Applying Cinematic Virtual Reality with Adaptability to Indigenous Storytelling

Cinematic Virtual Reality (CVR) is a style of narrative-based Virtual Reality (VR) experience built on filmed or computer-generated 360-degree videos. Since CVR is becoming more popular and widely accessible, researchers and practitioners have been trying to address challenges such as the conflict between the viewer’s freedom of choice and the creator’s control over where to look, or the risk of missing key story elements due to such freedom. As part of the solution, CVR creators employ attention-guiding cues, introduce viewer interaction, and combine these two techniques into all-encompassing CVR production frameworks. However, there are very few CVR projects that embrace the various differences in the backgrounds, preferences, and expectations of each individual viewer. Further to this, they do not consider the content creator/owners’ perspective when presenting and digitizing stories from the real world, especially when considering viewer’s connection to the cultural significance contained. In this article, a case study is presented to explore the use of adaptability to viewer situations and the coherence to Māori (the indigenous people of New Zealand) storytelling contexts in CVR experiences. In the case study, we began with co-design sessions with storytellers from Te Rau Aroha Marae (an active Māori cultural heritage site in the deep south of New Zealand), about appropriate features to collect from visitors to a virtual storytelling event, then co-built personas as representative tools. 360-degree videos of pūrākau (stories) were then captured and presented via an adaptable VR system. Evaluations were conducted with the storytellers to validate the system and to collect reflections and opinions on both the use of CVR in Māori storytelling and the cultural appropriateness of CVR with adaptability. We conclude this article with a discussion of possible improvements for future CVR frameworks.


INTRODUCTION
People now frequently use 360-degree videos to experience Virtual Reality (VR).Wearing VR headsets; they can freely rotate their heads while being immersed in omnidirectional scenes.The phrase Cinematic Virtual Reality (CVR) has been coined to describe those narrative-based VR experiences that use either real-worldcaptured or computer-generated 360-degree videos to portray a complete tale [42,76].Many scholars have attempted to address the Narrative Paradox (NP, the conflict between the viewer's freedom of choice and creator's control over where to look) [82] and Fear of Missing Out (FOMO, the risk of a viewer missing key story elements due to their freedom to look around) issues in CVR [37] from a practical perspective by using various attention guidance cues [6,7,19,76] to direct viewers to story elements in the scene.Creators are looking at designing CVR experiences involving viewer interaction in the progress of the narrative, moving away from the traditional linear passive format [20,60,63,72].Comprehensive frameworks have been developed to aid CVR creators in the development of immersive content as well as the design of the content delivery system for improved viewer interaction and immersion [9,36,78,80].However, we noticed that there were not many CVR systems that can deliver experience with adaptability to various viewers' backgrounds, tastes, and expectations.Instead, they normally considered all viewers as a uniform and featureless template group.Moreover, although following the guidelines from various frameworks created by predecessors in the same realm, many CVR projects failed to place enough emphasis on the content owners when they captured digitized stories from the real world.
To address these shortcomings, we worked closely with Māori (the indigenous people of Aotearoa, New Zealand) storytellers from a marae called Te Rau Aroha (a marae is a fenced-in complex of carved buildings and gathering grounds for religious, spiritual, and/or social purpose, belonging to a particular iwi (tribe), hapū (sub-tribe), or whānau (family)).We used CVR to create a digital experience based on the framework Adaptive Playback Control by Tong et al. [78].In our study, we explored if and how can we implement a CVR system with viewer adaptation to appropriately deliver virtual Māori storytelling in a way similar to the real-world face-to-face experience.The case study presented in this article started with a series of co-design sessions with the storytellers.We chose to work with this specific group of Māori storytellers because of the geographical uniqueness of their marae and the Bluff township.Te Rau Aroha Marae is located in the southernmost region of Aotearoa, New Zealand, called Motupōhue or Bluff, built in the 1800s, with a rich recollection of tīpuna (ancestors) and stories embedded in carvings and tapestry (Figure 1). 1 The rūnanga (tribal council) of the marae connects people and acts as the governing council to those who live in Bluff and on the islands off the south coast of NZ such as Ruapuke Island and Rakiura (Stewart Island).Statistical data2 shows more than 40 percent of the population there self-identify as Māori and maintain a traditional lifestyle, including the seasonal collection of tītī (mutton-bird) and oysters.This uniqueness of the Bluff community sets a strong cultural background for this case study and emphasizes the necessity for being culturally appropriate and following Tikanga Māori (Māori customary practices, behaviors, and protocols) when applying digital technologies to traditional cultural activities.
The context of this study also supported a second motivation.It sought to determine whether digital storytelling can help Māori people better connect with stories when it is built following their protocols.With these motivations in mind, the following steps were drafted to conduct this indigenous digital storytelling case study: (1) Planning the overall structure of the CVR experience with a focus on the narration's adaptability to each individual viewer's traits and background; (2) Conducting a series of co-design workshops with subject-matter experts (SMEs; in this case, Māori Storytellers of Te Rūnanga o Awarua from Te Rau Aroha Marae in Bluff, New Zealand), around what elements are needed to enable adaptability to bring it close to the activity of oral storytelling; (3) Using the elements summarized from the previous step as key points to guide the capturing of a series of customizable CVR content; (4) Developing a CVR system that delivers an adaptive storytelling experience; (5) Evaluating the system and its experience by the SMEs to see if (1) it matches the expectation of storytellers, like the oral storytelling experience they would like the viewers to have; (2) it is culturally appropriate.In the following sections, the process of narrowing the exploration of this project will first be elaborated.It starts from a design space about introducing the viewer as having an active role in CVR storytelling, continues onto the co-design sessions conducted with Māori storytellers, includes outcomes and reflections, and concludes with an evaluation and discussion of the CVR system.

BACKGROUND 2.1 Cinematic Virtual Reality
When VR emerged into the public's notice around the year 2016 and both the VR headsets and consumer-level 360-degree cameras became more widely available, creators put their focus on 360-degree video as a simple version of VR and quickly adopted it as an immersive storytelling medium [4,24,76].In that time period, the 360-degree video was mostly employed for storytelling projects, experimentation, and exploration [42].In contrast to this mainly short and simplistic footage for fleeting thrills, creators now strive for content with complete narrative arcs that let the audience feel involved in the story world.The prefix cinematic is used to define these narrative-based VR experiences that feature a complete, multi-part story arc [13].The name CVR then came into use, and it is described as a type of VR experience in which the viewer watches pre-recorded or pre-rendered content in the format of omnidirectional (360-degree) videos, using a Head Mounted Display (HMD) [42].Creators swiftly recognized the distinctive nature of CVR compared to traditional cinema, identifying two fundamental challenges in CVR storytelling: the NP and the FOMO [37,82].NP encapsulates the conflict arising from the divergence between the viewer's autonomy to choose and customize their experience and the creator's authority in shaping the narrative structure.FOMO represents the viewers' apprehension about missing crucial story elements due to their freedom to look around the VR environment [34,46].These challenges have been associated with sub-optimal narrative comprehension and diminished emotional engagement within CVR experiences [28,51].Studies have shown that when viewers have excessive freedom in navigating the narrative, it can lead to fragmented storytelling and a reduced understanding of the intended plot, affecting the overall comprehension of the story [30,69].Moreover, the unstructured nature of exploration may contribute to a decreased emotional investment, as the viewer's attention becomes divided across various elements, rather than being guided towards pivotal narrative aspects [10,85].
Initially, the creators attempted to migrate their filmmaking skills by shifting the viewer's perspective from third-person to first-person to embody another person in the scene, which matches the intrinsic characteristic of CVR content and was proved effective in empathy-inducing stories [14,62,83].They then applied cinematic techniques to wider genres to convey expressive content and tell compelling stories [11,44] while using non-intrusive cues to guide viewers' attention without compromising immersion.However, some solutions such as post-processed rotations and snap cuts on orientations have been found to introduce uncomfortable viewing experiences [33,67].To overcome this limitation, researchers have explored the use of diegetic elements in the scene as guiding cues [15,61], and more recently, the use of human characters as cues [6].Such methods were also visible in several CVR projects that premiered on film festivals [2, 21,71].A promising approach is the use of Action Units (AUs) developed by Tong et al. [76] as a set of body language references to embed attention guidance into the narrative during 360-degree video filming, as illustrated in Figure 2.These cues have been shown to effectively maintain the level of presence while attracting the viewer's attention.

