• Frame Sub-Steps Limit - For fast moving objects interacting with fluid or to have fluid behave properly under high-speed pressure or other situations will be more accurate by increasing the Sub-Steps Limit. If your simulation has high velocities and you're seeing noise artefacts in the flow, allow the simulation to take more sub-frame time slices, if necessary, by increasing this value. This will allow the simulation to resolve the fluid motion better. It comes at a cost of simulation time and can drastically impact the look of your fluid simulation. Use it sparingly.
  • Pressure Iteration Limit - Wherever velocity deforms the fluid, pressure is build up. It is a fundamental part of the simulation to equalize this pressure. For simulations that have high velocities and/or collision objects or closed container boundaries, it is recommended to use values between 2 and 10. Values larger than 10 should rarely be necessary. For simulations that don't have collisions and mostly moderate velocities, the simulation time can be kept lower by using the minimum of one iteration.
  • Adaptive container - If enabled, TFD simulates the smallest possible part of the container in order to save time and memory. The Clip Below parameters in the Velocity, Temperature, Density, Fuel and Burn tabs control when TFD may shrink the container.

Velocity/Channel Advection

The advection accuracy affects the sharpness of the details in the fluid.

1st order - produces a somewhat blurrier result where vortices get smoothed away quicker, but it's also faster and uses less memory.

2nd order - produces sharper details, keeps vortices alive longer but uses more time and memory.

  • Velocity Advection Accuracy - You can choose between 1st order and 2nd order accuracy here.
  • Channel Advection Accuracy - You can choose between 1st order and 2nd order accuracy here.
  • Adaptive Tracer - When advection of the fluid takes place, the simulation traces a trajectory through the velocity field for each voxel. The fluid moves along these trajectories. The default tracer approximates the trajectories with a single line segment. The Adaptive Tracer uses many line segments, such that none of them are longer than the voxel size.
  • Cubic Interpolation - Moving the channels along the flow requires repeated interpolation. This interpolation slightly blurs the channel each time. Using cubic interpolation reduces this blur at the cost of somewhat higher simulation times.
  • Use less memory but more time - This is effectively a compression being applied in real-time to the cache data. Since fluid simulation can be very memory intensive, you are likely to exhaust your system's memory with large simulations. In order to allow you to run larger simulations on machines with less memory, checking this option will let TurbulenceFD use less internal caches which results in less memory being used but more simulation time being required. Render times are not affected. This will produce smaller .bcf files on a stored drive.
  • Collision Objects Enlarge Container - The adaptive container will be enlarged to contain all collision objects by default. This may not be necessary, especially if the objects aren't moving. In such a setup, you can uncheck this option to keep the container smaller.
  • Smooth Collision Surface Rendering - When rendering solid obstacles emerged in the fluid, extrapolation will avoid artefacts on the surface of the obstacle. If you have invisible or transparent obstacles, disable this option.
  • Close Boundaries - By default the sides of the fluid container are open. The fluid can just move out of the container and disappear. If you want to simulate a closed or partially closed space, you can select the sides of the container that will be closed. The checkboxes are labeled according to the half-spaces of the coordinate system. That is, +X is the side of the container in the positive X direction and so on. For example, a ground explosion can be simulated in a container with the -Y side closed to avoid having fluid leave the container through the ground. Making use of these check boxes in certain scenes can reduce the need for a "wall, ceiling, floor" collision object to be used in the Emitter Tab as a collision object when the effect of having one is desired, thus reducing simulation times.

  • Upres Scale - When running an Up-Res'ing post-process by clicking the Up-Res button, the resolution of an existing simulation will be increased. This parameter specifies how much larger the resolution will be.
  • Fine Turbulence Intensity - The Up-Res'ing post-process also adds small scale turbulence to the fluid that will create additional detail. This parameter specifies the strength of this small scale turbulence.
  • Fine Turbulence Small Power - Specifies how stong small turbulence detail is compared to the next larger size. This works just like the Simulation/Turbulence/Small Power parameter but it's applied only for the additional detail added during Up-Res'ing.

  • Time Scale - This value controls how fast time will pass for the simulation. At 1.0, the simulation will run at the same speed as the animation. Values below 1.0 slow the simulation down, values above 1.0 speed it up. You can for example animate this value down to 0.0 to create bullet-time-like animations.
  • Frame Range - This setting specifies what part of the time-line to simulate. It corresponds to Lightwave's Preview Range and Render Range.
  • Start Clean Simulation - Clear all channels before starting the simulation at the frame specified above.
    • If not checked, you can continue from a previously-saved cache.
  • Load Simulation State From File - If Start Clean Simulation is unchecked, channels will be initialized from the file specified here. The file may be any .bcf file from a TurbulenceFD cache directory. Channels that are not available in the .bcf file will be cleared as if Start Clean Simulation was checked. See the parameters in the Cache tab for information about how to control which channels are saved to the disk cache.
  • Continue Simulation From File - If checked, the next simulation run will load the state from the file specified above and jump to the frame following the one cached to the file. Use this to continue a simulation at any cached frame, possibly after changing some parameters.

