VR-Jam Postmortem

The Game - “Scalar”

  • Scale using an artifact!
  • Solve Puzzles!

The VR-Jam Competition

  • 3 Weeks
    • Week 1 - deliver a screenshot
    • Week 2 - deliver a video
    • Week 3 - deliver the game
  • Panel of judges pick the winners

The Team

  • 4 members
    • People came and went
    • Full-time jobs - little time
    • No experience making games
    • No skilled artists

Week 1

Week 1

  • Brainstorming
  • Picking tools
  • Cost us a week of dev time

Ideas

  • Puzzle/Adventure game for the slower pace
  • Manipulation of the small environment
  • Inventory items scale with the player

Scale

  • Had to pick scales for player to inhabit
  • Affects asset design, puzzle layouts
  • Ratio had to be just right
  • Settled on 1.0x and 0.1x

Unity

  • Picked Unity for the game engine
  • Free 30 day Pro trial
  • UE3 integration felt unfinished

Bitbucket

  • Private git repository on BitBucket
  • Free for 5 members
  • Pro version of Unity required for version control
  • Diffing Unity projects still a huge PITA

Week 2

Week 2

  • Asset Workflow
  • Basic Gameplay

Tiles for rapid prototyping

  • Only need few composable assets
  • Had to work for normal and small player
  • Edit -> Snap Settings in Unity for grid resolution

Blender

CGTextures

  • Royalty free images
  • Great for textures
  • Has images from Cambodia!

Asset Workflow

  • Rough cuboid shape for collision mesh
  • Create prefab in Unity for prototyping
  • Refine model, create UV maps, and texture
  • Update prefab in Unity

Programmable Animation

  • Animating programmatically over a number of frames
  • Unity coroutines. See C# yield keyword
  • E.g. Open a big stone slab after a key is inserted

First Puzzle

Week 3

Week 3

  • More assets
  • Puzzles
  • Final level layout

Problem: Hierarchical prefabs

  • Unity does not have hierarchical prefabs
  • Prefab status lost of the smaller pieces of large prefab
  • Resorted to constructing a “prototypical wall” and copy pasting it

Problem: Performance

  • Tiles updateable and composable, but small
  • A LOT of them on-screen
  • Performance left much to be desired.

Problem: Performance

  • Tiles great for prototyping
  • Should eventually be “baked”
  • A tool or plugin would be ideal

Problem: Merging Unity Scenes

  • Duplicate the current scene
  • Create new room in an unoccupied area
  • Commit this new scene as a separate file
  • Another person finally merges all the rooms into final level

Crunch time

  • Severe crunch-time during the last weekend
  • Sunday was a 14 hour long marathon
  • Didn’t want to look at it any more! :P

General Advice

Set the smallest next goal

Do the simplest thing to make a step forward. It doesn’t mean it has to be done sloppily - just don’t be overly ambitious.

Know when to limit scope

If implementing something is taking too long and is not critical to getting your idea across, just drop it.

Delicate Balance

Managing a team has to be a delicate balance between democracy and dictatorship.

Oculus/VR advice

Educate

Make sure everyone on your team is on the same page about the design space for VR content.

Anchor

When exploiting new VR phenomena, give the player some anchor so they are not discomforted/hurt.

Adapt

Many existing graphics techniques don’t work in VR. Adjust your asset creation accordingly.

Simplify

Keep assets simple for fast/consistent framerates.

Questions?

Author : Alexander Kondratskiy

Email : akondratskiy﹫siggraph·org

Website : “KholdCode” - kholdstare.github.io