The making of MIRAI
"MIRAI" 100 years ahead in time, we can see the old garage of a mechanic where he keeps his old friend, a robot model T-LOW 91. The 4th WW Robots Agreement has been just broken, now robots and nuclear weaponry are allowed again. Robots will have to meet up for what the future may bring.
Update - 10 Mar 2019
I made "MIRAI" as part of my final project for Vancouver Film School. The main challenge was to create this huge environment along with the character (and at the same time design it) in a really short period of 6 months. The asset count went over 60 without counting the robot (T-LOW 91) so I tried to keep my geo and my textures as light as I could.
The idea started with the story, that one leaded to the creation of many concept proposals for the robot. It had a hard time doing it but the most complex part after having figured out the visual design of the robot was the "functionality". I thought "it was easy" making something just look good, but making it felt real was my approach, for that I wanted to give movement to it, therefore animating it. But I was wrong it wasn't easy to make something look cool and it was even harder to make functional something I designed. For that I had to make some test and researching of other real and fictional robots and types of joints in order to make it feel believable. The movement and the proportions of the original design were a challenge so I made some rigging test with a proxy with the dimensions of the robot.
(You can jump this first paragraph is just personal stuff...) The first time I pictured the idea in my head sounded pretty cool but then I notice I didn't have the slightest idea of how to do it. There were some assets that were a big challenge and I am really grateful with my teachers, mentors and classmates for helping me with most of my questions. But sometimes one must find their own answers when no one around you can help (deadlines are a monster).
Dumpling- (No dumplings were wasted in the making of this film...) Photogrammetry... don't do photogrammetry, you can make it faster just modeling them. Yes but I wanted to learn how to do it! So I did it. After taking MANY pictures and failing "n" times trying to scan it I noticed there is no better tool for an artist than YouTube. 4-5 tutorials of 30 minutes, 1 session of taking pictures, a couple of hours with Photoscan, some cleanup and retopo in Zbrush and Maya, trackers removal in Substance Painter and DONE! Never under estimate the power of tutorials and self-teaching.
Third Hand Tool- (Yes, the thing with the tweezers...) After making a considerable amount of assets I noticed that my scene was pretty heavy (even using references), so even if it is not what I think was the best decition I broke each asset into two parts: what needs to be smoothed and what does not need to be smoothed. How close it is going to be from the camera and how it is going to be affected by the light. That reduced my polycount and my time spent in each asset considerably. You can barely see what it is smoothed in this asset and what is not.
The Oscilloscope- (This one was fun...) I didn't know how to do it so I just thought for a second just to remove it. But I really wanted to do it only because I thought it was going to be fun (even that I wasn't pretty sure if it was going to work out inside the scene). After that 3 bad tutorials, 2 decent and one adequate, I could do it. And I think that one is one of my strongest assets. Oh yes. I did it with a image sequence as a texture (you can see the animation made in After Effects in the video). There was a lot of researching for the look of the screen (but it was FUN).
Hero assets (renders)
These are just some hero renders I did when I was tired. These renders gave me a momentary feeling of satisfaction having achieved something small.
Thank you for the time you spent reading this nonsense. I apologize if I made you feel dizzy and I'm glad in the remote case you enjoyed my project.