For Very Long Baseline Interferometry (VLBI), the fringe spacing is extremely narrow compared to the field of view imposed by the primary beam of each element. This means that an extremely large number of resolution units can potentially be imaged from a single observation. We implement and test a technique for efficiently and accurately imaging large VLBI datasets. The DiFX software correlator is used to generate a dataset with extremely high time and frequency resolution. This large dataset is then transformed and averaged multiple times to generate many smaller datasets, each with a phase centre located at a different area of interest. Results of an 8.4 GHz four-station VLBI observation of a field containing multiple sources are presented. Observations of the calibrator 3C345 were used for preliminary tests of accuracy of the shifting algorithm. A high level of accuracy was achieved, making the method suitable even for the most demanding astrometric VLBI observations. One target source (1320+299A) was detected and was used as a phase-reference calibrator in searching for further detections. An image containing 13 billion pixels was constructed by independently imaging 782 visibility datasets covering the entire primary beam of the array. Current implementations of this algorithm and possible future developments in VLBI data analysis are discussed.
Download