Articles

Houdini Fluid Advertising Workflow: From Setup to Final Composite

ARTILABZ™

ARTILABZ™ gives you unlimited access to all Houdini courses, 3D assets, simulation files, textures and tools. updated every month.

Everything You Need to master Houdini.

01

Premium Houdini Tutorials

Full access to every course — fluid simulation, procedural FX, brand visuals and more.

02

Monthly New Content

Fresh tutorials and assets added every month — your library grows with you.

03

Instant Access to Everything

The moment you join, the full library is yours — no drip-feed, no waiting.

04

Project Files Included

Every tutorial comes with the full Houdini scene file — open every node, learn every detail.

FROM 14.99€/MONTH

Includes one exclusive complete course

The exclusive course — a full production tutorial you won't find anywhere else, never sold alone.

Best Seller
Most Loved
Tutorial Camera Rig

ADVANCED CUSTOM CAMERA RIG

ANIMATION · CONSTRAINTS · CUSTOM UI

BUILD A FULLY CUSTOM CONSTRAINT-BASED CAMERA RIG IN HOUDINI WITH A CUSTOM UI PANEL. DESIGN FLEXIBLE SYSTEMS FOR PRECISE, CINEMATIC CAMERA ANIMATION ON ANY PROJECT.

€29.99

Freebies
Free Studio HDRI Pack box by Artivoxa showing 60 studio lighting setups with softboxes wrapped around the packaging

Studio HDRI Collection

ASSETS · EXR & HDR · 60 HDRIS

DOWNLOAD 60 STUDIO HDRIS CAPTURED IN A REAL PHOTO STUDIO. LIGHT YOUR PRODUCT AND BEAUTY RENDERS LIKE A PHOTOGRAPHER — SOFTBOX, LANTERN, STRIP AND GRID SETUPS, READY FOR ANY RENDERER.

FREE

Houdini Fluid Advertising Workflow: From Setup to Final Composite

A production-ready Houdini fluid advertising workflow begins with the camera and product message, not the solver. Plan the fluid’s role, establish real-world scale, build reliable sources and colliders, and test the primary motion at low resolution before committing to final caches. Once the action is approved, treat meshing, look development, rendering, and compositing as distinct stages with their own versioned outputs.

This guide explains how to structure that process from setup to final composite. It covers FLIP simulation, art direction, collision troubleshooting, liquid surfacing, shading, render passes, color management, and production decisions that keep revisions manageable. The goal is not merely to create realistic liquid, but to produce a controllable shot in which the fluid supports the product, survives close inspection, and integrates convincingly with the final image.

Plan the shot before opening Houdini

A product advertisement is judged by composition, brand visibility, timing, and surface quality as much as by physical realism. A technically convincing simulation can still fail if it hides the label, misses the hero frame, or produces detail that cannot be controlled downstream.

Define the visual target and delivery constraints

Break the brief or reference into separate production requirements:

  • the fluid’s primary behavior and material character;
  • the product’s position, movement, and hero feature;
  • camera framing, lens choice, and movement;
  • lighting direction and background design;
  • the timing of the impact, reveal, splash, or settling motion;
  • secondary effects such as droplets, foam, mist, or residue;
  • delivery resolution, frame rate, aspect ratios, and frame range.

Describe the fluid action as a readable gesture. It might wrap around a bottle, form a crown behind a container, create a directional pour, or hold a suspended shape for a hero frame. Each action implies different source geometry, collision requirements, solver settings, and resolution priorities.

Camera decisions must happen early. Focal length, camera distance, angle, movement, and depth of field alter the apparent scale and speed of the liquid. A simulation that feels balanced through one lens may appear too slow, thin, or visually crowded after a camera change. If the final camera cannot be locked immediately, identify which parameters remain provisional and avoid refining details that depend on them.

Plan timing as both an editorial and simulation problem. Mark when emission begins, when the fluid reaches the product, when the main gesture becomes readable, and when it settles or exits. Significant retiming after simulation can create inconsistent velocity, motion blur, or secondary-particle behavior. If the action must be substantially faster or slower, changing emission timing and resimulating is often cleaner than forcing the final cache into a new duration.