Involving Viewers in Storytelling Experience
Creators of CVR recognize the viewer's new role as an active participant within the scene and narrative, offering increased agency and interactivity through head or chair movements [22,76].However, CVR is still primarily viewed as a passive, "lean-back" medium with limited interaction possibilities [20,63].Achieving full interactivity and enhancing the viewer experience has been a subject of ongoing discussion, considering the complex hardware and viewer effort required [52,74].In parallel, fields such as museums, tourist attractions, and theme parks are striving for adaptability in traditional tours to cater to different situations [47].Understanding incoming viewers and profiling their characteristics is crucial for adaptability.Personalized data, including demographics and visitor behaviors, help comprehend their motivations and further categorize them into five groups based on characteristics and motivations [17].The choice of features for profile generation depends on the stability or dynamism required.Various methods have been employed to collect variables and create or match profiles for viewers [47].These methods include: -Explicitly asking visitors via a questionnaire before the experience [3,54,64].This is similar to real-world rituals, such as the pōwhiri (a formal process of being called onto the marae with a formal welcome and whaikōrero (speeches)) used by Māori to remove tapu (sacredness) and welcome new visitors before entering the marae.
-Gather information implicitly through a brief interactive interview rather than direct inquiries, aiming to avoid visitors' discomfort when approached abruptly in the museum and asked for personal information [1,29].-Implementation of objective-based methods [25,40,41,47], whereby visitors are invited to choose an object among several thematically related choices.The choice represents the visitor's preferences in terms of theme, language, character, and so on.-Classification-based method [64].The content creator anticipates and prepares a list of expected visitor profiles in advance.The system collects a few key features from the visitor and matches them to the profiles to categorize each visitor.Therefore, determining the features to be used for profile generation and the method of collecting these features is crucial in understanding visitors' motivations and creating customizable experiences for them in cultural heritage settings.Notably, the Classification-based method frequently utilizes the tool "personas, " a common method in user-centered design, to summarize and construct a set of typical viewers for museum professionals to refer to during the development process [39,65].For example, in the CHESS project [54,65], the team first created a set of personas to represent typical museum visitors, based on Falk's theory [17], then used the variables collected from visitors to match them to one of these personas.Although relatively coarse, the classification method reduces the number of profiles created and saves time spent on matching content to profiles.

Frameworks for CVR Storytelling
In previous work, researchers have explored the use of CVR for effective storytelling.They also considered how viewers, who are a necessary part of the CVR experience, will be given an appropriate role in the narrative process.The outcomes, however, remain as separate roles; they either sit on the creator's side looking at content capturing [42,82] or sit in the realm of viewer interaction with the content cultivating a viewer-centered mindset [20,63,77].In other words, the CVR creators currently have to use a trial-and-error approach to identify any guidelines for solutions when they encounter problems producing content.Since consuming film, including CVR, is more of a "lean-back" experience for many, striking this balance is key to its appeal.The production of interesting and enjoyable material while maintaining the creator's authorial control over the storyline and topic requires a framework that supports both content creation and viewer interaction design.This will allow viewers the freedom to interact alongside narrative progression, with their agency matching expectations.
Theories such as the "co-construction of the story" and "ludo-narratives" (narration delivered via playing) have gained considerable traction in the immersive storytelling domain [31,80].These theories emphasize the indispensable role of the viewer in shaping the narrative progression and enveloping them within a complete and engaging experience.Contemporary research within this realm predominantly concentrates on refining story structures that position viewers as co-authors of the narrative [9,59,79].The proposed frameworks advocate for adaptable, non-linear story architectures, often comprising self-contained "mini-stories" that empower viewers to manipulate the sequence of presentation while retaining narrative coherence.
In parallel with the design of these narrative structures, extensive research has been undertaken on viewer interaction methods, primarily integrated into the initial phases of story composition [23,36,45].These efforts have focused on employing frameworks that seamlessly merge interaction design with content creation.Central to this approach is the viewer's interactionencompassing gaze, gestures, and selectionas a pivotal instrument for advancing the narrative.The typical procedural approach begins by structuring the story, preparing content for individual story segments, formulating an action schema for viewer interactions, and harmonizing the viewer's participation with the narrative progression, all aimed at crafting a unified and engaging experience.
However, an essential, often underrepresented aspect is the evaluation and perception of user interaction methodologies, despite their integral role in shaping the overall viewer experience.Literature exploring the subjective and objective assessment of user interactions in CVR remains relatively sparse.For instance, research by Johnson and Garcia [27] illustrates the significance of gaze-based interactions in enhancing user immersion and narrative engagement within VR environments.Conversely, Smith and Lee [70] looked into the potential drawbacks of over-reliance on gesture-based interactions, underscoring challenges related to user fatigue and diminished engagement in prolonged storytelling experiences.
Generally, while the existing CVR frameworks prioritize narrative structures and interaction design, there is a discernible lack of robust evaluation methods and subjective user perceptions.Based on this work by others who used them as design references [18], Tong et al. [78] established a two-part framework called the Adaptive Playback Control (APC) as a set of guidelines for interactive CVR producers, emphasizing the necessity of making the viewer's experience the first priority.CVR content creators are expected to use APC as a reference for their production process, selecting story structures to support their goal with storytelling, preparing content segments using Action Units (AU) as directional cues, and assembling them with interactive elements made for viewers into an enjoyable and satisfying immersive storytelling experience with interactivity.Additionally, it guarantees that the creators maintain essential authorial control of the story throughout the process while simultaneously allowing viewer agency and for the storytelling experience to be adapted to viewer expectations.

DIGITIZING M ĀORI STORYTELLING USING CVR WITH ADAPTABILITY
Māori storytellers tend to not tell a fixed story, but instead assess their audience and select the necessary story elements from a pool in their head, assembling and arranging them to create a narrative that fits the current context and embellishing it with the teller's unique touches [49].This is similar to the non-linear storytelling introduced for some interactive games [9].Therefore, appropriate frameworks and tools are critical when capturing Māori stories and converting them into digital experiences while maintaining their authentic formation.

The Unique Face-to-face Māori Storytelling
When telling a story, storytellers select conventional imagery and dramatic devices, arranging them based on the audience, situation, and storytelling objective [43].Different tales emphasize different characters and events, each with the storyteller's unique style.This process involves connecting and assembling kernel aspects, simple stories, narratives, and disclosures for the listeners [35,53].For example, the Tama Rereti story of Māori ancestors tells of Tama Rereti's journey to discover the origins of the Aurora Australis, featuring prominently in the decorated wharekai (dining hall) at Te Rau Aroha Marae.The kernel stories include Tama Rereti's voyage with young chiefs and experts, his exploration of the Southern Ocean, and his unfortunate demise from choking on a shrimp.The same story differs when delivered by different iwi (tribes), different storytellers, and towards different audiences [49] (as different tellings), including to remember Te Waka o Tama Rereti, which is a set of stars for navigation.A kanohi ki te kanohi (face-to-face) oral storytelling like this is how Māori people transfer and pass on mātauranga (knowledge) and Tikanga Māori [5], in the past and also today.
In the Māori context, oral storytelling also involves the audience.The storytelling practice is seen as a form of group knowledge sharing that strengthens the spiritual and interpersonal links that unite communities [68].Thus, both the stories and the ways of telling are as varied as the people telling them and the listeners they are speaking to.Oral storytelling is also chosen in this study above other media, because when the storyteller is actually present with the audience, the power and authenticity of the story come through as mana (the authority, influence, and charisma of the storyteller, in this context).The spiritual depiction is crucial to Māori storytelling, as it provides the narrative context and significance [55].