  • Clip Below - Specifies the minimum velocity in voxels per second that will prevent the container from adapting its size. Velocities below that will be cut off when the container adapts its size. The default value is chosen such that it works with most simulations, including explosions which need lower values than other simulations. However, at the default value the container may be lept larger than necessary. If you want to speed up the simulation by letting the container adapt more tightly to the channels, increase this threshold. A value of e.g. 1000 will effectively ignore the velocity when adapting the container.

    See also the threshold parameters of the channels below.
  • Damp Velocity - Specifies the percentage by which the velocity is reduced in each frame. This simulates drag or friction in the air. This is for example useful for explosions where you may want the fluid to leave the source quickly but then form a cloud that moves slower.

  • Wind Direction - Specifies the direction of a global wind in the container.
  • Wind Speed - Specifies the speed of the global wind in the container.

These two wind operators allow you to define a global wind that is blowing in the fluid container. It can be used to add wind as it would appear in an outdoor scene. To create small, local winds like from an exhaust pipe, use the Normal Force parameter on an emitter object (e.g. a plane or disc).

This set of functions provides amplification to existing curls to keep them alive longer. When using strong amplification, the curls will get stronger and stronger. After some time, the flow may become very noisy. To avoid that, the Intensity Channel and Mapping allow you to restrict the effect of Vorticity to a region that you control using another fluid channel.

  • Vorticity - Sets the strength of amplification for small curls.
  • Intensity Channel - You can control the vorticity in space using any of the fluid channels. Use this to amplify the curls only where the temperature is high enough, for example.
  • Intensity Mapping - For improved control over the intensities, the values of the Intensity Channel are re-mapped by this F-curve before scaling the vorticity.

Adds random velocity noise to the container, making it swirl and a chaotic way as you would expect from turbulent fire or similar. It is driven by a procedural noise function that changes over time. The following parameters control the intensity, scale, etc. of this texture.

  • Turb. Intensity - Specifies the strength of the turbulence.
  • Intensity Channel - You can control the turbulence intensity in space using any of the fluid channels. Use this to add turbulence only where the temperature is high enough, for example.
  • Intensity Mapping - For improved control over the intensities, the values of the Intensity Channel are re-mapped by this f-curve before scaling the turbulence intensity.
  • Smallest Size - Specifies the size of the smallest curls added to the fluid.
  • Largest Size - Specifies the size of the largest curls added to the fluid.
  • Small Power - Specifies how strong small curls are with respect to the next larger ones. A small power of 1.0 will make curls of all sizes equally strong. A small power of 0.5 will make small curls half as strong as curls of twice the size.
  • Speed- This value specifies how fast the turbulence field changes over time.

  • Active - Checkbox. Without this active, the following settings are ghosted. This option is active by default to enable simulation and caching of the temperature channel.
  • Clip Below - If the Adaptive Container option is enabled, this value defines the minimum temperature that will cause the container will consider non-empty. If this value is zero, no voxels containing temperature will be lost. However, often shading settings are such that low values aren't visible anyway. In this case, increasing this threshold allows the sim to keep the container smaller.
  • Temp. Diffusion - Air mixes even if it does not move on the large scale. This effect is called Brownian Motion. Diffusion accounts for this effect by basically blurring the temperature. As a result of high diffusion, the buoyancy force becomes smoother.
  • Cooling - Specifies the percentage by which the temperature is cooled down in every frame. Cooling works like exponential decay in the Burn channel.
  • Half-life - Instead of specifying the cooling intensity as percentage per frame, you can specify the time it takes for the temperature to cool down to half its value. This may be more intuitive in many situations. Note that "no cooling" would mean infinite half-life.
  • Buoyancy - Buoyancy is the force that makes warm air rise and cold air sink. This value specifies the force per unit of temperature exerted on a particle.
  • Buoyancy Direction - specifies the direction of the buoyancy force.

