IPP Software Navigation Tools IPP Links Communication Pan-STARRS Links

Changes between Version 72 and Version 73 of ppStack_testing_201111


Ignore:
Timestamp:
Mar 19, 2012, 1:39:18 PM (14 years ago)
Author:
Mark Huber
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • ppStack_testing_201111

    v72 v73  
    44
    55Discovered for many inputs into ppStack (i.e., deep and even refstacks), that the PSF in the output convolved stack grows with the number of inputs beyond the initial target PSF.
    6 Had been missed due to convolved stacks really only being used for magic, but was noticed with some M31 analysis and deepstacks of MD04. Memory use has also been an issue with the deepstacks and was simultaneously being investigated.
    7 
    8 Test example: MD04 skycell.055 somewhere, ~100 input warps
     6Had been missed due to convolved stacks really only being used for magic, but was noticed with some M31 analysis and deepstacks of MD04. Memory use has also been an issue with the deepstacks and simultaneously being investigated.
     7
     8Example: some MD04 skycell.055, ~100 input warps
    99 * FWHM input average from log = 4.48+-0.81 (pixels)
    10  * largest input has FWHM_MAJ of ~7.2 so clipped?
    11  * target FWHM = 6.88, kron~6.29 from log
     10 * largest input has FWHM_MAJ of ~7.2 (so clipped, since PSF.INPUT.MAX is 7.5 for STACK_DEEP? but PSF.INPUT.CLIP.NSIGMA is 5.0 so no)
     11 * target FWHM from log  6.88, Kron~6.29 from convolution part of log
    1212 * convolved stack FWHM_MAJ=14.36
    1313 * for comparison, unconvolved stack FWHM_MAJ=4.32
    … …  
    1616
    1717
    18 MD04.refstack.20111012 ~92 inputs
    19 
    20 MD04.deeptest.20111012 ~300 inputs
    21 
    22 May also vary
     18More specific examples:
     19
     20 * MD04.refstack.20111012 ~92 inputs
     21{{{
     22      Input FWHMs : 4.010976 +/- 0.334657
     23          fit ext:   1.439042 sec for 225 of 225 sources
     24          fit psf:   0.395260 sec for 172 of 225 sources
     25        apresid: 0.022497 +/- 0.005780 (systematic error) from statistics of 225 psf stars (172 used)
     26        chisq vs flux fit term 0: 0.000000 +/- 0.000000
     27        chisq vs flux fit term 1: 0.000000 +/- 0.000000
     28      try model PS_MODEL_PS1_V1, ap-fit: 0.022497 +/- 0.005780 : sky bias: 0.000000
     29      Target seeing FWHM: 4.706741
     30--> output convolved PSF, end of log
     31        fwhm (psf): 10.825166,10.697605 (moments): 0.000000,0.000000
     32}}}
     33
     34 * MD04.deeptest.20111012 ~300 inputs
     35{{{
     36      Input FWHMs : 4.835018 +/- 0.719810
     37          fit ext:   1.385216 sec for 225 of 225 sources
     38          fit psf:   0.441030 sec for 171 of 225 sources
     39        apresid: 0.055969 +/- 0.001243 (systematic error) from statistics of 225 psf stars (171 used)
     40        chisq vs flux fit term 0: 0.000000 +/- 0.000000
     41        chisq vs flux fit term 1: 0.000000 +/- 0.000000
     42      try model PS_MODEL_PS1_V1, ap-fit: 0.055969 +/- 0.001243 : sky bias: 0.000000
     43      Target seeing FWHM: 6.630503
     44--> output convolved PSF, end of log
     45        fwhm (psf): 14.636683,14.511685 (moments): 0.000000,0.000000
     46}}}
     47
    2348
    2449
    2550Other issues:
    2651 * hard limit on number of inputs (was shell descriptors limit of 1024 easily being reached with image,mask,weight,psf,cmf files for each input. Raised to 65535 now by default.)
    27  * selection of target PSF FWHM 
     52 * selection of target PSF FWHM, worse input PSF not necessarily best choice as could end up being rejected and only a few inputs may be that large (not representative)
    2853 * heavy RAM use (10-20GB typical for 100+ inputs)
    2954 * long run-time for 20+ inputs, 400 inputs order 12-14hr (use method similar to psphotStack?)
    … …  
    355380
    356381
    357