Digitizing Māori Storytelling Using the Matrix of Segments and Takes
Digitizing the stories told by Māori storytellers has been carried out since people noticed the necessity for preserving knowledge, maintaining the face-to-face experience and participatory characteristics [12,38,55].CVR was chosen for this task, because its viewers are fully immersed in the scene and provided with a feeling close to a real-world face-to-face experience [16].However, previous works focused on the temporal succession of content segments only [57,66], leaving the impact of viewer interaction with content disclosure untouched.As Fig. 3.The MST structure consists of segments presented linearly from left to right and one or more takes that are selected.Each segment along the timeline may have different numbers of takes available in parallel.On each playback pre-scripted by the creator or picked by the system, the viewer sees one take for each segment and, therefore, experiences only one version of the story with specific choices of takes assembled.As an example, there are five connected segments chosen for one playback, shown at the bottom of the figure.Notice each segment contains a specific take.
discussed above, the teller of face-to-face stories accepts change.Thus, in this project, we wanted to not only use the CVR technology to capture stories but also to create a more adaptive experience, resembling face-to-face oral storytelling activities.This introduced a necessity for a structural model to lead the production work and enable flexibility and reconfigurability in the captured content [35].
Several studies have introduced the concept of customizable stories for disclosure variants in the context of immersive storytelling.However, prevailing customization frameworks, as suggested by Zagalo et al. [84] and Roth et al. [59], were found to be impractical for application within CVR projects.Extensive research by Carstensdottir et al. [9] and Ryan [66] has shed light on the constraints of these frameworks within CVR.In a 360-degree video-based CVR project, the narrative structuring typically adheres to a linear, chronological presentation of captured videos rather than spatially scattered sequences.Unlike interactive video games where player input significantly influences the narrative, CVR, being a primarily "lean-back" experience, relies predominantly on the storyteller's choices and preferences for content disclosure.This demonstrates the unique narrative constraints faced in CVR projects and the infeasibility of existing customization frameworks for these immersive storytelling formats.Therefore, in this project, the idea of a Matrix of Segments and Takes (MST) for CVR was introduced to demonstrate that the content blocks, repetitive variations, and alterations on time sequences and their entire topology can all be consolidated into a complete toolkit for creators.The MST model includes the structure of a story in its finest granular elements, namely, the Segments and Takes.Figure 3 depicts their relationship during the formation of a story.
Segments are small content units within a story following its temporal succession.A segment is a unit containing a piece of the story that can no longer be divided into smaller parts to maintain its integrity and be comprehensible to the viewers.In other words, a segment is like a mini-story that if consumed by the viewers, should at least be understandable within this same segment.An example is a scene of a conversation among several actors on a set with several lines of dialogue, where the viewers can get the general picture of what those actors are discussing.Takes are the cluster of content clips of the same segment.Like the takes cinematographers capture onset in filmmaking, the takes here comprise multiple captured clips of the same story segment.Content-wise, they are the same story, but each is unique in detail.For example, the beginning of a story can be told in different ways, with different styles and patterns, by different storytellers, and facing different groups of listeners.In an interactive or immersive storytelling context, using takes can lead to the viewers experiencing variations between repetitions.In general, the MST represents a two-dimensional mutable story structure (as shown in Figure 3).The sequence, structure, and dependency of the segments are on one dimension; on the other, within each segment, there is a parallel array of sibling takes with the same material, but different presentations.In the disclosure phase, the CVR system selects content blocks under the MST, either by design or by considering the viewer's traits and preferences.Consequently, each viewer experiences a single version of the story during playback, which is made up of (1) a collection of segments in a single temporal order and (2) one particular choice of takes under each segment block.Such a combination is shown at the bottom row in Figure 3.The illustrated structure also supports a different story variation in which the method of selecting takes is random.
In this project, the MST supports the adaptation of a story for diverse audiences.The storyteller can record a story with multiple segments and capture several takes for that segment.Given that the viewers will have different knowledge bases, personal connections, and expectations towards the story, a careful choice of takes can allow everyone to find the content they experience fitting their expectations, yielding higher engagement and enjoyment if it is adjusted to match the viewer's profile (adaptability).The whole CVR experience also falls more in line with its face-to-face roots.

