IDL (www.rsinc.com) is the official language for working with AzTEC data. While IDL has a history of reasonable forward-compatibility, please note that the instrument team has written the official analysis pipeline with IDLv6.0 and makes no guarantees that the data reduction software produced will work with earlier versions.
AzTEC produces copious amounts of data. The raw data rate for all of the signals collected from the instrument is ~320MB/hr. Thus, a typical night of observing requires 4.5 GB of storage. While most users ultimately will not be interested in the raw data, initial users will need to be able to work with data structures of this size.
Users are encouraged to contact the instrument team prior to planning observations to discuss the data analysis process. Several options for instrument team involvement in reducing the data are available depending on the level of scientific collaboration.
The following diagram shows the first in two levels of processing in our standard reduction. The raw timestream data is "cleaned" by a principal componant technique to identify and remove modes in the data dominated by the atmosphere. This procedure produces "cleaned" timestream data which can then be turned into signal maps and noise maps (noise maps are made by randomly multiplying scans by +/- 1).
We also keep track of the effects of the atmosphere cleaning on the instruments point spread function (PSF). A typical AzTEC PSF, what we call the "point source kernel", is shown below.
Finally, we take the various products of the first stage of the analysis and use them to construct the final set of data products as shown below.