These methods do not, in themselves, cause the data to be read, but a mapping is set up, such that when the pixel data for a frame is required for display, compression, transmission etc., then it is read as needed. This is therefore useful for the conversion of long cine sequences to DICOM, as only one frame at a time is held in memory.
SetPixelFileExt uses a double instead of a long integer for the offset, allowing use of data segments within massive (>2 GByte) files. Although use of a floating point value is unusual for this type of application, double precision values can hold integers accurately to 48 bits (allowing files up to 256 TBytes!)
SetPixelFiles uses one file for each frame – all starting at the same offset (normally 0) within each file.
For all versions except SetPixelFileCompressed, the pixel data must be uncompressed in a format which matches the bits allocated specified in the DICOM data.
The number of frames is set, as usual, by element (0028,0008), which should be set before invoking these methods.