METHODOLOGY 4.1 Co-design Sessions for Persona Creation
Central to this case study is the strategic adoption of a co-design methodology, a deliberate choice stemming from the intrinsic significance of Māori cultural themes and storytelling traditions within the project's scope.The choice of such a method recognizes the pivotal role of indigenous perspectives and cultural authenticity.This iterative method, well-documented for its effectiveness in fostering inclusive and culturally sensitive solutions in various cultural projects [36,48,73,81], aligns with our commitment to actively involve Māori collaborators in the development process.By engaging directly with storytellers from Te Rau Aroha Marae in Bluff, New Zealand, our objective was to authentically capture and embed Māori storytelling elements into the narrative fabric of the CVR experience.
In this project, we engaged directly with two storytellers and several Māori collaborators, all from Te Rau Aroha Marae.We invited all of them to each of the design sessions and let them lead the conversation.In the sessions, they were introduced to the project and the idea of building a storytelling system with adaptability using digital technology.Tools to represent viewer variations were initiated, discussed, and iterated.Content capture plans were also formalized during these sessions.
During each session, co-design with experts of Māori protocols and knowledge was conducted to discuss and explore the following session questions (SQs) as a start, aiming to ensure cultural appropriateness throughout the process: -SQ1: What are the appropriate and main visitor features that a storyteller will consider?-SQ2: What design tools can be chosen to appropriately represent the visitors that fit real storytelling scenarios?During the first session, the tool personas was chosen to represent the typical visitors in the Māori storytelling context.In this project, the personas were used in the production stage to describe the detailed information on imaginary visitors constructed out of the storyteller's deep understanding of the storytelling and marae visiting experience, since in the digital project the actual experience of the digital content by the visitors will not happen until the contents were captured and constructed into a usable system.Notably, this is quite different from the CHESS project aforementioned (Section 2.2) in which the personas are internal to the system and used for mapping viewers during the visiting stage.
The research team first pictured a virtual Waitangi Day (the national holiday commemorating the signing of the treaty between Māori and non-Māori in New Zealand) event being held at the Bluff marae as a storytelling background (an actual event that took place at that marae one year earlier).With this virtual stage confirmed, the researchers asked the storytellers (both had participated in hosting the real Waitangi Day event) to recall the typical visitors that come to the marae during such an event from their memory and experience.From the gathered information, the researchers picked visitor types identified as "came before" or "will definitely come" by the two storytellers, and then put them into a list of preliminary visitor profiles to create personas.It was also verified with the storytellers that they would welcome those visitors, if they came to the marae, and tell them the stories of the marae, the tribe, and of the carvings in the wharenui (meeting house).They were also informed that later in this project they would conduct similar storytelling activities to be captured for the CVR experience.The purpose of the personas created, or virtual visitors, was to help storytellers deliver different versions of the story, as they do in face-to-face scenarios.For example, a storyteller from the marae was asked to look at one visitor from the personas list and assume they were the one visiting the marae.The storyteller then should imagine telling that visitor stories in the meeting house.Those stories would then be captured digitally.This set of prototype personas was then brought to the co-design sessions to be reviewed for any missing critical factors and whether they were appropriate for a Māori storytelling context.Because of the COVID-19 pandemic, all the co-design sessions were held as online video conferences.
The discussions during the sessions were all recorded and transcribed to text using an online tool. 3The scripts were then manually read through for possible errors, organized, and analyzed using Thomas' General Inductive Analysis [75] method as a guideline.However, due to limited time and the number of experts we could reach out to, we could not conduct a formal thematic analysis as described in the guide in full scale.From the scripts, insights and iterations towards the initial prototype personas were extracted and summarized as presented in the following paragraphs: During the first co-design session, the key fact highlighted in the discussion was that in the view of Te Rūnanga o Awarua storytellers from Te Rau Aroha marae, Storytelling is about the transfer of knowledge and reciprocal knowledge sharing.From the Māori storytellers' perspective, knowledge transfer, sharing, and exchanging all must follow "access control" rules, and different people would be provided access to different knowledge.They believed the mechanism of access control to be a process of accessing the viewer's knowledge background and their personal traits to determine the appropriate knowledge access level.Factors include the visitor's age range (Tamariki/children, Rangatahi/teenagers, Pakeke/adults, and Kaumātua/elders), whakapapa (genealogical connection, including Manawhenua those who have a connection to the family, and Those who do not), previous visits (including waewae tapu, i.e., first-time visitor or regular visitor) are considered critical for storytelling.The storytellers also added that the visitor's intention reflects what knowledge they are after.Besides, the storytellers proposed a change from personas representing an individual to a group of visitors, as their storytelling is mainly focused on groups.They also provided suggestions on each group's motivation and the corresponding length and details of the stories they would tell (listed with details in Table 2).For example, a normal visitor to the marae could be given access to only the generic story.However, a visitor who has a genealogical connection to the marae and had previously been there could be given access to a more specific story, such as around the tīpuna (ancestors) of the marae.
In the second co-design session with the same group of Te Rūnanga o Awarua storytellers, we moved on to discussing the design details of the group personas and the use of adaptive digital storytelling experiences in Māori contexts.Storytellers provided further feedback on the personas, including the use of the term Others to represent those who are not Māori, Ngā Hau e Whā (the four winds, north, south, east, west, people from all directions) to represent the visitors who are not genealogically connected to the local tribe; and the removal of children from the age groups, as they are regarded as a non-typical user of the VR technology.
After the two sessions, we updated the personas following the feedback and suggestions about their features and values, emphasizing ethnicity, age bands, previous visits, and genealogical connections.The final version is presented in Table 1 and Table 2.  Group Intention of the visit 1 They are here to hear the stories from the Bluff marae, including of the history and the carvings.They also will share stories from their own marae (story sharing and exchange, mostly generic).2 They have come to visit on a field trip from the school's indigenous art class.They would like to learn about the carvings and decorations in the wharenui.3 Their visit has been arranged as a part of the Ocean Resource Summit held nearby in Invercargill.4 They are honorable veterans invited for a local ceremony about remembering the people's past and looking into the future, on ANZAC day.They will visit the marae after the event.5 They are representatives of the Rūnanga (tribal council), coming to attend the Waitangi Day event at the marae.6 They have come to visit the marae, look for re-connection with the family.

Content Preparation
In content preparation, the Māori collaborators at the co-design sessions first made a choice on the topic of the stories.The MST model was used as a structural reference to assist this process.They chose to tell the stories about the whānau (family) wall of the eight carved walls within the meeting house on the marae, which features an abundance of stories.The whānau wall has significant meaning to the local family, as it illustrates the lineage of generations, from the First Family in Māori folklore to those who settled in Bluff and established the family.
It encompasses different carvings and weaving patterns, symbolizing various sub-threads and sub-segments of stories, which also makes it suitable to be applied with the MST model.During co-design sessions, the two invited storytellers also stated that they would mostly use a three-segment model (Generic, Detailed, and Depth) to tell stories.However, in this project, they would only use the Generic and Detailed segments (as shown in Figure 4), because they expected the digital experience would not completely replace the physical visit, and thus should not cover all story content.We also needed to cope with the limited time and resources available for the production.The storytellers also requested to treat personas 5 and 6 as special cases outside the three-segment model.The same story was told repeatedly towards Personas 1 to 6 to generate multiple takes of each segment.
It was suggested to the storytellers to use different disclosure strategies for different visiting groups (personas), either telling the same story in a different style, or a slightly different story, but on the same topic.For special cases 5 and 6, there were only lengthier Generic segments.Under the structure of Segments and Takes, the choices of actual content to fill in were in the hands of the storytellers.They were free to choose from the pool of stories that all linked to the whānau wall.In total, six Generic segments and four Detailed segments were planned for the study.The Generic segments were designed to be 1-2 minutes, and the Detailed segments to be 2-5 minutes long.For personas 5 and 6, their Generic segments were 8-10 minutes long.Transitional content, such as the introduction, welcoming message, and transitions between blocks to make the experience more fluid, were included in each segment.While in some other cultural projects, the story is adaptively assembled following different visitors' profiles [50], the stories in this study are captured as a whole package via the storytellers.The system does not alter the sequence of the segments once they are made.All content segments with their fixed structure for this study are summarized as an MST table in Figure 4.
With a capture plan ready in hand, the research team physically traveled to Te Rau Aroha Marae in Bluff where the storytellers are genealogically connected and captured the stories about the whānau wall with an Insta360 Pro 2 360-degree camera running on full 8 K (7,680 x 3,840) HDR at 30 fps to maximize preservation of details. Figure 5(a) shows a behind-the-scenes view of the capture work, including the whānau wall, the storyteller, and the 360-degree camera.These captured 360-degree videos were later down-sampled to 5 K (5,120 x 2,560) during the stitching phase to reduce the load on the PC during playback.The camera was stationary in all scenes.
As shown in Figure 5, the positions of the camera and the storyteller were carefully chosen following the content creation guideline of the APC framework.Two storytellers from the marae did the narrations for the content.
Before the capture started, one of the six personas was presented to the storytellers.The persona descriptions were printed on paper, listing all the key features from Table 1, plus a group photo collected online matching the persona features, as shown in Figure 6(a).The storyteller was instructed to spend a few minutes to become familiar with the persona, looking at its key features and imagining the group described were actually visiting, so he could prepare a story for them.The process was repeated for each group persona until all were captured.In every take, the storyteller was instructed to first tell the Generic part and Detailed part following it with a 10-second pause in between.Each story was tailored by the storyteller himself on site to reflect the characters of its corresponding group.The highlights and overview of all six stories are listed in Table 3.The clips were then trimmed and post-processed as segments, grouped under takes, by cutting them into two clips at the cut point.

The System
The system to control and present the recorded 360-degree videos was developed in Unity 2020.3.36f1 and was run on a laptop with an Intel i9 CPU and an RTX3080 graphics card to ensure the smooth playback for our experience of the relatively high resolution 360-degree footage.The captured 360-degree video footage (customized content) was loaded into the content library of the system.We also added a user interface (Figure 6) for the viewers to choose which visiting group s/he belongs to using the VR controller.
Since the focus of this case study was on exploring whether digital adaptability would be appropriate in the Māori cultural context, rather than how to accurately match viewers to their persona groups, the module that was supposed to match end-users with personas had not yet been built into the system.The prototype system was deployed on an Oculus Quest 2 headset.A ray-cast pointing method was implemented for the viewers to choose a visiting group from the six and to watch the corresponding 360-degree video (Figure 6(b)).Please note that in the interface, the texts under the photos cover only their age group, genealogical connections, and whether they are first-time visitors, due to limited space on the canvas.However, the full details of each group persona were printed and described on paper (Figure 6(a) and were provided to the storytellers at the beginning of the evaluation stage.
For the first four persona groups (1 to 4), a two-part story structure was used, with a branch at the end of the Generic segment to determine whether the experience should continue to the Detailed part or end right there (also depicted in Figure 7).While watching, the viewer would be first presented with the Generic segment.In the viewer experience, the system pauses the video at the end of the first segment and shows another panel asking if the viewer would like to continue to hear more story (the Detailed segment) or end the experience (Figure 7(b).If the viewer chose to continue, then the detailed segment would be presented.Otherwise, the system would reset the video player, remove the existing panel, and show the persona panel again (Figure 6).

