Are you struggling to achieve a smooth, realistic pour of Liquid Foundation in Houdini without spending hours on trial and error?
Do constant adjustments to particle settings and viscosity leave you with stiff, lifeless fluid that looks nothing like the dewy finish you need for beauty CGI?
When every millisecond counts and budget constraints loom, wrestling with the FLIP solver can quickly become overwhelming.
In this guide, you’ll move past guesswork and gain a clear workflow to nail your foundation simulation—no more wasted renders or patchy results.
Get ready to master key Houdini techniques, optimize FLIP parameters, and deliver that flawless liquid effect in record time.
What prerequisites and reference assets do I need to simulate liquid foundation in Houdini?
Before diving into a FLIP simulation of liquid foundation, ensure your environment is set. Houdini 18.5 or later provides improved FLIP workflow and GPU acceleration. Install SideFX Labs tools for quick emitter setups and shelf assets. Confirm your machine meets CPU core counts or has an Nvidia card compatible with Houdini’s GPU solver.
- Houdini 18.5+ with FLIP Solver and Labs toolset
- High-core CPU or CUDA-enabled GPU
- Face geometry: clean topology, subdivided low-res mesh
- Photogrammetry scan or retopologized UV-mapped head model
- Rheology data: viscosity vs. shear-rate curves
- High-resolution reference photos of foundation swatches
Collecting accurate reference is crucial. A photogrammetry scan yields true face contours for collision. Pair this with measured rheology curves—viscosity and surface tension—so your solver parameters mirror real fluid behavior. Finally, capture color swatches under standardized lighting to build a PBR liquid foundation shader later in the pipeline. These assets form the foundation of a physically plausible, production-quality simulation.
How do I set up a FLIP tank tuned for foundation: particle size, viscosity, surface tension, and seeding?
Begin by creating a FLIP Object inside a DOP Network and enclosing it in a tank geometry. This container defines boundary collisions and fluid containment. Assign a static object for walls, enable “volume velocity” if you plan on downstream render motion blur or mesh generation.
Adjust particle size via the Particle Separation parameter on the FLIP Object. A separation of 0.002–0.004 works well for beauty shots, balancing fine capillary detail against simulation cost. Smaller values capture small ripples; larger values reduce memory usage.
Under the FLIP Solver’s Physics tab, set viscosity to match liquid foundation’s thick flow—typically 0.003–0.01 Pa·s. Enable “Enable Viscosity” and choose a smoothing kernel radius of 1.2x particle size. For surface tension, enter 0.5–1.0 N/m and activate curvature sampling to control droplet cohesion and film behavior.
Seeding ensures stable initial fill: inside the Source volume node, use a Density Volume or VDB to drive emission. Seed particles uniformly, with a jitter of 0.1, to prevent lattice patterns. Emission rate should match the tank’s volume divided by the time to fill (e.g., 1 m³ over 2 s = 0.5 m³/s).
- Particle Separation: 0.003
- Viscosity: 0.005 Pa·s
- Surface Tension: 0.8 N/m
- Emission Jitter: 0.1
With these settings, your FLIP tank will simulate the smooth, semi-viscous flow characteristic of liquid foundation. Fine-tune based on shot scale and render tests to achieve realistic spreading and droplet breakup.
How do I simulate realistic application behaviors — pouring, spreading, shearing, and skin contact?
Pouring: To achieve a natural pour, use a Source Volume node to emit FLIP particles with an initial velocity matching your foundation bottle’s tilt. Connect to a Flip Object and set the emission region to a low-resolution proxy. Ramp up particle count over the pour duration to avoid sudden density jumps, and add a Gas Force for subtle turbulence.
Spreading: For realistic spreading on a surface, leverage the FLIP Solver’s viscosity controls. Switch to a non-Newtonian model under Viscosity→Model→Carreau. Tweak the μ₀ and μ∞ values to mimic foundation thickness. Enable “Slip on Collision” in the Collision tab to prevent particle locking, allowing the fluid to expand naturally into a thin, even film.
Shearing: Foundation shear-thins under brush or sponge pressure. Capture this by using an Attribute Wrangle in DOPs to drive per-particle viscosity based on strain rate. Compute strain via wrangle code like f@viscosity = fit(pow(length(v@vel - v@vbd), p), lowVisc, highVisc); and feed this attribute into the solver’s viscosity field. The result is fluid that thins under motion and thickens at rest.
Skin Contact: Accurately simulating foundation adhesion demands high-fidelity collision. Use a Static Object with a skinned SDF collision volume. In the FLIP Solver Collision tab, increase the surface VDB resolution to capture micro-grooves. Activate “Stick to Surface” and adjust the adhesion coefficient to fine-tune fluid attraction. Optionally add a point-based magnet force near the skin to simulate capillary pull into fine pores.
How do I create accurate foundation shading and look development: base color, subsurface scattering, sheen, and micro-detail?
For realistic liquid foundation, you need a shader setup that replicates the way pigment, oil, and light interact. In Houdini’s Material network (or Solaris LOPs), the Principled Shader offers parameters for base color, transmission, subsurface scattering, coat, and micro-detail. By adjusting these settings you can mimic the semi-translucent, slightly oily appearance of real makeup.
- Base Color and Transmission: Feed a scanned or hand-painted 2K color texture into the Base Color. Use the Transmission Color to tint the light passing through thinner regions. Match your texture’s gamma to 1.0 and clamp to avoid over-saturation.
