Category Archives: CapstoneProposal

Matthew Kellogg – SkyComics Checkpoint

Tweetable: An editor and viewer for immersive 3D environments for the purpose of storytelling.

For my project I am making a two part system for creating and viewing an immersive 3D environments composed of fully surrounding images (Skybox) and directional sounds. I have implemented the following features:

  • Scene Drawing Helper
    • In browser use
    • Draw on panable and zoomable spherical scene view
    • Export to high resolution cube map / skybox format
    • A gui for defining drawing guides, importing models, and controlling the view

I have promised with these features the following list of features. However, due to time constraints I will not be implementing features with slashes through them.

  • Environment Viewer
    • Web viewable
    • Clickable regions
    • Intuitive click and drag and pinch to zoom controls
    • Fading between scenes
    • Play directional sounds and background music
  • Scene Drawing Helper
    • Import models for shape definitions (Instead only simple guides will be used)
    • Import cube map (to add more stuff or make notes on drawn parts)
    • Layers support (Instead only adding to images will be supported)
    • Click region definition
    • JSON output of click regions and any other metadata
    • Scene linking tools (This will be done using JSON)
    • Full comic file export (an archive of metadata and media that can be opened in the viewer)(This will be a folder with all the images and a JSON file explaining the click regions and scene links.)

I will try to follow the following schedule.

  • Today I will
    • Click regions definition
    • Import cube map
  • By 4/23 I will
    • JSON to describe regions and actions
    • Play sounds
    • Viewer mode with controls
  • By 4/27
    • Polish everything.
    • Make an example.

This is the current look of my project. The background is a skybox provided with ThreeJS.



It can be viewed here.

John Mars

21 Apr 2015

Overall, things are going pretty smoothly. The hard parts have been mostly solved, namely the browser visualization, making manifold meshed, and creating 3D objects. The rest of the stuff is pretty much just plug-and-play busy work. It looks like there’s a lot of unchecked ToDos below, but they’re all connected, and should all go simultaneously (that’s poor list-making on my part, sorry).

  • Tiles need to be completely downloaded
  • Solution must be found for creating manifold meshes
  • Develop method for turning single-face meshes into solid meshes Mostly completed — mesh edges have been found, now I just need to make faces between them and a base
  • Build website locally In progress
    • Develop website backend In Progress
      • Determine website framework Node.js
      • Embed openFrameworks on server Using Three.js instead
      • Implement a map API for displaying available cities and picking them In Progress — City-level picker is almost done; decisions need to be made on displaying which locations are available
      • Implement Shapeways SDK
    • Develop website frontend In Progress; its convolved with the backend
  • Deploy DigitalOcean droplet
  • Purchase Domain
  • Deploy website on server
  • Test print from Shapeways URGENT
  • Open Shapeways store
  • Publicity
  • Profit

Thomas Langerak -Capstone Update

To say it the “Dutch way” (really direct and without reservation); the capstone project is totally not going as planned. The first two weeks went quite well, I made some quick mockups and found a chess engine I could work with. Than disaster struck, a computer crash and mockups mostly gone. Luckily I had backedup most things, yet a few things were lost.

About two weeks back I decided to go for the tangible route. I experimented with a raspberryPi with a camera and OpenCV. Unfortunately I did not make the deadline for myself today and therefore believe a tangible solution is highly unlikely to be achieved on time. Just for me working the first time with linux, a raspberrypi and CV proved to much of a burden. Therefore I have decided to pull the plug and focus fully on the digital solution. Let me make clear that I would definitely would have prefered tangibility, but I also have to be realistic. Next to this I want to make clear that though I am almost at the same stage as two weeks back, I have learned a tremendous amount during this time period. I definitely gained knowledge about technologies I only had heared of. I am certainly planning on learning raspberryPi for the future, as I definitely see potential in the medium.

So with about a week to go, what still needs to be done?
On a conceptual level I need to find something worth unlocking.
I had decided to unlock my computer, as this is highly personal. Yet now I need it to actually run my work.

On a making level I need to make a visualization in processing.
I actually need to unlock something.


20 Apr 2015

In summarized the conceptual idea of remapping in this prezi presentation. In this post I will update the current state of the capstone proposal in practical terms.

Basically, I spent my last 2 weeks dealing with data scraping and analysis with processing. Right now I have not only the basic attributes of the individual footprints (scale, area, perimeter, etc.) but also most of the analysis were implemented. The buildings are embedded with shape metrics such as dispersion, “rectangularity” and ortogonality. Besides, I also collected the neighborhood average income (by zipcode) and the elevation of the building. Along the next week I intend to implement similarity, in order to allow the buildings to find similar buildings.

The basic motion and collision functions are also defined, so the footprints can move around an dispute the space. In the video below, there is a simple example of their animation.

The next step is to use the attributes and analysis as the drivers of the movement, enabling a goal-oriented remapping. These maps will indicate spatially one or two criteria such as a similarity clustering or a income x size mapping. I hope that using the footprints as agents will not only enable these remapping but also will create a very interesting battle for space, emphasizing the differences and similarities of the city map.

Finally, I will try to make it more efficient, in order to enable remapping thousand of buildings. Probably, I will try to change collision detection – that right now is a brute force algorithm.


20 Apr 2015

Since the proposal, we made a big mechanical change. Instead of making the player navigate through the streets to the school each day, we are going to have the player text people (parents and friends) while the background whizzes by. This way we can convey more narrative and make the subtle changes of the background, well, more subtle. This can be seen as a step down in interactivity by getting further away from mechanics and instead focusing more on the narrative.

Currently, we have all movement implemented. The dialogue system is also in place, but the UI is not yet done to be displayed in the way we want yet. Some art assets are also in place, although not enough to have variety, it is enough to get by. The animations for the girl is also set. I am currently working on the dialogue.