Jump to: navigation, search
  • Software: the following MIRIAD programs are under renovation, mostly by Peter
    • varmaps: should be ok, but interpolation is tricky to set via beam size and cell size. Also perhaps a rounding error showing up as aliasing of rows/columns, that didn't show up in NEMO's snapgrid counterpart. Peter has a script in MIS/templates that will excersize this pipeline test.
    • sinbad: interpolation of two nearest OFF's might be useful. Right now it's using the last or if not available, next available
    • imstack: median filtering now done, including with 3D input cubes. Has options=mean now. also pay attention to missing data? only pays attention to flags now - mar 26 PJT
    • blflag: would be nice to have interactive flagging (explanation: blflag doesn't compile on default 64bit, PJT will look into this). Alternatively, viewing via uvimage could be instructive.
    • uvflag: can we flag the source data based on an amplitude cutoff of the CALIBRATED data (instead of the raw data) - no,
  • Antenna 20: Once the observatory has a solution for the position of antenna 20, we will need to go back through the data to correct its position and recalibrate the data to see if we can recover antenna 20. I would suggest that we go in reverse chron order since ant 20 is most likely to be good near the end of the E config and was definitely not working at the beginning of our project.
    • Update: New baseline solution received, and all E-array data has been re-analyzed by Shaye and Demerese. We will list our flags in the Data Summary section.
  • Flux Calibration: Currently the pipeline uses bootflux=1 in do_cal1, which assembles all information in bootflux.log files. See the pipeline script how this resulted in us concluding us that from 10 trials that have Uranus included, that the flux for 3c84 can probably be fixed at 12.5 Jy for March 2011. There are some hints the flux for 3C84 is currently decreasing again. Checking on calflux that appears to be a decrease of about 0.2-0.3 Jy/month, but their value for the n1333 campaign is more like 12.0 Jy!
  • Flagging Bad Source Data: how do we plan to flag bad on-source data (see uvflag note above)?
  • Linecal: are we going to apply linecal in the final data reduction? (there is now a linecal=0 command line option do do_cal1)
  • CSflag: should we flag shadow data? (there is now a command line option csflag=0 to do_uvflag, but right now do_uvflag is added to the two do_uvcat styles)
  • SD mapping: the final cube is currently derived from 10 good trials (see $MIS/cat/n1333sd.lis), but t least 5 more (and perhaps up to 10) have useable SD data where the pipeline currently chokes on. Current cubes look ok, but we might be able to improve a bit on them.
  • do_mos1: put in an option to run in parallel, useful to speed up on multicore machines (and with enough memory)
Personal tools