Establish scale and a camera-ready blockout

Use consistent real-world scale for the product, fluid source, colliders, camera, and lights. Scale affects the perceived response to gravity, viscosity, surface tension, collision thickness, absorption, and motion blur. It also determines whether small features are physically meaningful or below the practical resolution of the simulation.

Build a proxy product and simplified collision geometry before refining the asset. The blockout should include the camera, rough lighting direction, background framing, source position, and intended contact areas. This inexpensive setup can reveal whether the fluid obscures a logo, intersects an opening incorrectly, or leaves the frame before simulation and rendering become expensive.

Decide what belongs in simulation, rendering, and compositing

Simulate features that require coherent volume, physical contact, changing occlusion, shadows, reflection, or refraction. A droplet that collides with the product or refracts the environment normally belongs in 3D. A subtle glint, restrained bloom, or atmospheric accent may be more controllable in compositing if it does not need to affect geometry or lighting.

The primary solve does not need to generate every visible detail. Secondary particles can extend a splash, instanced droplets can add designed accents, and shader-level breakup can provide fine variation without increasing the density of the entire simulation. Separating these layers also allows the main silhouette to be revised without rebuilding every finishing effect.

This boundary should be agreed on before production advances. It is a central part of a broader advertising VFX workflow, because it determines which revisions require a new simulation, a new render, or only a composite update.

Build a reliable Houdini scene

A dependable scene separates source preparation, collision geometry, simulation, meshing, look development, rendering, and compositing. That structure makes it easier to determine whether a problem is creative, physical, or technical instead of allowing corrections to accumulate in one network.

Prepare sources and collision geometry

Check transforms, units, normals, topology, and thickness before bringing the product into the fluid network. Thin, open, or self-intersecting surfaces can create ambiguous collision volumes. Non-uniform scale and inconsistent transforms can also make collision tolerances difficult to interpret. For animated geometry, inspect the full frame range rather than validating only the rest frame.

The solver rarely needs the highest-resolution product mesh. A simplified collider is usually faster and easier to diagnose, provided it preserves the silhouette, contact areas, sharp transitions, and gaps that affect the fluid. Small cavities and thin shells need enough usable thickness to survive conversion into a collision representation. Always visualize that representation before launching a full simulation; the rendered product mesh does not prove that the solver sees the same shape.

Create the source around the intended action. Geometry can define a solid emission region, curves can guide a stream or nozzle path, and volumes can control a broader incoming mass. The source should carry intentional velocity and timing rather than relying on the solver to turn a static emission into the desired performance. Inspect it over time for sudden volume changes, discontinuous velocity, or particles born inside the product.

Choose the fluid approach based on the shot

A FLIP-based simulation is a practical starting point for pours, splashes, free surfaces, volume-preserving motion, and interaction with product geometry. It combines particle-based motion with a reconstructed surface for rendering. It is not automatically the best answer for every liquid-like element: a tightly controlled stream, stylized sheet, or nearly static hero shape may benefit from procedural geometry, guided particles, volumes, or a hybrid setup.

For FLIP, set particle separation according to camera-visible detail rather than choosing the smallest value available. Lower particle separation can resolve finer sheets and droplets, but it increases simulation time, memory use, cache size, and meshing cost. A background stream does not necessarily need the same density as a close product-contact splash.

Set substeps according to the fastest relative motion and the narrowest important collision feature. More substeps can improve temporal sampling, but they do not repair invalid geometry, inconsistent scale, or a poorly designed source. Similarly, aggressive reseeding may even out particle coverage while introducing unstable density changes in thin regions. Evaluate these controls through the final surface, not just the particle display.

Material properties must be assessed together:

  • Viscosity affects resistance to flow, stretching, and settling.
  • Surface tension influences cohesion, breakup, and bead formation, especially in smaller features.
  • Gravity establishes the dominant acceleration and strongly affects shot timing.
  • Drag and custom forces can suppress or redirect motion but may look artificial when they overpower the main flow.

There is no universal liquid preset. The correct values depend on scene scale, source velocity, material identity, camera distance, and the designed action. A practical example of these trade-offs can be seen in an olive oil pour workflow, where viscosity, glass interaction, and lighting must be considered together.

Art-direct the motion without losing plausibility

Advertising fluid is designed motion. Physical plausibility matters, but the liquid must also frame the product, reveal the intended feature, and form a readable silhouette at the delivery frame rate.

