VTuber Project: Creating Bora
I’ve decided that I wanted to do something different for my Graduation Project— something fresh and unique. Most importantly, I wanted it to be something extra challenging that I’d have fun figuring out to distract myself from the impending doom of the world highlighted by this pandemic.
With all of that combined along with my love for designing characters, I arrived at a project idea that I thought would be all of that and more: to create a 2D virtual avatar, dubbed as ‘VTuber’ or ‘Virtual Youtuber’. To put it simply, they are avatars that can be synchronously moved by facial recognition programs. By warping or distorting the flat surfaces, the 2D avatars can feel more mobile and three-dimensional.
In order to create a VTuber, I would need to learn (or rather, teach myself) a whole new program altogether, the program being Live2D. Even though this project may seem like it does not correlate directly to the field of Illustration, a lot of the skills, techniques and theories I’ve learnt during my time at university were applicable to this project. Trying to create a moveable 2D character, and figuring out a fun and creative design was an enjoyable challenge.
VTubers are often animated by a specific streamer, and a persona is adopted to fit the character. Thus, I began with a few character designs that were based on different parts of my own personality.
After gaining feedback from professors and fellow peers, I proceeded with the 4th design and continued to test out various colors and design changes.
Again, I received feedback, applying changes and edits accordingly and finally went with the 5th design, as you will see below in the final character sheet of the avatar, now dubbed at this stage as Bora.
You may notice in my final design that I made small edits to Bora’s design in order to be more adaptable to the new process and program I was trying to learn. Below, you will see the final character art, and the texture atlas of the illustration, showing how individual pieces of the avatar had to be separated into different layers.
Finally came the fun (or perhaps excruciatingly difficult) part of trying to learn how to rig a 2D illustration into moving synchronously with facial recognition. As my professors were just as new to the Live2D program, most of my questions and confusions had to be answered via the internet. Thankfully, the internet is so very resourceful that I was able to find what I needed. So began my process of applying the rigs to each part of the avatar. I created a short summary video of the steps taken to creating the final VTuber product of Bora. So, grab a snack, sit back, and enjoy!
During my process of learning how to rig Bora, I continued illustrating on the side. Below is a keyframe illustration I created that depicts Bora at the hat stall, trying to find a hat to wear. I thought it’d be interesting to show my VTuber character within a fictional setting to further ground her background and personality.
Thank you for reading until the end and I hope you enjoyed seeing the process and final result as much as I did!