The Evaluation
Evaluation sessions were conducted with storytellers instead of end-users for two reasons.First, end-users may be unable to perceive, compare, and then evaluate the variation between multiple story disclosures unless they are explicitly told and exposed to the backstage mechanisms [78].Running evaluation with end-users would also require the viewer to revisit the same story several times and try a series of versions, which introduces significant uncertainty about impressions, as the novelty factor wanes with repetition.Second, the storytellers were the experts and could best assess whether the digitized adaptability had helped them to capture their stories appropriately.They were all well versed in storytelling and familiar with most Māori story content, especially those originating from the marae to which they all have genealogical connections.
To investigate whether a digital storytelling system can be a viable replacement for face-to-face storytelling, we conducted an evaluation that compared the characteristics of the digital system to those of oral storytelling activities.To ensure cultural sensitivity and relevance, the evaluation was conducted using Māori protocols,

Group Summary of the story to that group 1
The storyteller first welcomed the group to the marae using Te Reo (Māori language) phrases and introduced them to the whānau wall and told them it emphasizes the First Family and all families after it.The storyteller then moved on to tell the story of how Ranginui (the sky father) and Papatūānuku (the earth mother) were separated by their children and how the tale is represented through the carvings on the wall. 2 The storyteller first used English to welcome the group and give them a basic overview of the wharenui and the walls, then he moved on to tell a short version of the Māori myth tale about the creation of the world.Then he introduced the statues and the carvings on the wall, emphasizing the process of making them, the materials used during the process, and the features brought in by the artist who carved them.3 The storyteller first used English to welcome the group and give them a basic overview of the wharenui and the walls.He then told a short version of the Māori myth tale about the creation of the world.He also talked about the ceremonies, especially the funeral-related rituals that will be held in front of the wall.4 The storyteller first used English to welcome the group and acknowledged their contributions to the country as servicemen.Then he introduced the characters carved on the wall and specifically talked about the war that happened between the two brothers in the first family and reflected on the value of peace at homeland.5 The group was first welcomed formally using Te Reo.The storyteller then told the story of how Ranginui and Papatūānuku were separated by their children.He then talked about how observing the Waitangi Day has a significant meaning in the community and how it was connected to the story of the first family.6 The storyteller welcomed the group to the marae and started the storytelling by citing several previous events that happened in the wharenui.Then he moved on to talk about the story of the first family and referred to the reunion of the people in the present day among the community.
which are well understood by the participants.A group session was held on-site at the Te Rau Aroha Marae with four storytellers and three researchers, with one researcher hosting and asking questions while the other two recorded responses.The participants tried and reviewed the system with content from one storyteller loaded and then repeated the process with content from the other storyteller loaded before engaging in a group discussion.
It is important to note that one of the storytellers was also a participant in the study but was not present during the evaluation of their own content.The session had three phases: the briefing, the experience, and the debriefing (group discussion).Inspired by the approach by Kriz [32], the procession of the phases and the questions for the group discussion were both designed in a fashion that started from open and casual talks before directing the discussion towards the vital factors expected to be pushed through evaluation.During the briefing phase at the beginning, the digital storytelling system and its feature of adaptability were introduced, followed by the use of personas and how the content was captured.The motivation for developing such a digital system and the purpose of running the evaluation session were also explained.Finally, the use of the VR headset, the controller, and the digital system itself were all demonstrated.
During the experience phase, four storytellers were present and rotated through the virtual storytelling system.Each storyteller chose two group personas from the menu and watched the content.It took 15 minutes for each viewer to go through the experience.While one was using the HMD, the laptop connected to the HMD mirrored the VR view to a monitor nearby, so the others were able to follow the viewpoint and discuss freely.This non-immersive exposure was specially designed for this evaluation to give users pre-exposure to the experience when they were not actively participating, so conversations and thoughts would start to formulate during the experience phase.It was also easier for the storytellers to talk about their thoughts and feelings when we moved onto the debriefing phase, cultivating an atmosphere of free discussion, since they had seen how others were doing and had a certain amount of communication beforehand.
After all the storytellers had tried the experience, they were first instructed to freely talk about their feelings and thoughts on the digital system, then moved on to a semi-structured interview, guiding the discussion towards comparing the adaptive digital storytelling experience with their knowledge of storytelling face-to-face in the real world.The conversation during the interview was recorded and then transcribed using an online tool 4 and manually adjusted to correct Te Reo Māori words and expressions.During the interview, the two key questions were phrased as follows: -Now imagine that one of the visiting groups you selected during the experience is actually going to visit the marae, and you will tell a story to them face-to-face.At the last moment, however, they need to cancel their visit because one group member tested positive on COVID.How viable do you think the digital version you just experienced is to use in place of the physical visit for this group?Is this an appropriate replacement?-What do you think is necessary to be added or improved in this digital experience for it to be actually put into use in the marae, such as for a setup for virtual visiting?
Other flexible questions helped keep the interview flowing.These, as well as open questions at the end for extra suggestions, are not explicitly listed here.Opinions and feedback from transcription were then summarized into four main aspects, presented in the next section.

RESULTS AND ANALYSIS
As described in the previous section, the evaluation sessions were conducted with storytellers at the Te Rau Aroha Marae.To compare a digital storytelling system with oral storytelling, a group session was conducted onsite, following Māori protocols and involving four storytellers and three researchers.The sessions were recorded, transcribed using an online tool, and manually reviewed, correcting Te Reo Māori words and expressions.The recordings facilitated a unique dynamic where storytellers interacted with the mentioned objects, enhancing the realness of the storytelling experience.The transcriptions were then processed through a semi-thematic analysis (instead of a formal one that will be in full compliance with Thomas' General Inductive Analysis [75] method) because of limited time and the number of experts to whom we could reach out.The analyzed transcripts and extracted themes were subsequently used to inform the study's findings, as presented in section 5.2.

Reflections from the Co-design Sessions
In the Māori worldview, storytelling is a process of reciprocal knowledge sharing, as not only the teller but also the listeners define the content and structure of a story.Listeners hold various base-knowledge levels and have different levels of access to certain knowledge.In a face-to-face scenario, the storyteller gets feedback from the listeners in the interactions and will gain an understanding of the listeners' knowledge baselines and other characteristics by observing their reactions.Meanwhile, such real-time interaction is not available in digital storytelling with pre-recorded content.
In this study, we aimed to construct a digital experience that still maintains the flow of information from the listeners to the storyteller even when they were separated in time.Besides, we also intended to follow the Māori protocols and traditions by acknowledging the storyteller's authorial control in the storytelling process and to create a visitor-centered system, which are commonly seen in the GLAM (Galleries, Libraries, Archives, and Museums) sector, customized and altered based on the visitor's behavior, intention, and preferences [54].
Therefore, personas are used in this study as instruments to bridge the pre-recorded telling of a story and the later viewing of the story.The personas served as references for the storytellers, to presume the knowledge access level, during the recording stage of digital storytelling.When working with the storytellers, we printed personas on paper as an assistive element to their storytelling, which is a departure from the assumption of typical user representatives, and design anchors familiar with modern design workflows.We also admit that a component to collect information from the listeners for an adaptation stage was not implemented in this study, which will be discussed further in Chapter 6.

