SCAD X Harbor Picture week 5

SCAD X HARBOR PICTURE

2/4/20255 min read

January 28, 2025

Feedback / Notes

Mentor Vi:

  • Consider adding movement to the flower in Shot 1.

  • In Shot 3, focus more on the metal drip on the flower.

  • The framing in Shot 5 is improved.

  • Shot 4 could be shorter.

Mentor Hailey:

  • Fix the dark grey center of the flower in Shot 1.

  • Punch in on Shot 3.

  • Adjust the yellow and greenish tones on the bottle in Shot 4.

  • The rays coming in Shot 5 are a good idea.

Mentor Joleen:

  • Should the ground in Shot 1 be water?

  • In Shot 3, have the camera follow the drop.

  • Maintain the drop's shape more in Shot 4.

  • Add more variety to the flowers in Shot 5.

Mentor Kyle:

  • The depth of field and textures in Shot 1 need improvement.

  • Shot 1 should have a more graphic approach; it feels too busy right now.

  • The story and layout are good, but the shaders and lighting need work—refer to references.

  • Remove the empty space in Shot 3 and bring the camera closer to the petals.

  • In Shot 4, slightly rotate the bottle, add more detail to the water droplet (it looks too even right now), and consider adding extra shots.

  • In Shot 5, fill in more flowers and brighten the sky.

  • For Shot 5, ensure the logo transition does not overlap and move the bottle up slightly.

Mentor Billy:

  • Add caustics in Shot 5.

  • For Shot 1, consider a top-down artistic approach, direct the space, and adjust the camera placement accordingly.

  • In Shot 4, refine the water roll effect and adjust the IOR of the glass.

Adjustments

Shot 1:

  • Add subtle movement to the flower.

  • Fix the dark grey center of the flower.

  • Improve depth of field (DOF) and texture quality.

  • Simplify composition to make it less busy.

  • Consider a top-down camera approach for better composition.

Shot 2:

  • flower textures in the center.

Shot 3:

  • Emphasize the metal drip effect on the flower.

  • Remove empty space and bring the camera closer to the petals.

  • Adjust framing to better follow the drop’s movement.

Shot 4:

  • Shorten shot duration.

  • Adjust the yellow and greenish tones on the bottle.

  • Ensure the drop maintains its shape throughout.

  • Slightly rotate the bottle.

  • Add more detail to the water droplet (currently too even).

  • Refine water roll effect

  • adjust IOR of the glass.

Shot 4.5:

  • new shot?

Shot 5:

  • Improve framing.

  • Increase variety in flowers and add more to the scene.

  • Brighten the sky.

  • Improve transition of the logo to prevent overlapping.

  • Move the bottle slightly upward for better placement.

  • Add caustics.

  • Ensure light rays interact well with the scene.

  • longer.

FX Tasks:
  • Refine water droplet shape in shot 4.

  • Adjust ripple thickness and increase ring spacing in shot 5.

  • Slow down fluid simulation and resolve mesh issues in shot 5.

  • Add bubbles to fluid simulations in shots 4 and 4.5.

February 07 , 2025

Shot 4 - Water Droplets

Vellum Fluid -

  • Viscosity

  • friction

Animation

February 06 , 2025

Shot 4.5 - Fluid + bubbles

So this week, I started working on bubbles for the water. Since Shot 4.5 is a new close-up shot, we thought it would be interesting to add bubbles. However, based on the real perfume, the bubbles should stay on the water's surface and move with it. I was thinking of using a DOP network to control the bubbles since I want them to disappear when the water starts moving, but I’m also looking for other methods that might reduce simulation time.

But I am not satisfied with the result this week. However, I ran out of time, so I will try another approach. I want the fluid to return when the bubble breaks, but since I am animating it, it is really hard to control right now.

get the point on surface

randomize pscale in a rang by fit01

delete points by using threshold

The way I make the bubbles follow the bottle animation is by using Point Deform. After that, I use Point Wrangle to make the bubbles disappear frame by frame.

bubble reference took by me

February 06 , 2025

R&D for Shot 5 – Fluid

Animated Collision

After getting the collision movement to work for the particles, I encountered an issue—loss of volume, which prevents the bottle from filling properly. This happens when particles escape or disappear during the simulation, leading to an unrealistic result.

loss of volume

I found some solutions to address this problem:

  • Ensure the collision object is thick enough to avoid leaking. If the collision geometry is too thin, particles may pass through, causing volume loss. Increasing the thickness helps contain the fluid properly.

  • Increase substeps. More substeps improve the accuracy of the simulation, reducing errors that lead to particle loss.

  • Apply and decrease particle separation. Lower particle separation increases resolution, making the fluid more stable and reducing gaps between particles.

  • Turn off reseeding. Reseeding can sometimes remove particles unexpectedly, leading to volume loss. Disabling it prevents unwanted changes in particle count.

  • Decrease the particle radius scale. A smaller particle radius scale helps maintain proper volume distribution and prevents excessive shrinking of the fluid.

I found this page to have some useful information on FLIP simulations.

https://www.cgchannel.com/2020/03/10-expert-tips-for-better-houdini-flip-fluid-simulations/

I was trying to adjust the particle separation and scale to achieve the best simulation.

The way I set up the FLIP simulation: since my collision will be animated, this is how I approached it. I cleaned up the geometry and used a Time Shift node to make it the emission source for the DOP network. I also used a Collision Source node to generate both geometry and VDB for the FLIP solver. Additionally, I ensured the collision is thick enough to prevent volume and particle loss.

After that, I used Extract Transform to get the translation and rotation from the bottle and used a VOP to apply the force. After that, I used DOP Import and applied a matrix in an Attribute Wrangle to make the FLIP simulation move with the collision.

Optimizing FLIP Simulations for Small-Scale Collisions

Since we're working at real-world scale, the bottle is quite small, making the particle values very low. To optimize the simulation, I used Fluid Compress to delete unnecessary attributes and set it to 16-bit to reduce sim time. After Fluid Compress, I used Particle Fluid Surface to transfer values back to the particles.

Refining Surface Meshes with Boolean Operations for Better Results

For a better surface after the simulation, I used a Boolean operation with the bottle to clean up some meshes.

Reference for Shot 4 - Water Droplets
Reference for Shot 5 - Fluid in the bottle

Reference pictures and videos were created by me. Special thanks to Professor Gaynor for providing the Chanel bottle.

full bottle fluid reference

Full Bottle Fluid Moving Following the Motion in Previs Reference

Reference Shooting by Myself
Reference Shooting by Myself

slow motion

slow motion