Control the primary gesture first

Judge the main action without droplets, mist, foam, or procedural surface noise. The stream, wave, wrap, or impact should already communicate a clear direction and relationship to the product. If it does not, adding resolution usually increases cost without correcting the composition.

The source is the first art-direction control. Its size, shape, orientation, emission window, and initial velocity determine whether the liquid begins as a focused stream, broad sheet, or swelling mass. Animate these properties smoothly enough to preserve continuity in velocity and volume. Abrupt source changes often read as an edit inside the simulation.

Colliders can also shape the performance. Product geometry, a pedestal, or an invisible guide surface may redirect fluid away from a label or toward a controlled silhouette. Guide geometry should remain simple, watertight, and physically defensible from the camera. If it produces abrupt changes in direction or unexplained acceleration, it will expose the art direction rather than support it.

Velocity fields, attractors, drag, and turbulence can refine the broad flow. Use them locally and evaluate acceleration as well as position. A field may place the liquid correctly on one frame while producing implausible motion on the frames around it. Large-scale direction should remain coherent before small-scale turbulence is introduced.

Protect product visibility

The fluid should support the label, cap, opening, silhouette, or material feature being advertised. Maintain deliberate negative space around those areas, even when a less controlled simulation would naturally cover them. Possible solutions include reducing fluid volume, changing the source angle, introducing a restrained guide, or adjusting the camera. The best intervention is usually the one that creates the desired composition with the fewest downstream exceptions.

Review at the final output size. Separate strands visible in the viewport may merge into one opaque shape after shading, motion blur, bloom, and downsampling. Conversely, small droplets that consume substantial simulation and render time may disappear entirely in delivery.

Add secondary detail as controllable layers

Once the main action is approved, add details according to their visual purpose:

  • Splashes and sheets should clarify an impact or change of direction while remaining connected to the primary motion.
  • Tendrils can frame the product or form leading lines, but weak strands often create clutter and unstable meshing.
  • Droplets work best as selected accents rather than a uniform cloud of noise.
  • Foam and bubbles should correspond to the material and contact conditions instead of hiding an unresolved base surface.
  • Mist and fine spray require coordinated simulation, lighting, sampling, and compositing because low-contrast detail can disappear or become noisy.

Secondary elements must share the velocity and acceleration of the main liquid. A droplet layer that appears for one frame, changes direction without a force, or carries incompatible motion blur can make an otherwise convincing shot feel synthetic.

Cache for controlled iteration

Do not begin with a final-resolution, full-duration simulation. Use low-resolution or reduced-range caches to validate emission timing, collision behavior, fluid direction, contact points, and the broad silhouette. Visualize particles, velocity, collision fields, and active simulation bounds when diagnosing the result; a beauty render can hide whether an artifact began in sourcing, solving, or meshing.

Once the motion is approved, cache the particle state before expensive surfacing and rendering. Preserve the attributes needed downstream, which may include velocity, age, id, viscosity, or material classifications. The exact set depends on the meshing, secondary-effects, shading, and motion-blur requirements.

Keep simulation and mesh caches separate. This allows the surface to be rebuilt with different filtering or polygon settings without rerunning the solver. Version the source, collider, camera, scene file, solver settings, frame range, and outputs together so the approved state can be reproduced.

A useful production distinction is:

  • Creative change: alters source shape, timing, volume, path, or fluid behavior to redesign the action.
  • Technical fix: corrects scale, transforms, collision thickness, unstable time stepping, cache dependencies, or missing attributes without changing the approved performance.
  • Detail decision: increases resolution or adds a secondary layer only where the camera and final image require it.

This distinction prevents an expensive technical recalculation from being mistaken for a creative improvement. It also helps production decide which downstream caches and renders have been invalidated.

Mesh the fluid for stable rendering

An approved simulation is not yet a finished render asset. Surface reconstruction can introduce holes, disconnected islands, inflated joins, temporal flicker, or thin regions that refract unpredictably.

Reconstruct and inspect the surface

For FLIP particles, Houdini’s Particle Fluid Surface is a common starting point. It reconstructs a continuous polygon surface that can reflect and refract the environment, receive shadows, and interact with the product in the renderer. If the fluid is already represented as a volume, a VDB-to-polygon workflow may be more suitable.

