Design Challenge: Design, build and exhibit an interactive window display spanning two windows, based on the theme of time
Gal Nissim: Body Scan, Fabrication, Mixamo Design, Max Patch, Unreal Engine Integration of 3D Avatars, Scenic Design
Lindsey Johson: Body Scan, Chat App Designer and Developer, Scenic Design
Angela Perrone: Body Scan, Fabrication, TV stand design, 3D Mesh Work: merging meshes, texture baking & mesh & texture editing,
Michelle Hessel: Body Scan, Fabrication, Unreal Engine Scenic Design, Unreal Engine Integration of 3D Avatars, Scenic Design
Project & Concept Overview
Our concept and design, Current Times is a social commentary exploring the current state of human communication, where we exist in a physical world yet are trapped in the digital world, only to communicate to the outside through digital devices.
Window 1: Physical World: This window features a live person, sitting on their computer. They do not communicate or acknowledge the out-of-the-window world, as they are communicating with this outside only through the digital world, found in window two. The actor, portrayed by either myself, Michelle Hessel, Gal Nissim or Lindsey Johnson, controls the avatar and chat in the second window. The aesthetic is a bleached canvas of a bedroom, where everything is white except the person. When a live person is not in the room, a sleeping dummy version (color) is staged.
Window 2: Digital World: This window features two large screens against a white background. Screen one hosts an avatar of the live person in the window. Screen two features a chat box, where visitors can log in to http://currenttimes.us/ to chat in real time, with the avatar. The avatars can react via chat and/or with actions that showcase everything from emotions to dancing. The actions and chat are controlled by the actor in the first window. When the live person is not present, an avatar of the dummy is viewable on screen sleeping on a cloud in the sky. Icons of the avatars social media also appear on clouds, showing the numbers of missed messages and posts that accumulate over time.
I took on the role of merging 3D body meshes together, baking textures, and exporting the 3D Avatars before they were animated in Mixamo and Unreal Engine. Merging multiple meshes together is simple, however merging textures of multiple objects is a complicated recipe. Thanks to the help of Quest Kennelly, resident Blender wizard, I was able to learn and master this process. This was my journey.
Gathering the Meshes
To create the avatars we needed 3D body scans of each individual who would be sitting in the window plus the avatar. To accomplish this we used
Skanect (mac software)
Structure (ipad App for scanning)
Preparing & Merging the Meshes
To get the highest quality resolution, we needed to capture two separate scans: one of just the face, and one of the entire body. This meant that the head and body meshes needed to be merged together into one mesh.
I was able to bring the two meshes into Blender, line them up, and joined the meshes by pressing CTRL-J.
HOWEVER, blending meshes does not blend textures. To make the textures match, specifically for the new blended meshes, I needed to bake the textures onto the new mesh, an involved process that I will outline in detail below.
The Process of Baking Textures
Baking the textures took a lot of investigation and testing in Blender. Quest Kennelly, blender wizard, was a huge help in teaching me blender and how to accomplish this task. This is an in-depth tutorial that I put together that dives into the step-by-step process used to create our final 3D models.
The textures baked nicely but not perfectly. So once the texture was baked and matching the merged mesh, I went through the entire surface, filling in any missing missing holes.
This was a bit tedious but worth it to make the meshes as true-to-life as possible. Viewing the UV map of the texture while also looking at the mesh in edit mode, I was able to highlight the empty or incorrect vertices in the edit window, and see their location highlighted in the UV map. I then moved the highlighted vertices to the appropriate, corresponding color patches in the UV map.
[INSERT IMAGES OF FINAL EXPORTS]
Other Noteworthy Info for Blender
Things learned that one needs to consider:
scale when importing it into Unity or Unreal
One unit (1uu) in Unreal is equal to 1 cm. So when scaling your object
This is a great tutorial for sizing.
Designing the Screen Mounts
Two large screens were needed in the digital window, but with a shallow depth and no wall to mount on to, placing the screens posed a challenge. Consulting with Marlon Evans and Shir David‘s mounts in other windows, I came up with these mount designs for the two, different sized and weighted screens:
Setting Up & Connecting
The installation was up from November 11th through 13th at the NYU Skirball Center for the Performing Arts, located at the corner of 3rd Street and LaGuardia Place.
Throughout the process we faced various challenges that we were able to solve Our first issue was with the quality of the body scans. The reason for merging meshes and baking textures was born from the issue of not getting the proper definition and high quality in the facial features. We decided to use two body scans – one of the head and one of the body, and the merge the two together. Figuring out how to bake textures to merged meshes was another hurdle that was tackled with the help of Quest Kennelly.
Another big challenge was getting screen mounts to fit into the designated window depth without access to a wall to mount on. We designed new mounts to fit the space.
On the day of the opening, we had an issue with connectivity. The internet was required for running the chat app and remote desktop, which worked as long as the lap top was not in the window. As soon as it crossed the threshold into the window, the internet disconnected. We did some trouble shooting and found that the macbook model that we were using was too old and the specs of that model were preventing the internet from working. We traded out laptops and had a smooth run.