Methods and apparatuses for performing Z-buffer granularity depth calibration in graphics displays of three-dimensional scenes

Abstract

The weight given to each object in a scene in the computation of the z-buffer granularity is determined based upon a variable normalized risk parameter which may be supplied by the application. A recursive composite parent object process is invoked upon the two children of the scene object in order to allocate graphics resources to objects A and B. Primary and secondary hit value means corresponding to objects A and B are computed. The hit value means are the average sum of hit values recorded for all the leaf objects contained by the objects over a predetermined number of frames. The statistical variances of the primary hit value and secondary hit value are also computed over the previous predetermined number of frames in order facilitates the mapping of the normalized risk parameter to a non-normalized risk parameter indicating the optimal risk for objects A and object B. A quadratic parametric variance equation is solved for the optimal fraction of the remaining graphics resources to be allocated to object A. The optimal fraction is multiplied by the available graphics resources, resulting in the resources allocated to object A; the remaining graphics resources are to object B. If either object A or B is a composite parent object, the recursive parent object process is invoked on the two children of object A or B with the available graphics resources for the instantiation of the recursive parent object process set to the graphics resources allocated to object A or B. The resources allocated each object are then used as weighting factors in the computation of the z-buffer granularity. In an embodiment, the z-buffer midpoint depth is determined by dividing the sum of the multiplicative products of the distance from the observation point, the allocated resources, and the importance coefficient for each object by the sum of the products of the observation point and the allocated resources for each object.

Dependencies are inferred from the claim text.