SCAD x Harbor





Home/ Blog / SCAD x Harbor / week 7

    ② Week 7


            Bulgari Serpenti Ring      


    Feedback from Mentors (Class 13 )

    1. Position & Composition
    Shot 2-2: Adjust log/ring position to avoid leaning too far left.
    Shot 2-1: Enhance initial camera movement and increase macro feel (DOF).

    2. Color & Lighting
    Shot 1-2: Dragon’s black levels feel off (C6), try rotating the HDRI.
    Shot 3-1: Sky looks overexposed, increase bokeh, and remove ease-in on Bezier.

    3. Particles & Animation
    Shot 2-1: Improve particle pullout at 12'.
    Shot 1-2 & Pop shot 1-2: Fine-tune animations.

    4. Overall Suggestions
    Consider a camera push-in at the start for more impact.
    Great progress so far—keep refining!



    Week 7 Task
    1. Fixing shot1 and shot2 jumping issue
    2. Adjusting the shot2 particle based on new camera
    3. More particle for shot 3 fx

    1. Mesh jumping issue (trouble shooting) - shot1 & 2

    Week6 jumping issue at 01:00:03 ( scale transformation fx by Yilin Zhao; shader&camera by Mia Esparragoza; lighting by Zane Kidwell, particle fx by Soi Park, Compositing by Veronica Ramirez)
    Week6 jumping issue at 01:00:10 ( ring morphing fx by Soi Park; shader&camera by Mia Esparragoza; lighting by Zane Kidwell, ring anim& snake shape particle trail by Yilin Zhao, Compositing by Veronica Ramirez)


    From week7, the dragon model in shot1 and 2, as well as the ring, have a very noticeable model mesh change problem every time the effects are exchanged. 
    Firstly shot 1, I checked all the relevant nodes and found that the main problem is with the attribute delete in the fx of the Soi. In her nodes, particle and mesh are merge together and added an attribute delete node after that, and deleted some attributes (including vextex uv attribute of dragon model). sometime when merge occurs, Houdini tries to align the properties of different objects. for example, If the model has P, N, and Cd, and the particles only have P and v, the merge may result in some attribute values becoming 0 or undefined.

    The same problem was solved for shot2, with the slight difference that in shot2, the morphing fx (by Soi Park) also used remesh for the ring model, but by turning off remesh and modifying merge and attribute delete, the jumping issue was solved.

    trouble shooting (example of shot1)
    Original nodes(by Soi Park): everything merge together then adding a attribute delete node
    Fixed nodes (for particle): Separate particle and mesh, add attribute delete nodes  individually  
    Fixed nodes (for mesh): Separate particle and mesh, add attribute delete nodes individually (keeping the uv attribute of model)


    Week7 jumping issue fixed (scale transformation fx by Yilin Zhao; shader&camera by Mia Esparragoza; lighting by Zane Kidwell, particle fx by Soi Park,  Compositing by Veronica Ramirez)
    Week7 jumping issue fixed (ring morphing fx by Soi Park; shader&camera by Mia Esparragoza; lighting by Zane Kidwell, ring anim& snake shape particle trail by Yilin Zhao)


    2. Adjusting shot2 snake particle trail based on new camera


    shot2 This week's particel trail fx is mostly just a few minor tweaks based on the new camera moverment and timing

    Week7 shot2 snale particle trail (shader&camera by Mia Esparragoza; lighting by Zane Kidwell, ring anim& snake shape particle trail by Yilin Zhao,Compositing by Veronica Ramirez)



    3. More Particle for shot3

    Regarding this week's shot3, I mainly readjusted the glowing particles section and lowered the particle seperation to make the particles denser, although the cache time got longer (4 hours, despite using attribute delete and camera culling)

    week 6 shot3 render result ( particle seperation 0.03, and render in redshift)
    week 7 shot3 render result ( particle seperation 0.02, and render in redshift)




    WIP of shot 3 for shinning part (more detaile about this absrtact partilce background please see week5 blog)
    1. base part of abstract particle (overall nodes, more detail please check week5 blog)
    2. adding attirbute vop （for extract some of the color form the base)
    3. inside vop: Noise-based generation of colors (Cd), controlled by Ramp
    4. delete the black part of particle, then using attribute transfer (transfer color from base abstract particle to the extracted particle)
    5. Control the transparency of the particles through ramp, so that the particles have a smooth fade-in/fade-out effect during the life cycl
    6. color correction for shinning particle part




    Yilin Zhao / SANM560 / SCAD x Harbor