WISH

Wish invites audiences to enter the magical kingdom of Rosas, ruled by the sorcerer King Magnifico (Chris Pine), who has the power to grant the wishes of his loyal subjects. When young idealist Asha (Ariana DeBose) makes her own heartfelt wish, her plea is answered by a dazzling ball of cosmic energy called Star. Uniting with Star, Asha sets out to save Rosas from the tyranny of King Magnifico and help its population to learn the true value of following their dreams.

“We were able to see up-close the original effects animation artwork from films like Cinderella, Peter Pan and Sleeping Beauty,” said Dale Mayeda, head of effects animation. “We appreciated the craftsmanship and warmth of imperfection you get in hand-drawn effects drawings. We wanted to incorporate what inspired us by adding variation and imperfection into our effects, as well as graphic iconography, focusing animation on arcs, and caricatured timing.”

The artistic approach demanded control over the design of every frame of animation. As a result, the Wish effects team relied less on simulation and more on procedural animation. Artists combined 2D effects animation techniques with modern 3D procedures to generate a look that was simultaneously fresh and familiar.

“We utilized the twelve principles of animation introduced by Disney animators Ollie Johnston and Frank Thomas,” said Erin Ramos, head of effects animation. “Applying these principles to our effects work allowed our team to stretch their artistic abilities.”

Star Dust Magic

Critical to the film’s success was the appearance and animated performance of the childlike and highly energetic Star, in which the effects department played a vital role. Work on the effects began well before the character made it into the hands of the effects department.

“We worked in collaboration with the animation and rigging departments,” said effects supervisor Jake Rice, “developing a Houdini Engine rig to create controllable and easily animatable silhouette outlines for the character. Since Star is non-verbal, it became increasingly important to ensure that the character’s form was easily readable against any background, and to make sure that poses were readable against the character’s own body. The ease of use of Houdini Engine made the whole process extremely extensible and robust, leaving the animators to work their magic with very few hiccups.”

In basic form, the character of Star appears as a solid color against the background environment, enhanced by a twinkling dust trail. However, the simplicity of the underlying form made it difficult to read the form of the body in more extreme poses. “We added in controllable silhouette lines,” commented Rice. “These helped the audience understand the depth and form of the character without adding in extra shading.”

The silhouette lines were generated in Houdini. “From Maya, we passed the Star geometry as an input into Houdini Engine,” Rice explained, “together with a series of Null locators denoting position. We passed the shot camera information to the Houdini Engine rig through parameters. These were hooked up to a camera within the Star outlines OTL.” Each element of Star’s body required its own set of silhouette lines. “We generated a level set on the mesh by dotting the camera’s eye vector with the surface normal. From here, we used a VEX solution to extract the silhouette curves using a marching squares-type algorithm.”

Further refinements made it a straightforward business to clip the silhouette lines and center them with respect to the body part’s locator in screenspace. This gave animators the high degree of control they needed to create the required stylized look. “Changing which side of the body part a silhouette line was centered was as easy as moving a locator in camera space,” noted Rice. Animators also had the ability to carve out the main line around Star’s body with respect to the centers of Star’s limbs. “This allowed for animators to control and fake the amount of perceived overlap a given limb had with respect to the body.”

Once the outlines of Star were baked in, the effects work on Star’s signature dust trail could begin. “Star’s dust was a semi-automated Houdini rig which took in the character and outline,” Rice revealed. The Houdini rig generated tiny particles of dust emitting out from the body, based off areas of screenspace where the outline did not appear. “The rig had lots of flexibility, giving effects artists a solid starting point for their work while leaving lots of room for fine tweaking and adding custom flourishes. One key component was the rig’s ability to incorporate hand-drawn effects and layer them into 3D. We took great advantage of the trace SOP along with all of the NDC VEX functions to help place our hand-drawn artwork into 3D.”

The final step was to add a light amount of dusting to Star’s silhouette line. To achieve this, the effects team took advantage of the fact that the silhouette line was generated within Houdini. “We used the same rig with parameters supplied from animation,” said Rice, “to create PolyWire geometry onto which particulates were stuck. Because the amount of outlines could change from frame to frame, we came up with a solution that allowed us to generate a constant number of unclipped outlines. This was then scaled to match the clipped geometry, which meant that the dust would track with the outlines without flickering or requiring a simulated solution.”

Star Gets the Zoomies

During moments of excitement, Star takes flight with sudden bursts of hyperactive energy, like an overstimulated puppy getting the ‘Zoomies.’ When faced with such frenetic movement, artists might typically use motion blur to smear the effects based on velocity. However, inspired by storybooks from the early 1900s, the production design directive for Wish took motion blur well and truly off the table. “Not using motion blur created interesting challenges for our rig,” remarked effects lead Ian Coony. “We had to simulate fast motion without strobing artifacts.”

The team researched options, ultimately drawing inspiration from classic hand-drawn effects animation, which simulates the feeling of fast-moving objects by squashing and stretching character shapes, and by adding ‘dry-brushed blur.’ The end result was a dedicated Houdini rig for Zoomie scenes.

“We based our procedural Houdini network on the wind rig for the Gale element from Frozen 2,” said Coony. ‘The rig allowed animation and layout to see the amount of streak or blur using elongated teardrop-shaped geometry to represent speed and scale, using published input from the animation department.” The effects team resampled the upstream animation into a simulated motion path. “We had a Houdini-based network of controls for path-shaping, length, width, noise, texture, and magic dust emission. We added graphic silhouette lines based on camera angle using controls for Star’s overdrawn rough edges.”

