Manual Modules Dynamics Solver Object
Solver Object Basic Coord. Main DetailsMain Tab
These two parameters define when the dynamics animation will begin and when it will stop. The Star and Stop parameters are independent of the project’s animation length. If you set Stop to a time point beyond the project length, you will need to increase the Maximum value in the Project Settings dialog.
Dynamics must solve a number of complex differential equations in order to calculate the motion of bodies in the dynamics system. The scientific term for this entire calculation is ‘integration’. Various integration methods can be used to solve these equations. The Integration Method parameter defines which particular method the solver will use.
The following integration methods are available:
The least accurate but fastest to calculate method. It is about half as accurate but twice as fast at Midpoint. Use this method for simple collision animations where no soft body objects are involved.
Fast to compute, but not very accurate.
Generally about 10 times more accurate than Midpoint, but four times slower.
The most accurate method; slowest method to process.
Use this method if soft bodies are involved. Usually you can set Oversampling to a low value when using this method, such as 2 or 4. The method computes quickly, especially with soft body collisions, and can be as much as 100 times faster than the other methods. (Although some of the other methods are generally fast to compute, they would be slowed down by requiring a much higher Oversampling value than is needed with the Softbody method.)
No matter which method you choose, the motion will be approximated to some degree. Your aim when setting the Integration Method and its associated parameters is to find a suitable balance between accuracy and processing time.
For the Midpoint and Runge-Kutta integration methods, the Oversampling value defines the number of times the motion will be integrated per frame. Hence to increase the accuracy of the motion, select a higher Oversampling value. Note that a higher value will increase the processing time.
For the Adaptive integration method, the Oversampling value defines the minimum number of integrations per frame. The actual number will depend on the Subsampling value.
The bottom curve represents the perfect motion. The two other curves use the same integration method, but the middle curve uses a higher Oversampling value, lending it greater accuracy.
You can define Subsampling for Adaptive integration only. Subsampling works together with Oversampling to define the number of integrations per frame for Adaptive integration. Oversampling defines the minimum number of integrations per frame. The Oversampling value multiplied by the Subsampling value defines the maximum number of integrations per frame.
For example, with Oversampling set to 4 and Subsampling set to 8, there will be a minimum of four integrations per frame and a maximum of 32 integrations per frame. With Oversampling set to 2 and Subsampling set to 4, there will be anywhere from 2 to 8 integrations per frame.
Subsampling concentrates on critical areas of the motion, as illustrated below. The actual number of integrations will vary from frame to frame.
It is thanks to Subsampling that Adaptive is the most accurate integration method. With subsampling, integration will be concentrated in the most critical areas.
With subsampling enabled, integration is concentrated in key areas of the animation. In the diagram, to start with only oversampling is applied; then, as a critical point in the motion is reached (the peak in the diagram), which might be for example a change in direction, more and more subsampling is applied, up to the maximum selected. Eventually the motion becomes less critical and we go back to oversampling.
This parameter is necessary since dynamics engines tend to add energy into the system when the Integration Method is not accurate enough; adding a small energy loss will help to compensate for this slight inaccuracy.
This is similar to real life motion which continually loses energy due to various drag forces. However, if you should notice that, for example, spring oscillations are damped too quickly, reduce the Energy Loss value or the drag or damping within the force object.