Week 2 - Previs
This week we were focused on applying the feedback from the mentors after our pitch, and creating cg previs for each of the shots. We got great feedback on the pitch, the mentors liked our concept and most of our plan. Some of the feedback we got was to simplify one of our shots (the flower shot), think about the transitions between shots and flow of the camera to make sure it feels natural and smooth, consider reordering shots, and to think carefully about the colors and light in each shot. For this week, we focused mainly on simplifying the flower shot, and working with the cameras and motion to try and make the shots blend together seamlessly, as well as creating FX tests to use as previs.
Simplifying the flower shot:​
In comparison to our other shots, the flower scene was very complicated. Gracie, who will be doing the effects for this shot, did a lot of conceptualization of simpler flower effects that we could use instead, but they all felt either too complicated, or not very exciting. After running through many different options, we decided to instead switch to a different concept. Instead of flowers, this shot will take place in the clouds. The ring will fall through the stormy clouds into a bright, sunny scene with soft white clouds, connecting to the rings ability to track stress levels.
Gemini AI generated concepts for the flower shot
Our contact sheet for the new cloud shot
Camera work:
​
In our initial plans for this project, we considered making it a one shot project, where the ring travels through different environments with no cuts. This week, in talking more about cameras, we shut that idea down, because of the complications it would create with rendering, lighting, and scene set up. We are instead going to base our editing around other smart ring ads. In ads similar to ours, where the ring is very large on screen, the cuts feel natural because even though the environment and colors completely change between shots, the orientation and position of the ring are relatively similar, and the type of movement the ring is doing remains the same. If the ring is falling and spinning in a forest, it can cut to the ring falling and spinning through the water without feeling too broken up.
We are basing our camera, and our ring animation/simulation, off of this concept. The ring will, for the most part, be falling through each of the environments, which we hope will make the camera work feel smooth.
Screenshot from an initial camera test for the cloud shot
New reference photos of Oura rings for lighting and look dev
FX Testing:
​
Most of my time this week has been spent working with cloth FX for shot four. In this shot, the ring is going to fall from above and land in a moving bedsheet. The ring will roll around for a moment, and then the cloth will be pulled taut, launching the ring upwards to transition into the final shot.
I did a previous FX test for our pitch that showed the ring in the cloth, but it did not include the ring being launched up. This week I tested different ways to create the launching effect.
​
​
​
​
​
​
​
​
​
​
​
​
​
Issue: When the cloth has the settings to look the way we want it to, the ring isn't pushed into the air, because the cloth has too much stretch to it.
​
Troubleshooting: I ran a second simulation on the ring that just pushed it upwards, and then switched from the first simulation to the second simulation when the ring should be launched into the air, but the timing of the switch wasn't exactly perfect, and didn't feel natural.
I changed tactics and worked on changing the settings of the cloth (poly count and stretch amount) so that the cloth could actually launch the ring into the air. I managed to get a few different options that did launch the ring, but as expected, it changed the quality and look of the cloth too drastically to use.
​
Solution: My current plan to finish the FX previs is to use a combination of multiple passes. I'll render a cloth pass, simulating the cloth and the ring, with the cloth looking the way we want it to, and I'll also render a ring launch pass, where a stiffer cloth launches the ring upwards. These will then be combined, starting with the cloth and the ring from the cloth pass, and then when the cloth goes taut, switching just the ring to the one from the launch pass.
​
*** Another small issue is that in the launch tests so far, the ring is only launched 2-3 inches away from the cloth. Depending on the cameras and the editing of our shots, which haven't been finalized yet, this may be okay, but if not, I may have to animate or simulate the ring an additional time, to increase amount of time it spends moving upwards. ***
Previous FX test of the ring in moving cloth
Photos of FX tests for cloth and ring using different settings