Evaluation Results
Impressions of the CVR experience.All storytellers stated that watching 360-degree video via an HMD gave them a more immersive experience compared to watching a conventional video on a TV or mobile phone.They pointed out that the immersion CVR provided gave them a feeling of standing there with the storyteller at the scene while listening to the stories, such as "...what I particularly liked about the 360 experience, is sort of [you] sit there and I actually sat there and really listened to [the stories]..it makes sense as if I am really there... " [P01].The storytellers all stated that they paid much more attention to the content presented in CVR, compared to similar content presented via other types of media, such as photos, books, or video recordings.One example would be from one interviewee who said, "I find myself it's really really difficult to watch a static [flat video], like even on Zoom, and you are getting zero feedback...but here you get the feedback on facial expression and body movements and the environment... " [P02].Such attention differences were not expected by the storytellers, since they all thought they were already familiar with the content.Additionally, the fact that the stories were recorded in situ (in this case, in the meeting house where the storyteller was describing the carved or woven walls as well as the stories related to them), further helped to enhance the feeling of spatial presence, as one of the interviewees pointed out "I learned heaps on all sorts of content was just so interesting..

.it was relevant and regarded in that virtual reality, he [referring to the storyteller] actually had things to show us. So that cemented the knowledge that he was talking [about]... " [P01].
In addition, because the recording happened at the scene, the storyteller could at the same time interact with the objects being mentioned, such as pointing at and touching the walls.In the viewers' eyes, this also helped to increase the realness and the feeling of being there with the storyteller, mentioned as "...was the art one [group 2] where he was actually interacting with the wall and touching that thing, stroking the different parts and then pushing on the bits and he was explaining... " [P02] and "for an art class [group 2], I was glad to see specific patterns/art being explained in their original context location" [P03].All of those are not achievable in an audio recording or conventional 2D video.
One extra thing to point out is that the storytellers also discussed how the host talked to a virtual visiting group, instead of just an individual visitor.This observation came from the fact they looked at how the storyteller in the scene did not stare at the camera all the time, but instead acted as if there was a group of people, scanning evenly on all the virtual viewers there, across the camera.One person from the interview said "...he [the storyteller] was looking at all the different people who were potentially there and you are one of a group... " [P02] and "...in that virtual world, I think I only saw him look at me two or three times, but everybody else had his gaze until he looked at the camera straight away..

. " [P04].
Impressions on the use of system adaptability.All storytellers who experienced the digital storytelling system were aware of the characteristic of adaptability.Their first direct impression was the storyteller's variations when presenting the story to different viewer groups.They thought it was an excellent opportunity to see how the knowledge in the stories was structured and stacked differently from one to another, like one of them said, "..a really good thing I noticed about it is that you could actually stack up different users...and if they could group the content according to that... " [P04].By cross-referencing the personas and the key features listed, they noticed that the variation and reconstruction of knowledge from one group to another were mainly brought on by the differences in ages and visiting intentions of the groups.They especially noticed the relatively obvious change in the topic and theme of the stories from group 2 (a group of young art students) to group 4 (Army veterans visiting on ANZAC day).
The storytellers further discussed whether the introduction of adaptability made the CVR experience more like a face-to-face experience compared to a standard CVR.They proposed that the adaptive digital system could provide appropriate content to various visiting groups, such as school and community groups, which are the two most frequent types, making up about 80% of all regular visits.They thought it could be used as "one of the many ways to share knowledge" and could replace at least the activity of giving the visitors basic Māori stories, such as the story of how Ranginui and Papatūānuku were separated by their children.It was highlighted in the interview as well, e.g., "...then someone says for example, [name hidden] who has been here for the first time and it has given the overview kind of story [generic version]... " and "...ensure that they [the kaikōrero at the marae] were not overburdened [with] these types of [storytelling] and get the concern out of the head as well... " [P01].The storytellers also made the point that utilizing the digital system would free them from having to give the same generic story to every group that came to visit, allowing them to concentrate on telling stories with in-depth information targeted to specific groups.
They added, however, that a digital storytelling system, even with adaptability enabled, could not completely replace a real visit, because the latter involves more than just storytelling.As one of the storytellers put it during the interview, "...visiting a marae is more than listening to story in the meeting house, there are other experiences...when people come to the marae you offer them the kai (sharing food together), the pōwhiri, the chatting, etc.You offer them the full experience...This [referring to digital storytelling] is only part of the experience... " [P04].In the eyes of the family (in this conversation refers to the people from the marae), welcoming people visiting the marae is about showing manaatikanga (hospitality, care for people) and building connections between the visitors and the hosts.They do not think this can be achieved simply by letting visitors put on an HMD and go through the digitized experience.
Interaction capability as a viewer during the experience.The storytellers' feedback converged on viewer interaction during the storytelling experience when asked about future upgrades they would like to see in the system.When asked about the reason behind it, they stated that in Māori protocols they treat the storytelling activity as a type of bilateral interaction, with viewer involvement in multiple factors, including the pace, topic, and one thing unique to their culture, challenging the storyteller over authenticity (as one story will have different disclosures and versions between different tribes and families).Therefore, they expect the visitor to be able to interrupt the storytelling process, raise their hands, ask questions, or push the topic towards something else they would like to see, e.g., "...when he [the storyteller] was talking about Māui [the great hero in Māori mythology] and then his brothers on one of the carvings I wanted to say 'wait, talk more about Māui... "' [P02] and "..like he touched on the Tukutuku panels [a distinctive art form of the Māori people, a traditional latticework used to decorate meeting houses] with the art students...he talked about the different materials, how the colors were made...there could be people wanting to pause and ask follow up questions around the specifics... " [P01].These observations all show that they are comparing the adaptive digital storytelling experience with a real-life one and trying to have the capabilities of a physical visiting experience.The storytellers also added that interactivity was also seen as the process of building connection before knowledge sharing, as one interviewee put it, "...started to talk about that connection... the connection was pivotal before we can even get the content.So the Tikanga [protocol] of that is required to come on to the marae is going to be part of the learning experience as well, not just the fact of content that [storyteller] has given us... " [P02].
Other interaction elements were also brought up during the discussion.One was the change in the height of the camera.Especially for the teenager's group, the storytellers pointed out that in a marae visit [physical one], it is common for the teenagers to sit on the floor and look up at the storyteller.Such a scenario was not considered in the APC, as maintaining the camera height at eye level was initially recommended in the APC guidelines.The other is the control over the time length and pace of storytelling.For some groups, such as teenagers, 10 minutes or more would be too much for them, especially for a passive learning style.But the storytellers added that since the system has enabled self-control over the pace, it could mitigate the issue of time length over visitors' attention span.This additional interaction feature would allow the viewer to choose whether to stop after a generic segment or continue listening to a detailed version of the story.Other than that, since it will be a self-guided tour style, if implemented, then a pause function would be favorable in case the system user is interrupted by something else.

DISCUSSION
During this case study, we noticed that the APC framework could not only be used to design a CVR experience, including the content preparation and system deployment introduced in previous chapters but also to analyze the experience and interpret the evaluation results.Such multi-aspect applications are reviewed and discussed here in two parts, which are: (1) the use of immersive content with the APC for Māori storytelling; and (2) the adaptation of the APC to appropriately tell a Māori story in digital media.

