Content

Comparison of psana, cheetah and cctbx

Comments

Chun Hong Yoon <yoon82@stanford.edu>   
Wed 11/22/2017 3:33 PM
To:   Dubrovin, Mikhail;  
   Monarin Uervirojnangkoorn <monarin@stanford.edu>;  
Cc:   O'Grady, Paul Christopher;

Hi Mikhail and Mona,
Shall we try to compare cxic0415 run84? This is a dark run.
Run85 is a light run.
Cheetah bad pixelmap is here.
/reg/d/psdm/cxi/cxic0415/res/cheetah/hdf5/r0101-se_84_new/badpixelmap.h5
Best,
Chuck

Monarin Uervirojnangkoorn <monarin@stanford.edu>
Wed 11/22/2017 6:29 PM
Hi Mikhail,

Here is the mask for bad pixels generated from cctbx.
/reg/d/psdm/cxi/cxid9114/scratch/mona/s1/discovery/mask_cxic0415.txt
The 0’s are the bad pixel. I used r86 as Dark run and r100 as light run.
I hope my understanding of their 64 x 185 x 194 format is correct. I also sent an email to Aaron. I’ll let you know again if the format is not right.
Cheers & Happy Thanksgiving!
Mona

Files

Compare cxic0415 run 84 - dark, 85 - light

• Cheetah bad pixelmap, shape=(8*185,4*388):
  /reg/d/psdm/cxi/cxic0415/res/cheetah/hdf5/r0101-se_84_new/badpixelmap.h5
  extracted with get-mask-cheetah-h5.py
  has ~50K edge and central pixel masked
  extracted by get-mask-cheetah-h5.py

• cctbx, shape=(64,185,194)

  /reg/d/psdm/cxi/cxid9114/scratch/mona/s1/discovery/mask_cxic0415.txt
  includes bound pixels
  extracted by get-mask-cctbx-txt.py

psana pixel_status

• /reg/g/psdm/logs/run_proc/pixel_status/CXI/cxic0415/work/nda-cxic0415-r0084-CxiDs1.0:Cspad.0-e001514-sta.txt
• /reg/g/psdm/logs/run_proc/pixel_status/CXI/cxic0415/work/nda-cxic0415-r0085-CxiDs1.0:Cspad.0-e157676-sta.txt
• mask is generated by get-mask-from-status.py

psana calibman for dark

• work/clb-cxic0415-r0084-peds-sta-CxiDs1.0:Cspad.0.txt
• work/clb-cxic0415-r0084-peds-msk-CxiDs1.0:Cspad.0.txt

psana pixel_status bits

Bits assignment and statistics in calibman and det_ndarr_data_status for cxic0415 run84,85

Mask pixel statistics

• cxic0415-r0084
• get rid of "geometry mask" pixels
• total number of cspad pixels 2296960 = 32*185*388
• masks are presented in cheetah style tables, shape=(8*185, 4*388).

file	zeros	ones	xpm	png
cctbx	3638	2293322	badpixel-cctbx.xpm
cheetah	2187	2294773	badpixel-cheetah.xpm
calibman	3391	2293569	badpixel-dark.xpm
data (r84)	1030	2295930	badpixel-data-r84.xpm
data all(r84)	1329	2295631	badpixel-data-r84-all.xpm
data (r85)	1034	2295926	badpixel-data-r85.xpm
data all(r85)	7049	2289911	badpixel-data-r85-all.xpm

Comparison of masked pixels

Pixel maps are presented in cheetah style tables, shape=(8*185, 4*388).

Color coding of pixels for comparison

0 - black - non-counted pixels at 2x1 edges

1 - purple - good pixel in both masks

2 - red - bad pixel in the 1st file

3 - green-yellow -  bad pixel in the 2nd file

4 - white - bad pixel in both files

Comparison of bad pixel masks in pairs

file1	file2	coinside	only in 1	only in 2	xpm	png
cctbx	cheetah	1243	2395	944	mask-comp-cctbx-cheetah.xpm
cctbx	dark	2152	1486	1239	mask-comp-cctbx-dark.xpm
cheetah	dark	1162	1025	2229	mask-comp-cheetah-dark.png
dark	data	1030	2361	0	mask-comp-dark-data.xpm

References

• Pixel status in data