If the hardware configuration is too limited for smooth movement control at full resolution, higher frame rate can be achieved at the price of a lower image quality until the interaction is stopped.
Sub-sampling consists of generating an image smaller than the viewer size (i.e., lower resolution), which is then stretched to actual size. Rendering is proportionally faster as fewer primary rays are shot from the camera. Note that in cluster mode this also saves network bandwidth, contributing to higher frame rates.
Once sub-sampling is enabled, low resolution is displayed while the scene graph is modified (animation, interaction with manipulators, viewer navigation...). Full resolution is restored after a small delay when interaction stops. Note that this is different from the viewer's draw style options available with OpenGL rendering, e.g., “move bounding box”, which apply only during navigation interaction (camera updates).
This feature is enabled by the SoPreferences SoPreferences SoPreferences DIRECTVIZ_SUBSAMPLING parameter or SoDirectVizManager SoDirectVizManager SoDirectVizManager subSampling field. The sub-sampling factor applied to the original image size is a float between 0. and 1. set via the SoPreferences SoPreferences SoPreferences parameter DIRECTVIZ_SUBSAMPLING_FACTOR or SoDirectVizManager SoDirectVizManager SoDirectVizManager subSamplingFactor field.
