== IRC imaging toolkit version 20150331 ==

This is a major update of the IRC imaging toolkit for Phase 1 & 2, 
which has been developed and tested with IRAF v2.16 & 2.16.1. 
All the users are recommended to use this version.
The changes from the previous version (20131202) are as follows.


++ Summary of updates ++

(1) Process logs
   'prepipeline', 'pipeline1', and 'pipeline2' record the computer 
   environments and parameters used in the pipeline processing. The 
   name of the log file (default='ircpipeline.log') can be controlled 
   by the new parameter 'pllog'.

(2) Neighbor dark frames
   For each pointed observation, new MIR-S and -L dark frames are created
   by combining pre-dark frames taken within +/-5 pointed observations.
   They are called neighbor dark frames and added to the data package
   (i.e. not included in the toolkit). A new option for the parameter 
   darktype='ne' is available as default in 'pipeline1'. Note that they 
   are for long exposure frames only.

(3) New MIR-S flat frames
   Before 2007 January 7, a bean-shaped pattern called 'soramame' 
   appeared in the bottom right of MIR-S FoV. Past toolkits used only 
   one flat frame for each MIR-S filter to correct this pattern, but its 
   shape was found to vary with time. The period with 'soramame' is 
   divided into four periods (p1, p23, p4, p5) and a new flat frame for 
   each period and filter is created. Since p1 was too short to create 
   high S/N flat frames, new flat frames only for the remaining three 
   periods (i.e. p23, p4, p5) are included in this version.

(4) Hot pixel masking
   A pixel exceeding a certain threshold in the neighbor dark frame is 
   defined as a hot pixel and masked out. The threshold is controlled 
   by a new parameter 'ndhot' (default=500.) in 'pipeline2'. The masking 
   can be switched off by setting ndmask=no.

(5) Sub pixel sampling
   By default, a single pixels is divied into 2x2 pixels in 'aspect_ratio'
   called by 'pipeline2'. This procedure can be turned off by setting 
   double=no in 'pipeline2'.

(6) Calculating relative shift values between frames
   For pointed observations without dithering (i.e. AOT=IRC00 or IRC05), 
   shift values in x- and y-direction and rotation angle of each frame 
   relative to the first frame are examined. From their histograms, 
   we define the lower and upper limits to be accepted in the shift
   calculation. For those with dithering (i.e. AOT=IRC02 or IRC03), 
   histograms of shift in x- and y-direction are created for each 
   dithering cycle. As a consequence, the lower and upper limits are 
   defined for each cycle. If calculated shift values are outside of 
   the limits, the toolkit regards this calculation is doubtful and 
   exclude the the corresponding frame from the list for staking.
   In addition, the minimum number of sources to be used for the shift 
   calculation is changed from 4 to 7, in order to reduce a chance of 
   false matching

(7) WCS registration for MIR
   WISE W3 (12um) catalog is used for WCS matching of MIR images.
   For NIR images, 2MASS catalog is used as before.

(8) Flux calibration
   Flux calibration factors by Tanabe et al. (2008) are applied to the 
   stacked images. A data unit of the output images is [microJy/pix].

(9) Set instrument in IRAF v2.16.1
   In the latest IRAF (v2.16.1), an instrument should be explicitly 
   specified. A new blank file 'lib/setinst.dat' is added and is set 
   as the instrument file when loading the IRC package.

(10) Memory effect in MIR
   If (an) extremely bright object(s) fall(s) within the MIR FoVs, the 
   corresponding pixels decrease the sensitivity for a certain amount 
   of time. As a result, a weird negative pattern appears in the successive 
   observations, which is called memory effect. If such objects were 
   observed in pointed observations, copying a saturation mask for the 
   bright objects and adding it to the affected data often masks the 
   affected pixels out and improve the quality of stacked images. New 
   tasks 'copy_MIRmask' and 'add_MIRmask' in IRC_TOOL can be used for 
   this purpose. Information on the plausible affecting observations is 
   in README of each observation in the processed data package.
   Please note that if bright objects were observed in the all-sky survey 
   or in the slow-scan mode, the memory effect appears as vertical stripes 
   or arc-like shapes. In these cases, the new tasks mentioned above 
   are not applicable.


++ Known issues ++ 

(a) Multi-pointing
   This toolkit is not tested for stacking data from multiple pointed 
   observations.

(b) Drift
   Due to the settings of star trackers installed in AKARI, the telescope
   sometimes drifted during pointed observations. In the worst cases, 
   point sources appear as elongated shapes in NIR long exposure frames.
   In addition, some MIR frames fail to be stacked because relative shift 
   values fall outside of the accepted range by the toolkit. 

(c) Earthshine light
   A fully automatic scheme to remove the EL is not yet available.
   As for the memory effect, tasks for making and subtracting the EL 
   pattern ('mk_el' and 'sub_el') is available in IRC_TOOL.

(d) Anomalies
   NIR column pulldown and muxbleed are not removed. Ghosts appear on 
   small (~1arcmin) and large (~1deg) scales but are not masked out.
   A warning for small-scale ghosts in MIR-S is recorded in the process log.


++ Computer environments ++

This toolkit has been tested under these environments:

- Mac OSX 10.6.8, Perl v5.10.0, IRAF v2.16,   csh
- CentOS 5.10,    Perl v5.18.1, IRAF v2.16,   tcsh
- Mac OSX 10.9.5, Perl v5.16.2, IRAF v2.16.1, zsh
- CentOS 7.0,     Perl v5.16.3, IRAF v2.16.1, tcsh


++ Appendix ++

(i) Running the toolkit with data package released in Mar. 2015
   - obtain rawdata from DARTS
   - copy "dark/" directory in the same directory as "rawdata/" 
   - create a working directory in the same directory
     (i.e. rawdata, dark, and the working directory should all 
      be under the same directory)
   - run the toolkit under the working directory

=== EOF