GroundLink: Exploring How Contextual Meeting Snippets Can Close Common Ground Gaps in Editing 3D Scenes for Virtual Production

Paper Title

GroundLink: Exploring How Contextual Meeting Snippets Can Close Common Ground Gaps in Editing 3D Scenes for Virtual Production

Publication Info

• Topic area: Enhancing collaboration in virtual production through meeting-derived knowledge integration.
• Keywords: Virtual production, common ground, 3D scene editing, Unity, feedforward, in-situ information, collaboration tools, meeting-derived knowledge, onboarding, transactive memory systems.

Background and Problem

• Problem / challenge: Virtual production professionals face challenges in accessing tacit knowledge and creative intent from past meetings, leading to gaps in common ground, especially for newcomers. Existing tools fail to integrate meeting-derived knowledge directly into 3D editing workflows.
• Significance: Addressing these gaps can improve collaboration, reduce errors, and enhance efficiency in virtual production workflows, which are increasingly remote and distributed.
• Motivation and related work: Prior research highlights the importance of common ground (shared mental models and transactive memory systems) in collaboration. Existing tools provide meeting summaries or annotations but fail to embed this knowledge directly into the work environment. This paper aims to bridge this gap by integrating meeting-derived knowledge into 3D editing tools.

Solution

• Proposed approach: GroundLink, a Unity add-on that surfaces meeting-derived knowledge within the 3D editing environment to support common ground formation.
• Novelty:
  1. In-situ feedforward mechanisms that proactively surface constraints and decisions during editing.
  2. Cross-modal synchronization linking meeting snippets, decisions, and 3D scene elements.
  3. A meeting knowledge dashboard that consolidates decisions, comments, and meeting recordings with AI-generated summaries.
  4. Empirical validation of the system’s impact on common ground formation and task confidence.
• Procedure and key techniques:
  1. Meeting-derived knowledge is preprocessed into a JSON format linking decisions to 3D scene elements.
  2. The Unity add-on provides visual feedforward cues (e.g., encouragement/restriction zones) and inspector feedback for styling compliance.
  3. A dashboard synchronizes meeting snippets, summaries, and 3D scene elements, enabling bidirectional navigation.
  4. AI summaries allow scoped queries about specific objects or decisions.

Results

• Concrete findings:
  • Participants using GroundLink reported higher perceived shared mental models (SMM) and transactive memory systems (TMS) compared to a baseline (Microsoft Copilot).
  • Significant improvements in perceived confidence (mean confidence rating: 4.08 for GroundLink vs. 2.42 for baseline, p = 0.014).
  • Reduced perceived conflict with prior decisions (mean score: 3.08 for GroundLink vs. 2.17 for baseline, p = 0.062).
• Advantage over baselines:
  • GroundLink provided better contextual understanding, reduced task difficulty, and enabled more confident edits compared to the baseline tool.
  • Proactive, in-situ feedforward and cross-modal synchronization were highlighted as key differentiators.
• Experiments / evaluation:
  • Comparative user study (N=12) with 3D editing tasks in Unity, comparing GroundLink to Microsoft Copilot.
  • Expert evaluation (N=5) with virtual production professionals to assess domain-specific applicability.
  • Metrics included NASA-TLX, UES-SF, task completion rates, and qualitative feedback.
• Limitations and future work:
  • Manual preprocessing of meeting data limits scalability.
  • Visual clutter and information overload were noted as challenges.
  • Future work should focus on automating the authoring process, deploying GroundLink in real-world settings, and exploring its generalizability to other domains.

Summary

GroundLink is a Unity add-on designed to bridge common ground gaps in virtual production by embedding meeting-derived knowledge directly into 3D editing workflows. Through in-situ feedforward, cross-modal synchronization, and an integrated dashboard, it helps users access and apply past decisions and constraints. Comparative studies and expert evaluations demonstrated its potential to improve shared understanding, task confidence, and ease of editing. While challenges like visual clutter and reliance on manual preprocessing remain, GroundLink offers a promising approach to enhancing collaboration in virtual production and other creative domains.