  • Active - Check this to enable simulation and caching of the density channel.
  • Clip Below - If the Adaptive Container option is enabled, this value defines the minimum temperature that will cause the container will consider non-empty. If this value is zero, no voxels containing temperature will be lost. However, often shading settings are such that low values aren't visible anyway. In this case, increasing this threshold allows the sim to keep the container smaller.
  • Dens. Diffusion - Air mixes even if it does not move on the large scale. This effect is called Brownian Motion. Diffusion accounts for this effect by basically blurring the density.
  • Dissipation - Specifies the percentage by which density is dissipated in every frame. Dissipation works like exponential decay in the Burn channel.
  • Half-life - Instead of specifying the dissipation intensity as percentage per frame, you can specify the time it takes for the density to dissipate to half its value. This may be more intuitive in many situations.
  • Gravity - The more dense the fluid is, the more mass it has. Therefore, the more it gets affected by the gravity force. This parameter specifies the strength of that force.
  • Gravity Direction - specifies the direction of the gravity force.

Fuel works like a cloud of combustible gas like propane or gasoline mist. When it burns, it...

  • emits Burn which then represents the burning flame
  • expands the cloud (i.e. it emits pressure)
  • emits heat (see Temp. Emission)
  • emits soot (see Density Emission)
  • reduces the amount of fuel left

Fuel Masking allows you to control where fuel burns. The Fuel Mask is a cloud taken from any other channel, smoothed and re-mapped. Only within that cloud the fuel will burn. This allows you to control ignition and suffocation of burning fuel.

  • Active - Check this to enable simulation and caching of the fuel channel.
  • Clip Below - If the Adaptive Container option is enabled, this value defines the minimum temperature that will cause the container will consider non-empty. If this value is zero, no voxels containing temperature will be lost. However, often shading settings are such that low values aren't visible anyway. In this case, increasing this threshold allows the sim to keep the container smaller.
  • Fuel Diffusion - Specifies how fast fuel diffuses (see Temperature/Diffusion above for more details).
  • Burn Rate - Specifies the amount of the fuel that is burnt in every frame. The faster it burns, the more Burn and Temperature will be created. Burn Rate works like linear decay in the Burn channel.
  • Fuel Mask Channel - If a channel is selected, fuel masking will be active. See the simple-ignition example for a setup using the Burn channel.
  • Fuel Mask Smoothing - Before using the selected channel as a fuel mask, it can be smoothed or blurred. This will also make the cloud a bit larger, thereby propagating the ignition front like you would expect it to.
  • Fuel Mask Mapping - Re-mapping the smoothed fuel mask gives you control over the shape of the fuel mask and how steep the ignition front is. A flat increase will let the burn start slowly the larger the fuel mask values become. A steep increase will make the fuel burn at the full burn rate as soon as it touches the fuel mask.
  • Expansion - A factor that specifies how much the gas expands per burnt voxel of fuel.
  • Temp. Emission - Specifies how much heat is generated per unit of burnt fuel. Density Emission - Specifies how much density is generated per unit of burnt fuel.

  • Active - Check this to enable simulation and caching of the burn channel.
  • Clip Below - If the Adaptive Container option is enabled, this value defines the minimum temperature that will cause the container will consider non-empty. If this value is zero, no voxels containing temperature will be lost. However, often shading settings are such that low values aren't visible anyway. In this case, increasing this threshold allows the sim to keep the container smaller.
  • Burn Diffusion - Specifies how fast Burn diffuses (see Temperature/Diffusion above for more details).
  • Decay Mode - Decaying a channel can be done by a constant amount or by a percentage of the current value. The Decay Mode specifies which will be used.
    • Linear - Decay the burn channel by subtracting a constant value in each frame. If you emit 1.0 into the burn channel, setting Decay to 0.1 in this mode will let the emission live for 10 frames before it drops to zero. This mode works well for flames that have a sharp contour.
    • Exponential -Decay the burn channel by subtracting a percentage of the current value (of a voxel) in each frame. As the current value becomes smaller and smaller, so does the value that is subtracted. Therefore the emission never drops back to zero - only very close. And the speed at which is drops will get slower and slower. This is the type of decay that temperature and density have. It models the way dust slowly dissipates in air. The decision of which mode to use is mostly driven by your shading settings. To get an idea of how the modes differ, compare the histograms in the shader's mapping curve editors. The exponentially decayed channel will have lots of low values and less high values, because the decay slows down the smaller the values get. The linear decay more uniform distribution of high to low values. This makes shading sharp contours easier with the linear mode. If you want the channel values to live long and build up a more smoke-like cloud, the exponential decay works better.
  • Decay - In linear mode, Decay specifies the constant value that will be subtracted from the current channel value in each frame. In exponential mode, it's the percentage of the current value that will be subtracted.
  • Half-life - Instead of specifying the decay intensity as percentage per frame, you can specify the time it takes for the channel value to decay to half its value. This may be more intuitive in many situations. For example, if Half-Life is one frame, the channel value drops to 50% within one frame. That corresponds to a Decay of 50%.