RBF Settings

The RBF Settings section of the RBF Drivers panel is where you control how the selected RBF Driver interpolates between all your poses.

---

Interpolation

Each of the available interpolation types will change how closely the driven properties stick to the poses you have defined, as well as how smoothly and quickly they transition between then. You don't need to know how each of them works, you can simply select which one works best for you by switching between them and playing around in the viewport.

Linear

Linear interpolation will simply transition linearly between the poses based on the current inputs. The driven properties will remain very close to the defined poses. This works well when you want a very tight fit to your data points, particularly for smaller data sets.

Gaussian

The gaussian option will provide a smooth, bell-shaped curve, easing the transitions between poses in and out. It's often a good choice if your data points are somewhat scattered. You can affect the transition timing by changing the radius setting, or providing a custom radius for complete control.

Multi-Quadratic Biharmonic

Multi-quadratic biharmonic interpolation generally provides a nice median between the smooth characteristics of gaussian and strictness of linear and is the default setting. Your mileage may vary depending on your data set but it's usually my preferred option.

Inverse Multi-Quadratic Biharmonic

Inverse multi-quadratic biharmonic offers a slight variant on multi-quadratic biharmonic, pushing things a bit further at the ends.

---

Radius

Mean

The radial basis function will use the average distance between all data points.

Variance

The radial basis function will use the variance of the distances between all data points.

Standard Deviation

The radial basis function will use the standard deviation between all data points. This is probably the most commonly used dispersion measure in RBF networks and the default for RBF Driver

Custom

By selecting a Custom radius, an input field will appear where you can input whatever radius you wish. You're unlikely to use this option unless you're familiar with your data set and how the interpolation functions work but feel to play around providing different values here to see the effect.

---

Smoothing

Increasing the smoothing value will relax the fit of the driven properties to the poses, essentially increasing the visual effect of the interpolation. Under most circumstances the interpolation alone will provide pleasing results, but do play around with the smoothing to become familiar with its effect.