CVR for Māori Storytelling
The suitability of incorporating Māori storytelling within the APC framework was assessed during the initial codesign sessions.The project was designed to utilize 360-degree videos and to focus on cultural heritage themes, while also allowing for viewer participation in the storytelling process.Personas were used to represent different types of listeners, and the digital experience was created under the Māori context.The resulting feedback from the storytellers as SMEs suggests that the stories captured through the CVR format, guided by the APC framework, are both usable and effective.The evaluation showed that the use of the APC in CVR production helped viewers feel more immersed with the storytellers in the virtual scene.However, the variable camera height is a new factor that emerged from this study.For example, when creating content for teenage visitors in this context, the 360-degree camera should be placed lower than normal to match their real-world experience of sitting on the floor.Special requirements like this should be considered when applying the APC framework to cultural heritage cases.The ineffectiveness of Action Units (AUs) was also noted when the storytellers were not visible in the viewer's Field of View (FOV).Adding extra visual cues visible to the viewer could make AUs more effective.Additionally, supporting the storytellers to maintain their pointing gestures for longer durations would help viewers avoid missing the guidance cues.
In the utilized system, storytellers appreciated viewers having the option to choose between proceeding into a detailed segment or concluding the experience after the generic segment.This feature empowered viewers to control the pace and duration of their engagement, showcasing the efficacy of providing self-regulation within the storytelling process, as emphasized in the latter part of the APC framework.However, resource constraints limited viewer involvement to two aspects: non-real-time configuration of attributes at the beginning and a midexperience choice to explore the detailed segment.Components focusing on more nuanced viewer engagement, such as viewer profiling functions, active input through text or voice, and the ability to interrupt the storyteller for an immediate change of narrative, were not implemented.While the Māori storytelling context aims for a conversational experience, where both storyteller choices and visitor responses hold significance, this balance was not entirely achieved in the implemented CVR experience.CVR cannot yet address viewer needs and intentions in the storytelling experience in the same way that an in-person knowledge exchange would offer.Functions including real-time viewer input with contextual awareness, finer control over the flow of the narrative, and narrative changes triggered by object interaction might all be considered as future enhancements.
Moreover, opportunities for future improvements also come from the necessity of addressing the need for a more balanced approach between a storyteller's top-down authoring power and a viewer's bottom-up needs from curiosity or personal interests.The challenge lies in designing interactive elements that seamlessly integrate the curated narrative with spaces for visitor contribution or content filling in in real time.This could involve incorporating viewer-generated content, such as personal anecdotes or questions, into the storytelling experience.Furthermore, enhancing the system's adaptability to dynamically respond to visitor inputs during the narrative could bridge the gap between the curator's predefined story and the viewer's individual interests.Striking a harmonious balance between curated content and visitor agency might be pivotal in creating a more inclusive and engaging CVR experience.The subsequent section will also deliberate on another direction, proposing a few potential improvements to the MST model, using emerging technologies for future studies.

Cultural Appropriateness of CVR with Adaptability
Although a comprehensive framework such as the APC can be used as a guideline to produce a CVR experience for Māori storytelling as used in this case study, its utilization by Māori (who are the stakeholders and the owners of their storytelling and Māori knowledge) and its appropriateness, requires further exploration and inspection.
The first step of our case study was the employment of adaptability into CVR.Instead of referring to a viewer's specific input, such as gaze or gestures, storytellers look at viewer participation at a higher level.They treat the process of storytelling as a way of reciprocal sharing of knowledge.While the viewer is learning knowledge from the story presented by the teller, the storyteller is also gaining new understanding and thoughts from the feedback from the listeners.They believe that the viewer's personal traits from the beginning influence how the storyteller is going to present the story, such as the choice of what content segments to use around a given topic.Such influence also continues into the storytelling process; the storyteller will dynamically adjust both the content choices and the style of disclosure of a story based on the viewer's responses and feedback.The viewer will also show various attitudes towards content segments based on the teller's performance and skill.Based on such knowledge, the viewer interaction guidelines from the APC were tailored for Māori storytelling by emphasizing viewer participation through personas instead of real-time input, accommodating the significant differences between visiting groups.As a result, adaptability was implemented on top of the APC, ensuring varied experiences for different viewers within the same story and creating a personalized connection to provide a digital alternative for remote family members.
The storytellers' thoughts on the characteristics of adaptability of the digital experience were especially highlighted during the evaluation to learn whether this characteristic (compared to other pre-recorded storytelling experiences) had contributed to making storytellers think that the digital storytelling system is appropriate enough to be used as an alternative when physical face-to-face storytelling is not viable.The implementation of adaptability also means creating customized content, in which the local storytellers are actively involved.It is a way they prefer: Letting the one who knows the story well, tell the story and tell it appropriately to the viewer attracts significant attention among Māori communities, indicated by statements from the interview such as "...when preparing class materials, one of the requirements is we want the content to come as localized...for content from here [Bluff] you would like to have someone from down here to talk about it... " [P02], "...we also need to go into the local marae and ask the knowledge holders, ask them what type of knowledge they would like to share and how they want to share it... " [P01], and "...because it is the content of mātauranga that the local marae would like to share and they are sharing it [being] really vibrant and different... " [P02].That feedback points toward the assumption stated at the beginning of this chapter, that the implementation of adaptability, letting visitor types trigger different versions of stories, is also a type of viewer (visitor) involvement in creating and delivering stories.It is safe to say that the APC guidelines have been successfully adapted to the cultural backgrounds and unique Māori protocols in this case.
If one looks deeper into the use of the APC powered digital storytelling system in the realm of cultural knowledge transfer, especially from the perspective of storytellers, then several outstanding observations emerge.First, the digital system used here succeeds due to the combination of CVR content with viewer involvement in the form of adaptability and has been proven to be usable and effective.However, when it comes to whether the digital storytelling experience can be used as a replacement for a physical storytelling experience at the marae, the storytellers' opinion turned out to be unexpected.In short, they found that the digital experience can have its place at the marae but is definitely unable to replace the physical face-to-face visit.Since, in Māori traditions, the visitor experience to the marae is more than just visually looking at the buildings and carvings or audibly listening to a story, there are other activities that have to happen face-to-face physically.Second, the storytellers repeatedly emphasized the use of digital tools only as a pathway to physical visits, as a teaser or preview, instead of a replacement.One of their concerns over the digital storytelling experience was that if it becomes available someday in the future, then it will stop people from visiting the marae physically.Later they were assured that the digital experience will not replace the physical one, since the physical visit to the marae is a unique experience for showing their hospitality and building connections with people.Instead, with the introduction of adaptive content in the digital system, one can capture and deliver the story deemed appropriate for each visiting group, helping to attract more people to visit the marae physically.The digital storytelling experience actually can co-exist with physical visits, instead of replacing it.Third, the storytellers did admit that the digital system could be applied as an option to deliver generic and entry-level stories to a wider group of visitors to gain a basic understanding of the story and Māori protocols, especially in the time of pandemic or when people are scattered in distant locations.When they later move on to the physical face-to-face storytelling with storytellers, the latter will know that those visitors have a certain background knowledge because they have gone through the digital experience.It also helps storytellers assess the knowledge level of the visitors, to avoid giving the basic story to all groups while not knowing which group needs it and which does not.
We should also acknowledge that the central focus of this case study was the use of adaptability and narrative variation in CVR.With CVR, the viewer's participation in the storytelling process is moved beyond the simple linear connection of story segments with adaptability enabled.It is now at the level of alternating the disclosure of the content.The assembly guidelines are extended beyond just viewer-interaction design.The co-design sessions also explored the most appropriate method, the use of predefined personas to help storytellers bring variation into the presentation and capture the customizable content.It adds to the content creation guidelines for CVR creators.This case study validated the use of adaptive CVR in Māori storytelling.It also illustrated that adaptation and careful consideration are necessary when using a comprehensive framework for different storytelling scenarios.More importantly, creators first need to learn from the stories' owners before finalizing the design of a CVR experience to ensure its appropriateness as a digital tool before handing it over to the end-users.

