Bias frames have an ADU level greater than the light frames

[LMLeitch]

Hi,

I created master bias frames using PI ImageIntegration process using the following settings: (there were 44 raw bias frames)

Image Integration section:
•   Combination: Average
•   Normalization: No normalization
•   Weights: Don't care (all weights=1)
•   Uncheck Evaluate Noise

Pixel Rejection sections
•   Averaged Sigma Clipping for < 10 frames
•   Winsorized Sigma Clipping for 10 to 20 frames
•   Linear Fit Clipping for > 20 frames

Here are 'ADU' values obtained from PI compared to values obtained with CCD-Stack

bias 1x1 raw,    CCD-Stack: mean = 1097.97
bias 1x1 master, CCD-Stack: mean = 1180.64

bias 1x1 raw,    Pixinsight, left click and move the mouse, K value = 0.0164-0.0167
bias 1x1 master, Pixinsight, left click and move the mouse, K value = 0.0165

light, 1x1 raw,   Pixinsight, left click and move the mouse, K value = 0.0018-0.0022
light, 1x1 raw,   CCD-Stack,  mean = 1262.41

Using CCD-Stack to calibrate:
light - bias = 1262.41 - 1180.64 = 81.77

Using Pixinsight to calibrate:
light - bias = 0.0020 - 0.0165 = NEGATIVE 0.0145


The result was that the lum image calibrated with PI was a very black background with some very bright white stars.

I am guessing that the settings used to integrate the bias frame are not correct.

Does anybody have suggestions?

Thank you.

Larry Leitch

[Juan Conejero]

From the values that you have specified, it seems that your light frames have not been loaded correctly in PixInsight. This looks like the typical FITS interoperability issue. Which data formats are you using? 32-bit floating point?

Also based on the numbers in your post, your data set may include or require pedestal values that you might not be applying for calibration (typically, PEDESTAL FITS keywords with values about 100 - 200 DN).

To say something more precise, I'd need more information, or taking a look at some original files.

left click and move the mouse

Instead of direct pixel readouts you can use the Statistics tool to get mean pixel sample values, if this is what you want.

[jkmorse]

Larry,

As Juan indicates, you may be running into FITs incompatibility issues.  If that is the issue, here is a post from a while ago where I explained a number of possible solutions (at least they helped me overcome the same issues):

 http://pixinsight.com/forum/index.php?topic=8021.msg52728#msg52728

Hope it helps,

Jim

[LMLeitch]

I have emailed a DropBox link to Jim. I do not know of a way to email the link to the forum.

The link points to some bias and lum files that I am trying to calibrate. I have tried all of the combinations of "Signed-Is-Physical" and "lower-range 0 upper-range 65535" for ImageIntegration and ImageCalibration that I can think of.

My goal was to calibrate the lums with only a master bias. Trying to calibrate with bias, darks, and flats seemed like too big a step. ;-)

If anybody else would like a DropBox link, please let me know.

Thanks to all!!

Larry Leitch