination of an informative sound design and audio-description greatly enhances the movie experience for a visually impaired audience, separately serving two different purposes. The film content was mediated to the target group in greater detail, facilitating a more nuanced perception of the depicted events. The movie experience was enhanced through the stimuli and detailed information mediated by the sound design, maximizing exposure to depicted events. The audio-description provided information necessary to follow the plot, and gave access to central visual film elements.

Our study indicates that combining fundamental ideas of universal design with adaptive strategies improve accessibility, compromising neither the creative process nor the quality of the finished product. Access can be improved through seeing the film audio not as supplementing visual information, but as an information provider in its own right. Our approach asserted that the intents and visions of the creative team were aptly communicated to the target group. Further research in these areas holds great potential for increasing the amount and degree of audiovisual media content available to a visually impaired audience.

8. References