Volume Viewer - precomputed subsamplings

The use of precomputed subsamplings in Volume Viewer decreases memory usage and time spent reading files.

The Step values (cell sizes) used to generate a subsampling are inferred by comparing the numbers of data points in the full data set and in the subsample. For each dimension, valid cell sizes S for full data size N and subsample size M satisfy (M-1)S < N < (M+1)S. If there are no valid cell sizes, the subsample file will not be accepted, and a warning message will be issued. If there is more than one valid cell size, the user is queried. Examples:

• If the full data set is (100,100,100) and the subsample is (33,33,33) or (34,34,34), the inferred Step values are (3,3,3). If the subsample is (32,32,32), a warning message will be issued.
• If the full data set is (20,20,20) and the subsample is (3,3,3), possible Step values in each dimension are 6, 7, 8, and 9. The user will be asked to specify the correct Step values.

The various settings in the Region size panel result in target Step values. If a subsample is inferred to have been generated with these same values, it will be used. If not, then another subsample that includes the desired subsample will be sought and used, if possible. For example, if a (2,2,4) subsampling is the target but does not exist, an existing (2,2,2) subsampling can be used to generate it. A (2,2,3) subsampling could not be used in this situation. If no suitable subsample is available, the full data will be used.