MST Model Improvement and Use of AI in CVR for Māori Storytelling
The MST model, while useful in structuring CVR content for personalized and adaptive storytelling experiences, does present limitations in certain scenarios.It is particularly effective in narratives where content variations are encapsulated within well-defined segments and takes.However, its applicability may be constrained in broader scenarios with more dynamic viewer-storyteller interactions.Future extensions to the MST model are worth considering as upcoming steps to ensure compatibility with a more diverse range of scenarios.
In previous cultural heritage-related projects, personalized responses were commonly implemented using rulebased systems [8,26,56,58], relying on predetermined structures and variations.However, as CVR technology advances, featuring real-time viewer interaction, there is a pressing need to develop solutions that can adapt to emerging requirements, such as responding to viewer intentions and dynamically filling in content during the experience.For instance, consider a viewer expressing interest in exploring a specific cultural element not explicitly covered in the prepared content.In response, the next iteration of the MST model could dynamically generate a segment and associated takes related to the viewer's query, seamlessly integrating this personalized exploration within the overarching narrative.From a technological standpoint, leveraging emerging technologies such as generative Artificial Intelligence (AI) and Large Language Models (LLMs) to create filler content in real-time offers potential improvements.This ensures a continuous and engaging narrative even when viewers venture beyond the prepared content, enhancing the adaptability of the MST model to a wider range of viewer interactions.
However, we do see the challenges and risks on this path.From a cultural perspective and adhering to Tikanga Māori about storytelling, numerous precautions must be taken before applying technologies such as AI and LLM to a CVR project in the Māori context.As outlined in previous chapters, Māori people regard the storyteller as the custodian of the knowledge conveyed in the narrations.It is the storytellers' right to decide what can be shared and who will be the recipient.The involvement of generative technologies could potentially breach these protocols, raising concerns about ownership of the generated content and appropriateness in creating new content that did not previously exist or was not inherited.
There is also the potential conflict between the roles of AI and the real human storyteller if both are present in the scene.While one might argue that AI-generated improvisation could guide the narrative back to the core content while maintaining coherence, such improvisation also relinquishes control from the storyteller over the entire experience's progression, leading to unknown viewer impressions and consequences.Moreover, this approach might contradict the expectations of Māori collaborators, as discussed in the previous section, who desire digital experiences like CVR to serve as teasers, attracting attention and raising awareness to encourage physical visits and face-to-face interactions with the whānau at home.Therefore, we again would like to emphasize that the use of such emerging technologies on cultural storytelling projects will raise further ethical considerations that we cannot yet foresee, requiring careful navigation.Preservation of cultural authenticity, avoidance of appropriation, and respect for the sacred nature of narratives are important.If we plan to carry out these technological additions in the future, then we will no doubt proceed via close collaboration with Māori, running further co-design sessions from the ideation stage, carefully discussing and pro-actively consulting with Māori collaborators to address concerns such as data privacy and sovereignty, consent, and responsible technology use.We believe careful consideration and discussion will also be necessary for other teams who are working with cultural heritage matters and facing similar situations.

SUMMARY AND CONCLUSION
In this case study, the development of an immersive storytelling system incorporating adaptability, guided by the APC framework, offered a unique exploration into Māori storytelling within the realm of CVR.Co-design sessions with storytellers and the formulation of six personas representing various visiting groups were pivotal in tailoring content and showcasing the synergy of cultural relevance and technological innovation.The capture of 360-degree videos by the storytellers, guided by personas, set the stage for an evaluation that demonstrated the suitability of the adaptable digital storytelling system for sharing cultural knowledge within the marae.The study's significance lies in showcasing the potential of interactive CVR storytelling projects, guided by the APC framework, in presenting culturally resonant experiences within both spatial and social immersion, in the context of Māori storytelling.Yet, as with any pioneering exploration, this study lays the groundwork for future endeavors that promise an even richer and more immersive experience.
Looking forward, future studies can explore an array of avenues to enrich CVR storytelling.One future exploration will be 6DOF CVR experiences, utilizing technologies such as volumetric capture to build more dynamic and interactive environments.Another direction of future exploration will be enhancing the MST model for better coverage of various viewer and storyteller interaction scenarios that CVR systems will face, as discussed in Section 6.3.Incorporating behavioral-level viewer-storyteller interactions to capture viewer responses and adjust storytelling patterns accordingly is also worth exploring.Those enhancements seek to bridge the gap between the virtual and real, providing users with a more seamless and engaging narrative interaction.Simultaneously, increased involvement of Māori experts in refining storytelling content promises a more diverse and comprehensive representation of Māori cultural storytelling within the immersive experience.Upcoming research will progress into real group storytelling and multi-player viewing scenarios, aiming to analyze the collective experience of shared narratives in the CVR environment.Future studies will emphasize the application of the APC framework across diverse storytelling scenarios, ensuring consistent adaptability in immersive experiences.These advancements align with the core principles of preserving and honoring Tikanga Māori, ensuring cultural authenticity and integrity within complex storytelling narratives.The integration of Tikanga Māori will guide these experiences, fostering a sense of cultural appropriateness within the immersive world.

Fig. 1 .
Fig. 1.A map view showing the location of Motupōhue, Bluff in Te Waipounamu, the South Island of Aotearoa, New Zealand.An enlarged view shows the Bluff township and islands off its coast.Te Rau Aroha Marae is also presented as an inset photo, bottom right.

Fig. 2 .
Fig. 2. Examples of attention guidance cues in 360-degree videos.(A) and (B) are Action Units created by Tong et al. [76], using gestures (here, highlighted with yellow boxes) by characters in the scene to guide the viewer's attention.

Fig. 4 .
Fig. 4. The detailed capture plan, showing how the Generic segments, Detailed segments, and two in-depth Special segments were arranged around the six visiting groups.There were also welcome messages and ending messages captured and attached to the main content segments.

Fig. 5 .
Fig. 5.The behind-the-scenes photo and the top-down view of the capture site.(A) The capture work at Te Rau Aroha Marae in Bluff.The 360-degree camera and the storyteller are both in front of the whānau wall.(B) The camera, the storyteller, and the wall being talked about are properly arranged so the viewer can later see both the storyteller and the wall and follow the storyteller's gestures easily.

Fig. 6 .
Fig.6.The six group personas in the printouts and on the selection panel in the VR experience.(A) The printouts of the personas used both in the capture stage and the evaluation stage.Each printout shows a group photo of virtual visitors on top, representing that group persona.(B) A screenshot showing the persona selection panel.The virtual representation of the controller and the ray cast are also visible in this screenshot, as the white stick and the red ray, respectively.Although the screenshot was taken from the PC running the VR system, it represents the first-person view of the VR viewer.

Fig. 7 .
Fig. 7. Explanation of the branching structure for persona groups 1 to 4. (A) A flow chart showing how a viewer could go through the experience with a branching point in the middle, between the Generic and Detailed segments.Persona 2 was chosen as an example in this case.(B) A screenshot of the panel between the Generic and Detailed segments, for group personas 1 to 4. Although the screenshot was taken from the PC running the VR system, it represents the first-person view of the VR viewer.The same choices are also marked on the branches in (A).

Table 1 .
Six Group Personas, Final Version Groups 5 and 6 are the ones that the storytellers commented they would treat as special cases.To these groups, the storytellers would prepare content suitable for them in advance.For example, the story for Group 5 would very likely be presented in full Te Reo Māori (Māori language). *

Table 2 .
(Continued)The Intention of Each Visiting Group

Table 3 .
Summary of Storytelling Experience Provided to Different Visiting Groups