The effects team created additional tools derived from Star’s Zoomie rig, introducing such variants as Enchanted Wish Magic, Butterfly Flourishes, and Star Energy Blasts. “We created these variants for shots that needed more traditional glints and pixie dust, with a touch of the Star streak built in,” commented Coony, “and also as an interesting hybrid of Star Dust Magic and the Zoomie rig. These rigs had inputs for 2D hand-drawn flourishes and were designed to work in stereo using Z-Depth camera projections.”

Wish Bubble Magic

The captured wishes of the citizens of Rosas exist within a large repository of floating bubbles, all contained within the tower of King Magnifico’s castle. As a nod to the 100-year anniversary of Walt Disney Animation Studios, the tower contains a grand total of 1,923 wishes.

The appearance of the bubbles depends on the type of shot and the particular story beat. In certain close-up shots, the intricate contents of an individual bubble are clearly visible as a miniature animated tableau. In wider views – or in shots where too much detail would distract from the performance of a nearby character – the contents of the bubbles are obscured.

The effects department worked primarily on shots containing multiple wish bubbles, varying their approach as the script demanded. “There are many different ways we see the wishes,” observed effects lead Joel Einhorn. “We might see a few spheres up close, or numerous wishes filling a big space. We also see the bubbles in different states – ‘clouded,’ where just the bubble is visible, or ‘revealed,’ where you see the contents of the wish. For the final scenes of the movie we needed a ‘dead’ bubble look, which was sort of a bubble-cage without a shell.”

To develop the look for the clouded bubbles, effects animator Bruce Wright, under the guidance of effects designer Dan Lund, imagined stars and nebulae refracted from an infinite distance through the facets of an invisible internal gemstone. “The bubble was created using a reflection map on invisible shards inside the sphere to light up a noisy volume element,” explained Einhorn. “This created an angle-dependent sapphire-like appearance, with the feel of a refracted stellar nebula.”

With large numbers of bubbles visible in any one shot, the main challenge for the effects team was achieving a consistent look that was easy for artists to work with. “We created nested copies using our in-house instancer, Aurora,” Einhorn said. “This way we could easily make changes to the source wish bubble, which were then automatically propagated over all instances.”

The overall layout of the bubbles was established by a point cloud. “Each point had a bubble, an interior and an exterior light instanced to it,” commented Einhorn. “The invisible shards that create the glow were instanced inside each bubble. When wishes needed to be revealed inside a bubble, we teamed up with the crowds team. They gave us a setup to generate random characters, which we used to create 300 unique wish scenes. We used these for the instanced bubbles that required visible wishes.”

The effects team drove the whole process with point attributes, giving artists the ability to control a wide range of parameters. “We could change the shell color, volume color, light element intensity, amount, scale and color of the shards, how fast they move and spin, the reveal time of the bubbles, and more,” said Einhorn. “This was all wrapped up in an HDA which was part of a rig. The rig would read the artists’ animation of the point cloud for the bubbles, which could then be adjusted to achieve the desired look and feel. This approach enabled artists to focus almost solely on the look and performance of the wish bubbles, with all the technicalities neatly hidden inside the inner workings of the rig and HDA.”

In order to create a natural arrangement, it was important to vary the spacing between the bubbles floating in the tower. “We used VDB to Spheres on a noised-up volume,” noted Einhorn. “This created random-sized spheres that were all larger than a wish bubble. We used these larger spheres for our simulation in order to prevent interpenetration of the wish bubbles.” The team also made use of pre-made groups of wishes, prepared on a 1,000-frame animated cycle. After being placed by layout artists, these moved through the pipeline with little need for further adjustment.

A number of close-up shots reveal a character’s wishes as a hero animated scene nestled inside one of the bubbles. “We worked closely with the technical directors,” Einhorn remarked, “who created a way to fetch the hero wish scene and make sure it sat within its bubble seamlessly.” For such shots, the lighting team used the wish bubble with all settings exposed, giving artists ultimate control over all aspects of the look of the wishes, and therefore the ability to tailor it exactly to each individual shot. “The lighting artists could control the look of the hero scene, the bubble, and the amount by which the wish was revealed.”

Hero wish bubbles provided a basis for additional looks, used variously for certain key story moments. Effects animator Christopher Hendryx used hero bubbles without their shells to create the look of the wishes returning to their owners. Paul Carman employed them to create the cage used by King Magnifico to hold the bubbles captive, while Joel Einhorn used them to generate a specific look for the moment when wishes are destroyed.

Magnifico Magic

During the course of the film, King Magnifico puts on dazzling displays of his magical powers, referred to as Glass Magic, Flare Magic, Flow Magic, and Hands Magic. The effects animation team combined cutting-edge technology with Disney’s 100-year artistic legacy, collaborating with veteran animators whose credits included such iconic Disney films as The Little Mermaid, Beauty and the Beast and The Lion King.

One of the biggest challenges was devising ways to translate hand-drawn designs successfully into a 3D world. To achieve this, the effects team turned once again to Houdini. “The magic systems we developed were a seamless blend of procedural generation and hand-crafted keyframe animation,” said effects lead Daniel Clark. “Houdini’s modularity allowed us to build individual effect elements that we could kit-bash together into more complex sequences. For example, we designed several smaller, reusable components – like swirling particles, magical flares, and dynamic flows – that we could mix and match to create larger, more elaborate effects.”

This procedural approach delivered tremendous flexibility. It also sped up the iteration process when collaborating with the animation and lighting departments. “Houdini’s node-based architecture allowed us to rapidly prototype and test variations of the magic,” observed Clark, “especially when we needed to emphasize specific emotions or narrative beats.”

“The whole thing was truly a team effort,” Clark added. “One of the most exciting aspects of the project was collaborating with artists whose work extends over decades of filmmaking. Together, we were able to delve into the Disney archives for inspiration, pulling from the classic animation styles that shaped generations.”