- Subsurface Scattering: Enable Subsurface and set Scattering Radius based on pigment density (try 0.1–0.3 for shallow scatter). Use a thin layer thickness (Transmission Depth < 1 cm) to avoid excessive bloom.
- Sheen and Coat: Increase Coat Weight to 0.6–0.8 to simulate the oily surface layer. Keep Coat Roughness low (0.05–0.1) for a subtle highlight without mirror-like reflections.
- Micro-Detail and Bump: Use a high-frequency noise or grayscale pigment-clump map plugged into the Bump Input. Scale the bump to 0.05–0.1 cm to capture fine particle texture without breaking the smooth flow of the fluid.
Begin by creating a Material Builder node. Inside, drop a Principled Shader and connect your color maps. Switch the shader’s Specular Model to GGX for realistic sheen. Under Subsurface, choose “Burley” and dial in the Radius parameter to soften edges where the fluid pool is thickest. In the Transmission tab, adjust Roughness to around 0.2 so that highlights remain crisp but slightly diffused by oil content.
For micro-detail, project a tiling noise or scanned lipstick pigment mask across the UVs of your FLIP mesh. Use a UV Unwrap SOP to bake UVs on the fluid’s rest geometry. Then feed that map into both the Bump input and the Specular Roughness slot (inverted) to create tiny bright spots where pigment clumps scatter the light. This procedural layer will catch subtle ripples and surface tension effects when you ray-trace with Karma or Mantra.
Finally, iterate with test renders at 64×64 pixel resolution first, then step up to 512×512 for final look-dev. Compare your renders against reference shots of real foundation under LED softboxes. Pay attention to edge falloff: too much transmission will over-light the rim, while too little SSS will flatten the appearance. Fine-tune coat weight and bump strength until the fluid reads as a creamy, photorealistic foundation.
How do I optimize, cache, retime, and export FLIP sims for rendering and compositing in a beauty pipeline?
In a beauty pipeline, sim optimization starts at the DOP level. Use a tight FLIP liquid container by adjusting the dynamic bounding box to your model’s region of interest. Increase particle separation only as much as the final shot resolution allows—start at 0.01 m and tweak for smoothness versus performance. Activate the Particle Mask to drop unused regions, reducing memory and compute overhead.
Cache your sim early using a File Cache SOP or a ROP Geometry output node. Point it to a .bgeo.sc sequence named by frame ($HIP/flipsim/$HIPNAME.$F4.bgeo.sc) and enable “Incremental Save” so you can resume mid-sim. Bake custom attributes (e.g. thickness, vorticity) before caching via an Attribute Create SOP inside your DOP Import chain, ensuring all render channels are persistent.
- Fetch only P, v, age fields in the DOP Import SOP to minimize memory.
- Compress .bgeo.sc with gzip for faster disk I/O without quality loss.
- Convert to VDB via Volume Rasterize Particles for lighter surface geometry.
- Cache collision meshes separately to isolate dynamic updates.
To retime without rerunning the solver, load cached particles into SOPs and apply a TimeBlend SOP with a custom speed multiplier. Rebuild the liquid surface via an IsoOffset or VDB Combine workflow on the blended particle set. For slow-motion, set TimeBlend to 50% speed, then use a TimeShift to freeze or reverse select frames—all without touching the original FLIP sim.
Finally, export for rendering and compositing using a ROP Alembic or ROP Geometry to write .abc or .bgeo sequences. Include velocity channels for motion blur. In Mantra or Karma, define AOVs for thickness (subsurface masks) and vector (2D motion blur pass). Keep your liquid mesh packed to reduce memory footprint and ensure smooth handoff to Nuke or After Effects.
How can I get downloadable project files, Houdini Digital Assets, and training to reproduce this workflow?
What’s included in the download (scenes, shader presets, VEX snippets)?
This package delivers a turnkey set of assets to recreate a liquid foundation simulation. You’ll receive:
- Houdini scene files (.hip and .hiplc) covering base FLIP setup, emitter rig, and final render scenes.
- Shader presets for Mantra and Redshift: custom surface displacement, subsurface scattering profiles tuned for makeup.
- VEX snippets in .vfl files: attribute initialization for droplet size variation, wetmap generation, and adaptive viscosity blending.
- Documentation PDF: node graphs, parameter notes, and performance tips for production-scale sims.
- Video tutorial links: step-by-step screencasts demonstrating key procedures and troubleshooting advice.
How to install and use the HDA and integrate it into your pipeline?
Begin by copying the .hda file into a folder listed in your HOUDINI_OTLSCAN_PATH. In Houdini’s asset browser you’ll see “LiquidFoundationSolver” under the FLIP category. Drag it into any geometry network to auto-wire the emitter, velocity fields, and collision proxies.
To integrate into your studio pipeline:
- Version control: store the HDA in Git or Perforce, tagging releases to track updates and avoid breaking changes.
- Expose critical parameters: open the Type Properties, promote viscosity, surface tension, and wetmap resolution sliders so artists can tweak per shot.
- Automated setup: add a shelf tool that instantiates the HDA, applies the default shader from the presadv.json, and creates render ROPs with preconfigured camera and lighting.
- Documentation link: embed a clickable PDF reference directly in the asset’s help menu so artists can jump to specific sections without leaving Houdini.
With this HDA installed, any new shot inherits a standardized FLIP workflow. Artists focus on creative look-dev while the procedural rig handles caching, collision integration, and final shading passes.