Inspect the generated mesh at the intended camera distance and across adjacent frames. Look for:

  • holes or missing coverage;
  • unwanted isolated components;
  • blobby connections between separate droplets;
  • self-intersections and unstable normals;
  • thin sheets that collapse during reconstruction;
  • topology changes that create flickering highlights or refraction.

Smoothing and filtering can reduce particle-scale noise, while remeshing can improve polygon distribution. Too much smoothing, however, rounds sharp splashes, closes meaningful gaps, and erases directional detail. Refine hero streams, impact areas, and the camera-facing silhouette instead of increasing quality uniformly across the cache.

If a feature is smaller than the particle representation can support, meshing controls cannot reliably recover it. Either increase simulation detail where justified or build the accent as a separate layer. Manual patches may work for a still image, but they are difficult to keep stable through animation.

Develop the liquid look and lighting

Start with the intended material behavior rather than a generic liquid preset. A water-like material may rely on transmission, reflection, refraction, absorption, and an appropriate index of refraction. A viscous, opaque, or cosmetic liquid may depend more heavily on diffuse response, roughness, subsurface behavior, suspended material, or volume attenuation.

Verify scale and thickness before tuning the shader. Incorrect geometry cannot be made physically convincing by changing absorption or transmission alone. Judge attenuation through representative thin and thick regions, and confirm that the liquid maintains its intended color against both bright and dark backgrounds.

Use physically plausible shading as a stable base, then art-direct controls that remain consistent across the sequence. Roughness can soften distracting reflections, and absorption can give clear liquid visual body. Avoid using shader adjustments to conceal holes, unstable topology, or incorrect thickness; those defects usually reappear under a different light or camera angle.

Light the fluid and product as one setup

Large soft sources can establish readable gradients across glossy liquid, while narrow or shaped sources can create controlled highlight lines along a stream or splash. The product should receive compatible reflections, shadows, and color influence. If the fluid and product respond to different lighting logic, the result may look composited even when both were rendered in the same scene.

Contact areas deserve particular attention. Where liquid touches a bottle, label, surface, or cap, inspect shadows, reflected highlights, refraction, and local darkening. These cues must remain consistent with the material: an opaque fluid, transparent liquid, and suspended mist will not anchor to a surface in the same way.

For material-specific product work, the interaction between glass, liquid, and controlled beauty lighting is explored further in this skincare serum CGI breakdown.

Render useful passes without multiplying cost

Render passes should answer specific compositing needs. More outputs do not automatically create more control; they can also increase storage, render management, and the risk of rebuilding a beauty pass incorrectly.

Depending on the renderer and pipeline, useful outputs may include:

  • the beauty render;
  • fluid-only and product-only layers;
  • object, material, or Cryptomatte-style identification masks;
  • depth and motion-vector data;
  • shadow and emission contributions;
  • reflection, refraction, or transmission components when they can be recombined reliably;
  • separate mist, foam, spray, or volume layers.

Use the simulation’s velocity data for true rendered motion blur when fluid moves quickly. Post-process blur may be useful for limited adjustments, but it cannot always reproduce changes to silhouette, transparency, refraction, or interaction with the product.

Test both hero frames and fast-motion frames before rendering the sequence. Hero frames expose surface topology, thickness, shading, and refraction. Fast frames reveal motion-blur problems, temporal flicker, unstable highlights, fireflies, and noisy glossy or transmission paths.

When a render is noisy, isolate the cause before raising global samples. Test simplified materials and lighting, then evaluate reflection, refraction, volumes, motion blur, and geometry separately. Tiny disconnected mesh components, self-intersections, or unstable normals can create expensive light paths that additional sampling will not solve efficiently.

Integrate the fluid in the final composite

Compositing is where the fluid, product, environment, and photographic treatment are made to occupy the same shot. It should provide controlled finishing adjustments without becoming a place to hide incorrect scale, collision behavior, or motion.

Match camera, depth, and contact

Before grading, confirm that camera projection, focal length, sensor assumptions, framing, and object scale match across the fluid, product, and plate. Then inspect depth order and holdouts. The liquid should pass in front of or behind the product only where the geometry requires it.

Contact points are strong integration cues. A splash that intersects the product without corresponding shadows, reflections, refraction, or local occlusion will appear detached. Depth data can assist with atmosphere and selective defocus, but it should not replace correct geometry and occlusion.

