Files Not Being Calibrated Correctly

[kerrywaz1]

I have looked at Vicent's tutorial on master frame generation and have tried to calibrate a set of images using master bias, darks and flats. I have selected the master frames and checked the "calibrate" and "optimize" boxes in ImageCalibration, but the images appear to not have been calibrated correctly. The master flat seems to have been applied, but the bias and dark masters don't seem to have been subtracted (read-noise still visible as vertical lines - see attached). I am using a temperature regulated camera (SBIG ST11000) and light box flats.

Suggestions on what I am doing wrong here?

Thanks,
Kerry

[astropixel]

Hi Kerry. No need to check the calibrate boxes, just subtract the bias and integrate. Included in the DSLR workflow is an expanded description written by Vicent originally with CCD imagers in mind.

http://pixinsight.com/forum/index.php?topic=2570.msg19019#msg19019

PS nice image.

[kerrywaz1]

Thanks. I reviewed the thread and calibrated accordingly. Unfortunately, the vertical lines are still there for some reason.

Just to check, I calibrated the images in MaxIm DL and the result was the same there too. The lines have me puzzled. I have never seen them before in my images.

--- Kerry

[oldwexi]

Hi Kerry!
We suffered from the same problem of having vertical lines in the calibrated images of the 11K of my friend.
Checking the bias properly we found out that the bias change dramatically when the temperature changed.
5 Centigrade difference changed the bias already.
Especially when we captured the lights during night and made the darks, flats and bias during the day.
During day we could not get low enough to reach the temperature during night.
Recapturing all flats, bias and darks with the same temperature as the lights finally
let the vertikal lines disappear. PI Calibration was used with the 11K images in all cases.

May be you check the temperature of your lights and the calibration images.
By the way, the vertical  lines with the 11K is nothing new. There is another PI-thread about this
and also in the net its mentioned.

Aloha
Gerald

[astropixel]

A member on another site had the same problem and maintains that reducing the temperature a further 10C fixed the problem.

[sleshin]

Kerry,

The vertical banding artifact is a known problem associated with Kodak Interline Sensors like in the STL 11000 and described by Kodak in this Application Note:

 http://www.kodak.com/ek/US/en/VerticalTimingOptimizatioInterline.htm

As the paper points out the artifact can be eliminated by adjusting the vertical clock edge alignment. Give SBIG a call and talk to Bill Lynch. He's a good guy and does the repairs and can tell you what you need to do.

Steve

[kerrywaz1]

All,

Thanks for the inputs.

I have had the camera for about a year. It is a STL1100M Class 2 and I have never had this problem before now with the vertical lines.

I have a liquid chiller attached to the STL11K and it keeps the temp within about plus or minus 0.5 degree Celsius, even on the hottest night. All images including calibration data are taken at -25C. These particular images were 43 ten-minute exposures through a hydrogen alpha filter.

My darks and bias were taken back in May, so I am reshooting them tonight. Maybe that will make a difference, I don't know. If not, I will give Bill Lynch a call at SBIG and see what they can do.

[astropixel]

My darks and bias were taken back in May, so I am reshooting them tonight. Maybe that will make a difference

I suspect it will. More is better. You'll get a better result.

[kerrywaz1]

I reshot the darks and bias frames overnight and recalibrated the light frames - there was no difference, the lines are still there.

[kerrywaz1]

FYI. I exchanged emails with Alan Holmes at SBIG about this banding issue. Here is his response:

"Kerry - we have fought this on STL-11Ks.  We implemented a clock change fix about 3 or 4 years ago that reduced this effect - is your camera older than that?  [my camera was made in 2008] We also have a tool in CCDOPS to fix this.  Under the utilities menu, look for filters - fix vertical bars.  The filter is best applied to each image before stacking.  You will get better results if you stack fewer, but longer exposed, subframes.  Stacking lots of images, all precisely guided with good alignment, really brings out CCD artifacts at the 1 electron level.  However, our CCDOPS filter should fix this, or at least attenuate it.  We have applied all of Kodak's fixes - we think this is inherent in the CCD. Try the CCDOPS filter, and, if you can, dither back and forth by 16 high resolution pixels.  That will help immensely.  The artifact has a well known frequency of 32 pixels. - Alan"

[pfile]

can't they tell from your serial number or firmware level whether or not your camera has the fix?

[kerrywaz1]

I gave them the serial number on the camera and their response was a somewhat non-commital "we think it has the fix".

I noticed my luminance images do not have the banding, just the narrowband. I also have increased my dithering to six pixels. I tried 16 pixels, but the AO kept losing the guidestar after each exposure.

[kerrywaz1]

Since the vertical banding appears only in y narrowband images, I have found a workaround using PI's Cosmetic Correction script. I rotate the images 90 degrees and apply the Canon debanding and the images are fixed.