IRAC Test files
version 19-March-2001

- use WFPC2 files 800x800 single plane (2 axis) of R136 
- cut out a 256x256 section for the IRAC test data
- rewrite the fits headers to make them look more like IRAC data
---------------------------------------------------------------------------------
Test files for one Data Collection Event (DCE) (one of 4 IRAC bands)
This is the Basic set of data.

1)  BCD file
    - make it one plane for now. the new IRAC Pipeline Requirements v1.0
      seems to put the photometric uncertainty data in a separate file.
    - image includes
      - cosmic rays
      - dead pixels
      - saturated pixels
      - banding effects

Ancillary data

2)  Photometric uncertainties   
    - represents estimate of 1 sigma internal uncertainty in the flux 
      for each pixel (use err=sqrt(counts))

3)  Bad pixel map 
    0 = OK
    1 = dark current too high
    2 = dark current too unstable
    4 = responsivity too high
    8 = responsivity too low (dead)
   16 = responsivity unstable 
   32 = could not be linearized
   64 = missing pixel (pixel was filled in by SSC pipeline)
  128 = bad pixel that does not fall into above categories

4)  Data quality mask
    0 = OK
    1 = saturated pixel
    2 = pixels affected by banding (only in bands 3 and 4; correction applied 
        in SSC pipeline)  

5)  Ghost Image 
    Bright point source in one channel causes a ghost image in another channel
    SSC provides an estimate of the level
    
6)  Cosmic Ray map
    - SSC criteria:  >5 sigma above local bkg
     0 = OK
     1 = CR hit 

7)  Latent image mask
     0 = pixel OK
     1 = latent image - medium expected detectability (<75% probability) 
         (may still be able to use data)
     2 = latent image - large expected detectability (>75% probability) (cannot use data) 

8)  Latent image
    SSC provides estimate of latent image level

-----------------------------------------------------------------------------------------
Next:

Provide test data for all 4 IRAC bands - observing same field
Change the noise level in at least one of the BCD images

Next:

Provide test data for a revisit of the same field.  change image slightly (rotate, shift)
to mimic a revisit.

Next:

Provide several of the above sets, to mimic an AOR's worth of data.
(say a tile of 3 frames across, 20 down)