My initial target for Dust3D has been achieved, I can use it myself to finish a 3d model in minutes which could took me half day while using traditional 3d softwares, as a newbie modeler.
I made this software mostly for myself to speed up the game assets making. Now, I can easily make model and colorize it, and got the model auto-rigged, the next step is to speed up the short animation clips making, such as walk, run, dead, attack... and so on.
Just as the modeling process, I want make this animation progress’s workload as small scale as possible, then I can finish the whole progress without outsourcing it, say, as a indie game developer.
Currently, I have some initial plans,
1. Based on some physics simulation or other algorithms, auto generate skeleton animations.
2. Use webcam to capture video, and analyze the motion from video by using OpenCV or Tensorflow.
I prefer the webcam method, because I can easily do it myself, if I want some weird animation, especially for master characters, and maybe it is hard to represent as a simulation or algorithm.
This animation making should be very simple, no facial or muscles motion capture, just coarse motion, walk, run, dead, attack, for low poly model.
Just as I post my modeling software initial plan in 2016 from scratch,
I have shared my full progress here_, including how I back and forth in Qt/C++ choice, and how the Rust language is been used as the core algorithm implementation.
I wish this post is my new start of the game making journey, **please leave a comment on reddit website to share what's the best idea to achieve the monster animation task, thanks.**
(I will append the reddit link here when I made this post public on reddit)
