Puppeteering

See Also

Render Roadmap

Objectives

• Further the ability to communicate in Second Life by enabling residents to 'puppeteer' their avatars
• Create a Physical Avatar code layer which enables the addition of ragdoll physics and inverse kinematics to second life avatars

Overview

Puppeteering is real-time, user-manipulated animation of avatar joints.

It refers to nascent avatar expression technology. As with a real puppet, people will be able to manipulate the joints of their avatar, and others will be able to see these changes. The initial release will be rather humble -- the ability to temporarily adjust the position of individual joints. Enabling this is the deeper infrastructure of the Physical Avatar which, in subsequent evolutions, will allow (besides Puppeteering) a variety of new tools for real-time gesturing, inverse kinematics, rag-doll physics, in-world animation building, and scripted animation.

First Release: Expressive Puppeteering

• The first release is about laying the foundation for the future of avatar manipulation. Key is the development of the Physical Avatar, a code layer which models an avatar as physically-based joint balls connected by springs, and allows influence by external forces (from UI and from world). As this model is being manipulated, it continually converts back to the standard avatar representation so it can be rendered.
• The visible component of this is a simple UI enabling the momentary manipulation of avatar body parts. Initially this will be triggered by holding the CTRL key down. Mouse cursor proximity will cause joints to highlight, and a click and drag allows the positioning of that joint. As long as the CTRL key is held down the pose will be fixed, allowing the adjustment of several joints. All these manipulations will be visible to other residents. Movement, or the ESC key, will end puppeteering and revert the avatar back to original animation.
• A base set of normal joint constraints should be available to users that allow for realistic movement
• You will be able to save a puppeteering pose as an animation asset, for later use as a pose or in a gesture.

Future Releases

The Physical Avatar layer enables the development of several Grid-changing technologies.

• Real-time gesturing: An advanced UI, perhaps using keyboard chording or even alternate input devices to produce complicated avatar expression. Real time dancing.
• IK and rag-doll physics: Residents could walk hand in hand with realistic motion. Avatars tumble realistically when falling (or when shot, pushed).
  • Play awesomely with Havok. When bumped, ragdoll friction relative to impact force. :)
• In-world animation tools: Residents can build and save their own gestures, animations in world. No need for external modeling applications, exporting, and uploading.
• Scripted animations: Additions to LSL would allow the scripted adjustment of the Physical Avatar joint tree.
• Procedural animations: Ability to blend puppeteered actions into running animations including previously puppeteered loops.
• Touch a surface: Ride on the 'texture surface' code to move a joint to a surface on a nearby object, pulling IK chain along.

Project details

Puppeteering is a project within Studio Shiny. Cube Linden, Aura Linden, and Ventrella Linden are designing and implementing puppeteering. Periapse Linden is assisting with project management.

Over the past year puppeteering has had several starts and stops. The current team, however, has moved the project forward dramatically.

Roadmap

Tasks and milestones remaining:

• Fix bugs from latest QA pass on basic functionality
• Implement UI suggestions from Resident Experience team
• Implement code reworking for animation blending
• Iterate through another QA pass
• Validate server code for Main Grid safety
• Push puppeteering to several new test regions of the Main Grid (Linden access only)
• Verify Main Grid release safe for residents, create puppeteering beta viewer
• Open test regions up for residents. You will need to download a special viewer.
• Iterate on resident feedback from beta regions before planning any gridwide release