Difference between revisions of "AXIOM Beta TestsTODO"

From apertus wiki
Jump to: navigation, search
Line 1: Line 1:
#raw12 snapshot of colorchecker passport with 2010 offset (regs 87/88) gain x1 and x2 - exposure so nothing clips
#raw12 snapshot of colorchecker passport with 2010 offset (regs 87/88) gain x1 and x2 - exposure so nothing clips
#repeat Hutch and IT8 charts - same framing for both Nikon and Apertus, please! (surroundings do matter, and so does the camera orientation)
#repeat Hutch and IT8 charts - same framing for both Nikon and Apertus, please! (surroundings do matter, and so does the camera orientation)
#grey out of focus wall, gain x1,x2 with 100 frames of different exposure times (listed below), then run a script and upload the results (no need to upload the raw files)
#grey out of focus wall, gain x1,x2 with 100 frames of different exposure settings (listed below), then run a script and upload the results (no need to upload the raw files)
#one small video of the out of focus wall (5 frames or so), with exposure on the dark side (around 7 stops below white), gain x2 (to check FPN correlation between frames)
#one small video of the out of focus wall (5 frames or so), with exposure on the dark side (around 7 stops below white), gain x2 (to check FPN correlation between frames)
#raw12 snapshots of real world scenes (outdoors, people, etc.) with high & low gain, long & short exposure times  
#raw12 snapshots of real world scenes (outdoors, people, etc.) with high & low gain, long & short exposure times  
Line 14: Line 14:
* Please use a static scene (same background in both Nikon and Apertus), and try to take the picture from the same location (they should have the same framing *after* cropping the Nikon image to APS-C).
* Please use a static scene (same background in both Nikon and Apertus), and try to take the picture from the same location (they should have the same framing *after* cropping the Nikon image to APS-C).


Exposure times for the gray frames:
Exposure settings for the gray frames:


milliseconds: 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.10, 0.11, 0.12, 0.13, 0.14, 0.15, 0.16, 0.17, 0.19, 0.21, 0.22, 0.23, 0.25,
              0.26, 0.28, 0.30, 0.32, 0.34, 0.37, 0.39, 0.42, 0.45, 0.48, 0.53, 0.57, 0.61, 0.65, 0.70, 0.75, 0.80, 0.87, 0.94,
              1.01, 1.08, 1.17, 1.26, 1.35, 1.45, 1.57, 1.69, 1.83, 1.97, 2.12, 2.28, 2.46, 2.65, 2.86, 3.09, 3.32, 3.58, 3.86,
              4.16, 4.49, 4.84, 5.22, 5.62, 6.07, 6.54, 7.06, 7.62, 8.21, 8.85, 9.55, 10.30, 11.11, 11.98, 12.92, 13.94, 15.04,
              16.22, 17.49, 18.87, 20.36, 21.96, 23.68, 25.55, 27.56, 29.73, 32.07, 34.59, 37.32, 40.26, 43.43, 46.85, 50.54,
              54.52, 58.81, 63.45, 68.44, 73.84, 79.65, 85.92, 92.69, 100.00.
  exp_time reg: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 17, 18, 19, 21, 22, 24, 26, 28, 30, 33, 35, 38, 41, 44, 48, 52,
  exp_time reg: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 17, 18, 19, 21, 22, 24, 26, 28, 30, 33, 35, 38, 41, 44, 48, 52,
               56, 60, 65, 70, 75, 81, 88, 95, 102, 110, 119, 128, 138, 149, 161, 174, 188, 202, 218, 235, 254, 274, 296, 319,
               56, 60, 65, 70, 75, 81, 88, 95, 102, 110, 119, 128, 138, 149, 161, 174, 188, 202, 218, 235, 254, 274, 296, 319,
Line 27: Line 21:
               1692, 1826, 1970, 2125, 2292, 2473, 2668, 2878, 3105, 3349, 3613, 3898, 4205, 4537, 4894, 5280, 5696, 6145, 6629,
               1692, 1826, 1970, 2125, 2292, 2473, 2668, 2878, 3105, 3349, 3613, 3898, 4205, 4537, 4894, 5280, 5696, 6145, 6629,
               7152, 7715, 8323, 8979, 9687.
               7152, 7715, 8323, 8979, 9687.
To compute the exposure times, please write down the LVDS clock value used during the test.


Script to be run on those 100 frames:
Script to be run on those 100 frames:
  raw2dng *.raw12
  raw2dng *.raw12
  python2 [http://dl.dropboxusercontent.com/u/4124919/bleeding-edge/apertus/raw12_stats.py raw12_stats.py]
  python2 [http://dl.dropboxusercontent.com/u/4124919/bleeding-edge/apertus/raw12_stats.py raw12_stats.py]

Revision as of 16:19, 18 December 2015

  1. raw12 snapshot of colorchecker passport with 2010 offset (regs 87/88) gain x1 and x2 - exposure so nothing clips
  2. repeat Hutch and IT8 charts - same framing for both Nikon and Apertus, please! (surroundings do matter, and so does the camera orientation)
  3. grey out of focus wall, gain x1,x2 with 100 frames of different exposure settings (listed below), then run a script and upload the results (no need to upload the raw files)
  4. one small video of the out of focus wall (5 frames or so), with exposure on the dark side (around 7 stops below white), gain x2 (to check FPN correlation between frames)
  5. raw12 snapshots of real world scenes (outdoors, people, etc.) with high & low gain, long & short exposure times
  6. colorchecker passport zoomed in on the bottom half and defocused a bit (just the boxes area) - for FPN

For all samples:

  • Please set the black offset registers (87,88) to 2010. 1910 is too low (clipped shadows = data lost). To get a normal picture, use raw2dng --black=100.
  • Please save the register metadata block (on past samples, many files were mislabeled).

For the color charts:

  • Please use a static scene (same background in both Nikon and Apertus), and try to take the picture from the same location (they should have the same framing *after* cropping the Nikon image to APS-C).

Exposure settings for the gray frames:

exp_time reg: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 17, 18, 19, 21, 22, 24, 26, 28, 30, 33, 35, 38, 41, 44, 48, 52,
              56, 60, 65, 70, 75, 81, 88, 95, 102, 110, 119, 128, 138, 149, 161, 174, 188, 202, 218, 235, 254, 274, 296, 319,
              344, 371, 400, 432, 466, 503, 542, 585, 631, 681, 735, 793, 855, 922, 995, 1074, 1158, 1249, 1348, 1454, 1569,
              1692, 1826, 1970, 2125, 2292, 2473, 2668, 2878, 3105, 3349, 3613, 3898, 4205, 4537, 4894, 5280, 5696, 6145, 6629,
              7152, 7715, 8323, 8979, 9687.

To compute the exposure times, please write down the LVDS clock value used during the test.

Script to be run on those 100 frames:

raw2dng *.raw12
python2 raw12_stats.py