Collaborative Creativity in TikTok Music Duets

Title of the Paper

Collaborative Creativity in TikTok Music Duets

Paper Information

• Subject Area: Human-Computer Interaction (HCI), Distributed Creativity, Music Information Retrieval
• Keywords: TikTok, Duet, Distributed Creativity, Collaborative Content Creation, Music Information Retrieval (MIR), Creative Analysis, Social Interaction, Digital Creative Platforms, Cultural Evolution, Music Tagging

Research Background and Problem

• TikTok's Duet feature enables users to engage in open-ended collaboration by sharing and modifying existing video content, becoming a significant form of digital collaboration.
• Traditional collaboration typically involves defined participant groups and clear outcome goals, whereas TikTok's Duet feature introduces decentralized and asynchronous creative collaboration.
• Key questions addressed by the authors:
  • Does the Duet feature support broader decentralized collaboration?
  • How can "distributed creativity" and content evolution on the Duet platform be measured?
• Significance of the Research:
  • Digital tools are increasingly becoming mediums for creative collaboration, making it crucial to understand their impact on societal and cultural creativity.
  • As a rapidly growing social platform, TikTok's unique collaborative features make it an ideal subject for studying large-scale distributed creativity.
• Motivation and Related Work:
  • Previous research on collaborative creativity has primarily focused on synchronous or small-scale collaboration.
  • Theories of distributed creativity provide a framework for analyzing more complex decentralized digital interactions.
  • Existing studies lack in-depth analysis of indirect collaboration processes on platforms like TikTok.

Solution

• Methods and Innovations:
  • Proposed an analytical framework to quantify the evolution of creative works in collaborative processes by tracking "digital traces" and musical features in TikTok Duet videos.
  • Leveraged Music Information Retrieval (MIR) techniques to extract key features from video audio and analyze content changes using deep learning algorithms for automatic music tagging.
  • Combined tree structures and temporal distribution models to analyze relationships between videos and user interaction patterns.
• Implementation Steps and Techniques:
  • Data Collection: Gathered music-related videos with the "Duet" tag from TikTok, generating tree structures to capture video inheritance relationships.
  • Content Analysis: Used convolutional neural networks for music semantic tagging to quantify audio feature changes between videos.
  • Temporal Data Analysis: Analyzed time gaps between parent and child video postings to validate the dynamics of collaborative behavior.
  • Case Studies: Examined tree structures, paths, and content changes in three different Duet chains ("The Wellerman," "Misty Mountains," "Scream It Out").

Research Outcomes

• Specific Findings:
  • Tree structures revealed that Duet chains allow creative works to evolve along multiple parallel paths, fostering diversity and innovation.
  • Layered versioning of digital video content decentralizes collaboration, accommodating a large number of participants.
  • Musical feature analysis demonstrated that each path explores new feature spaces while inheriting elements of the original content, showcasing the dynamics of distributed creativity.
  • Temporal distribution analysis indicated "burstiness" in parent-child video posting intervals, with sustained dynamics in later collaborations.
• Comparison with Existing Solutions:
  • Unlike traditional collaboration tools, TikTok Duet supports nonlinear, decentralized development while preserving every version of the work.
  • The structure of creative chains enables users to work in parallel rather than converging toward a single final goal, enhancing diversity.
• Experimental/Evaluation Results:
  • Nearly all Duet chains exhibited distinct evolutionary paths, though the gradient of changes varied significantly across creative paths.
  • Collective attention decay over time was associated with the dynamic nature of Duet content.
• Limitations and Future Directions:
  • Data was exclusively sourced from TikTok, leaving collaborative mechanisms on other social platforms or in other artistic domains unexplored.
  • The study did not examine the influence of users' social networks or TikTok's recommendation algorithms on collaborative dissemination.
  • Future research could expand to distributed creativity in visual arts or literary domains.

Conclusion and Implications

This study reveals the operational patterns of distributed creativity on the TikTok platform, including how the openness of collaboration influences cultural diversity and creative development. The findings suggest that designing digital tools to support decentralized collaboration can play a significant role in fostering large-scale, diverse creativity. This has important implications for developing new creative platforms and enriching the digital cultural ecosystem.