Motion blur must also agree across elements. Match shutter behavior, blur direction, and timing between the fluid, product, and any live-action plate. If these differ substantially, the elements appear to exist at different moments even when their positions align.

Maintain a consistent color pipeline

Place the fluid and product under the same color-management transform before applying a shared look. Compare exposure, white balance, contrast, saturation, and black levels in the viewing space while retaining scene-linear data for operations that depend on physical light relationships.

Transparent liquid needs inspection beyond its silhouette. Refraction should distort the environment according to the fluid’s thickness and scale, and attenuation should vary through the volume. Check the result over bright and dark backgrounds; uniformly bright edges, missing internal color, or unnaturally clean contours often reveal a render-composite mismatch.

Apply lens treatment after the integration is stable. Bloom, glare, grain, distortion, chromatic treatment, and depth of field can unify CGI with a plate, but they cannot correct a wrong camera or missing contact interaction. Distortion should affect all relevant elements consistently, and grain should be introduced late enough for the product, fluid, and background to share the same image texture. For a broader treatment of these finishing stages, see the guide to CGI post-processing and compositing.

Review the composite at full resolution and at delivery size. Full-resolution checks expose unstable edges, refracted detail, mesh artifacts, and render noise. Delivery-size checks reveal whether droplets disappear, bloom overwhelms the product, or separate liquid shapes merge into an unreadable mass.

Diagnose common workflow failures

Fluid leaks through the product

Leaking usually begins with the collision setup rather than the render mesh. Check whether the collider is closed, transformed correctly, thick enough, and represented at sufficient collision resolution. Inspect animated velocity and determine whether the fluid or collider travels too far between time steps. If the collision field is accurate but fast motion still tunnels through it, adjust temporal sampling or substeps before changing the meshing stage.

The motion jitters or explodes

Review source animation, incoming velocity, scene scale, external forces, solver time scale, and substeps. Abrupt velocity changes can create unintended impulses, while fields that are too strong for the shot scale can produce chaotic acceleration. Use a diagnostic cache to remove optional forces and secondary effects, then reintroduce them one at a time.

The liquid looks blobby or melted

Compare the particle distribution with the reconstructed surface. Low particle density, an overly broad influence radius, excessive smoothing, or a coarse mesh can merge separate forms and erase thin features. Increase detail only when the final camera can resolve it; otherwise, redesign the silhouette or add selected secondary geometry.

Thin sheets and droplets flicker

These details may be near or below the reliable resolution of the simulation and meshing process. Inspect adjacent frames for changing topology, isolated components, and unstable normals. Remove detail that does not survive delivery, or generate important accents as separate particles or geometry with more predictable temporal behavior.

Refraction produces fireflies or excessive render times

Test the surface with simplified lighting and materials. Check normals, thickness, self-intersections, tiny components, bright environment sources, and the number of reflection or transmission paths. Target the responsible component rather than increasing every render setting or denoising the entire image indiscriminately.

The render is clean but does not integrate

Compare scale, camera, depth, lighting direction, contact shadows, reflections, refraction, black levels, depth of field, and motion blur. A physically credible simulation can still look detached if it does not share the photographic logic of the product and environment. Lens effects should be the last refinement, not the first repair.

Use stage approvals to keep production controllable

A robust workflow limits expensive recalculation by approving the shot in stages:

  1. Blockout approval: confirm the camera, product framing, scale, source placement, broad timing, and required negative space.
  2. Preview simulation: validate emission, collision behavior, main silhouette, and contact points with inexpensive caches.
  3. Simulation approval: version the chosen particle cache and record the source, collider, solver settings, dependencies, and frame range.
  4. Surface approval: check topology, thin features, temporal stability, and the final camera silhouette before look development.
  5. Look-development approval: test material, lighting, motion blur, and render quality on representative hero and fast-motion frames.
  6. Composite approval: verify camera agreement, depth, contact, color management, lens treatment, and readability at delivery size.

Version simulation caches, mesh caches, renders, and composites independently. A new grade should not require another FLIP solve, while a collision correction should not erase approved compositing decisions. When each stage has a clear responsibility and recoverable output, creative revisions remain flexible and technical fixes can be traced to their source—the defining advantage of a production-ready Houdini fluid advertising workflow.