While radio telescope arrays make some of the highest resolution maps of the sky, one thing they are not generally very good at is making images that cover large areas. But new telescopes that are planned or currently under construction (such as the Australian SKA Pathfinder, ASKAP, being built here in Western Australia) are going to make wide field images as standard. This is exciting stuff for radio astronomy as, for one thing, it allows surveys to be conducted much faster. If you can only look at very small patches of sky at a time, it's going to take you a long time to cover the whole sky, but if you can widen your field of view then you can survey the same total area in a much shorter time (it's kinda like doing a jigsaw - the smaller the pieces, the longer it takes). This means that with new telescopes we will be able to do both more sensitive surveys (staring at patches of sky for longer to detect fainter sources), and more efficient searches for transient sources such as radio supernovae and gamma ray bursts (and other things we probably don't even know about yet).
It is possible to produce wide field images from existing radio arrays, but it's not exactly "normal operating procedure". With any kind of telescope, the resolution is determined by the wavelength you're observing at and the diameter of your telescope. For an interferometer, the diameter used is the largest separation between your antennas, which can be several thousand kilometres for large arrays. This makes them capable of very high resolution images but, due (amongst other things) to averaging effects, the standard data outputs from a correlator are only usable over a relatively small field of view around the pointing centre, the point on the sky your telescopes were aimed directly at during the observation. In principle though, the observations are sensitive to sources over a much larger area, it's just that using normal methods we can't make reliable images far from the pointing centre.
Well, earlier this month we observed two fields in the nearby Andromeda galaxy (M31), one using the US-based Very Long Baseline Array, the other using the European VLBI Network. The data are on the way to us, and when we get our hands on them we are going to have some fun. Instead of just using the correlated data around the pointing centre, we are going to use a software correlator, running on a computer cluster here, to re-correlate the baseband (raw) data wherever we like over the area the observations are sensitive to, with the ultimate goal of imaging the entire primary beam of the array. This means we will be able to make reliable images of sources across the primary beam of the array, essentially the area covered by the beam size of the individual telescopes in the array - in the case of the VLBA observation, we will be able to make images of sources with a resolution of 5 mas (1 mas = 1/1000 arcseconds = 0.0000003 degrees) over an area about half a degree in diameter.
The nearby spiral galaxy M31 (Andromeda) in both optical (blue) and radio (orange). Our new radio observations will have a resolution some 7000 times greater than the radio observations used to make the above composite, but covering a similar total area. CREDIT: Optical: DSS; radio: VLA, Beck et al
Think about that for a moment. Half a degree - that's the size of the full Moon. If we were looking at the Moon, this is equivalent of being able to resolve features less than nine metres in size. We'd almost be able to spot the lunar landers! Imagine mapping the surface of the Moon at a resolution of nine metres. That's a lot of pixels.
So what are we hoping to see in our terabytes of data? Lots! Supernova remnants, star forming regions, planetary nebulae, and heaps of background quasars... may be something unexpected. This is going to be sooo cool!