Distributed Rendering and Collaborative User Navigation- and Scene Manipulation

Christian Kehl

University of Amsterdam

and UniResearch AS CIPR Bergen

Introduction: Virtual Environments for Decision Making

• Goal: help decision-making process with (3D) Visualizations and advanced interaction capabilities

• Challenges emerge from user feedback

• Generally: – Interactive discussion environment

– Medium size of participants (3 – 50) (+moderator)

– Heterogeneous output devices and interaction demands

3Di discussion with 7 participants (1 moderator), using multi-touch screen

Introduction: Virtual Environments for Decision Making

• Collaborating discipline: Flood Protection Planning

• Starting point: Immersive, single-user VR framework [DeHaan 2009]

VRmeer framework, in short

Introduction: Virtual Environments for Decision Making

• Challenges from Feedback: – Support rendering on flexible, large-scale rendering projection systems

– Integrative navigation for several, possibly remote users (video conference-like situations)

– On-spot, interactive manipulation (annotation) of massive datasets

Large-Scale Projection System (4.5m x 2.5m)

Multi-Modal, Multi-User Navigation

Interactive Annotation

General Approach

• Extension of [DeHaan 2009] to a network-based VE-framework [in this paper] – Remote, Distributed Rendering

– Remote, adaptive, collaborative Navigation

– Remote, collaborative Editing (Manipulation

• Further research: – Remote, Interactive Simulation and Processing

– Adaptive Rendering for arbitrary-sized output devices (e.g. mobile phones)

Idea of Distributed Graphics

Projection-aware Distributed Rendering

• Target output devices: – individual projectors (theatric setup)

– screen walls

– PowerWall

– mainly stereo-projection

• Server-Client configuration

• Synchronized Rendering via timestamp or frame number

• Adaptive projection for planar- and

spherical output environments

Designed, projection-aware network architecture

Collaborative User Navigation

• Particular challenge: unify device capabilities (similar to [Taylor2001])

• Framework of Function- and Device Abstraction

• Navigation synchronization on (dis-) appearance of clients

• Difference-coded view matrices, interpreted on rendering master

Collaborative Scene Manipulation

• Challenge: Modify Out-of-Core loaded datasets in real-time

• Server with high-resolution, Out-of-Core dataset

• Client draws simplified geometry

• Simplification method according to data structure (surface-/volume mesh, point set)

Render Server Render Client

Collaborative Scene Manipulation

• Modification idea based on [Kehl2013]

• Extension to 3D modification primitives

• Per – Vertex inclusion test via implicit function

Collaborative Scene Manipulation

Sample Rendering

Inclusion test via implicit function Box: −𝑎 < 𝑃′

𝑥 < 𝑎

−𝑏 < 𝑃′𝑦 < 𝑏

−𝑑 < 𝑃′𝑧 < 𝑑

Inclusion test idea via implicit function for prism

Technical Results

• Distributed Rendering: – Synchronization by timestamp: larger distortions, high framerates

– Synchronization by frame number: reduced distortions, low framerates

• Collaborative Navigation: – Framework implemented for various joysticks, spacemouse and keyboard

• Collaborative Modification: – Sphere- and OOBB-marker modification interactive

– Prism-based modification to complex for real-time execution (currently)

Distributed Rendering result for independent steering of a 3-screen setup; here: non-stereo

Technical Results

Impact on Decision-Making

• Distributed Rendering allows various rendering setups, actively in use in 4 projects

• Collaborative Navigation gains positive feedback; prime advantage: OS-interoperability

• Coll. Navigation and Modification allows interactive communication in meetings and workshops for Flood Protection Planning

• Techniques also applicable outside flood domain (e.g. seismic exploration, pre-operative planning)

Future Work

• Touch-based interaction poses challenges for distributed Rendering

• Issue with Modification: Simplification needed – > Out-of-Core loading for clients advisable

• Improve OOBB- and Prism-based an notations

• Markers in spatial hierarchies, as [Kehl2013]

• Remote, Interactive Simulation and Processing

• Adaptive Rendering for arbitrary-sized output devices (e.g. mobile phones)

Acknowledgements

References:

[DeHaan2009] G. de Haan, Techniques and Architectures for 3D Interaction, 2009: Delft University of Technology, Delft.

[Kehl2013] C. Kehl, T. Tutenel and E. Eisemann, “Smooth, Interactive Rendering Techniques on Large- Scale, Geospatial Data in Flood Visualisation,” in Information and Communication Technologie (ICT) Open, Eindhoven, The Netherlands, 2013.

[Taylor2001] R. M. Taylor II, T. C. Hudson, A. Seeger, H. Weber, J. Juliano and A. T. Helser, “VRPN: A Device-independent, Network-transparent VR Peripheral System,” in Proceedings of the ACM Symposium on Virtual Reality Software and Technology, New York, NY, USA, 2001.

• CGV group (TU Delft): equipment support, particularly for Dist. Rendering

• Donald Smits Institute (Groningen): collaboration on Dist. Rendering

• dr. ir. Gerwin de Haan: explanation on initial implementation

• Prof. Dr. rer. nat. Herbert Litschke and Dr. Simon J. Buckley: short-term review