PixInsight Forum (historical)
Software Development => New Scripts and Modules => Topic started by: Peter S on 2018 March 02 03:23:30
-
Hi,
following results for my sensor - I dont understand results concerning estimated thermal noise / skyfog noise.
results:
ISO speed: 1600
Image Size: 11608 x 8708
Bias frames exposure time: 1/4000 sec
Flat frames exposure time: 1/2000 sec
Dark frames exposure time: 30 sec
Light frames exposure time: 30 sec
Channel order: []
Gain per channel(e/ADU): [ 0.047, 0.047, 0.047, 0.047 ]
ISO for unit gain: [ 75, 75, 75, 75 ]
Read Noise(e): [ 1.24, 1.25, 1.25, 1.26 ]
Dark Current (e/pixel/sec) [ -0.0308, -0.0309, -0.031, -0.031 ]
Dark Current (e/pixel/sec) [ -0.0307, -0.0308, -0.0309, -0.0314 ]
Dark Current (e/pixel/sec) [ -0.03, -0.0298, -0.0301, -0.0304 ]
Dark Current (e/pixel/sec) [ -0.0297, -0.03, -0.0301, -0.0304 ]
Dark Current (e/pixel/sec) [ -0.0294, -0.0297, -0.0298, -0.0302 ]
Noise estimates for 30sec exposure
Read Noise(e): [ 1.24, 1.25, 1.25, 1.26 ]
Thermal Noise(e): [ -92233720368547760, -92233720368547760, -92233720368547760, -92233720368547760 ]
Skyfog Noise(e): [ -92233720368547760, -92233720368547760, -92233720368547760, -92233720368547760 ]
Mean values in ADU
Bias Frame 1: [1025.2, 1025.1, 1025, 1025.1 ]
Bias Frame 2: [1025.2, 1025, 1025, 1025 ]
Flat Frame 1: [10217.2, 13880.2, 13899.5, 4960.4 ]
Flat Frame 2: [10330.2, 14040.5, 14059.9, 5010.3 ]
Dark Frame 1: [ 1025, 1025.1, 1025, 1024.8 ]
Dark Frame 6: [ 1026.4, 1026.5, 1026.4, 1026.5 ]
Light Frame 1: [1197.3, 1328.1, 1328.4, 1144.1 ]
Light Frame 2: [1196.4, 1325.7, 1326.5, 1143.8 ]
DSLR Sensor Parameters v0.0.6 Completed.
Info:
Sensor has an offset / Black level: 1024
Any ideas?
Peter
-
I was responsible for that script. There's actually a version 7 which you can find here and fixes a bug in the skyfog calculation: https://pixinsight.com/forum/index.php?topic=10365.15
However the main problem is that it is calculating a negative dark current. This will be because the standard deviation of the difference of a pair of dark frames is lower than the standard deviation of the difference of a pair of 2 bias frames.
Why this is the case, I don't know. From the 100Mpixel sensor size I'm guessing this might be a Phase One Trichromatic camera? I wonder if the file format is being interpreted correctly or if some "cooking" of the raw files is taking place, for instance noise reduction on longer exposures?
Mark
-
Mark,
I´ll try the new version!
...it´s the IQ3 100 (not trichromatic), auto darks were surpressed, no auto noise suppression.
but the sensor temperature rises between and during exposures.
Peter
-
V7 does not change much, but the results (dark current) depend on which dark or light frames of a series I take, so it´s probably the rising temperature (?).
Bias frames exposure time: 1/4000 sec
Flat frames exposure time: 1/2000 sec
Dark frames exposure time: 30 sec
Light frames exposure time: 30 sec
Channel order: []
Gain per channel(e/ADU): [ 0.047, 0.047, 0.047, 0.047 ]
ISO for unit gain: [ 75, 75, 75, 74 ]
Read Noise(e): [ 1.25, 1.25, 1.25, 1.25 ]
Dark Current (e/pixel/sec) [ 0.1655, 0.1809, 0.1685, 0.1786 ]
Noise estimates for 30sec exposure
Read Noise(e): [ 1.25, 1.25, 1.25, 1.25 ]
Thermal Noise(e): [ 2.23, 2.33, 2.25, 2.31 ]
Skyfog Noise(e): [ 2.69, 3.61, 3.62, 2.12 ]
Mean values in ADU
Bias Frame 1: [1025.2, 1025, 1025.1, 1025.2 ]
Bias Frame 2: [1025.3, 1025.2, 1025.1, 1025.1 ]
Flat Frame 1: [10345.7, 14075.4, 14095, 5026 ]
Flat Frame 2: [10466.9, 14247.7, 14267.4, 5079.5 ]
Dark Frame 1: [ 1061.6, 1065.7, 1062.1, 1067.3 ]
Dark Frame 2: [ 1064.5, 1068.8, 1065, 1070.5 ]
Light Frame 1: [1216.5, 1343.6, 1342, 1164 ]
Light Frame 2: [1216.4, 1343.8, 1342.3, 1164 ]
DSLR Sensor Parameters v0.0.7 Completed.
Peter