See volumes for an explanation of standard volumes and OpenVDB volumes.
Unlike the VDB Combine SOP, this provides a multi-input so can merge more than two grids without needing an additional merge SOP.
Float and double SDF VDBs and fog volume VDBs can be merged, all other VDB grids are left unchanged.
Grids with different transforms will be resampled to match the first VDB in the collation group.
A subset of the input VDBs to be modified (see specifying volumes)
The criteria under which groups of grids are merged. In addition to these collation options, only grids with the same class (fog volume, level set, etc) and value type (float, int, etc) are merged.
Collate VDBs with the same grid name.
Collate first primitives from each input, then second primitives from each input, etc.
Collate all VDBs (provided they have the same class and value type).
The type of interpolation to be used when resampling one VDB to match the other’s transform
Nearest neighbor interpolation is fast but can introduce noticeable sampling artifacts. Quadratic interpolation is slow but high-quality. Linear interpolation is intermediate in speed and quality.
Merge operation for Fog VDBs.
Leaves input fog VDBs unchanged.
Generate the sum of all fog VDBs within the same collation.
Merge operation for SDF VDBs.
Leaves input SDF VDBs unchanged.
Generate the union of all signed distance fields within the same collation.
Generate the intersection of all signed distance fields within the same collation.