Problem with WBPP light masters

I have run into an issue producing Master L,R,G,B with WBPP 2.4.1. Please see the attached image (boosted stretch) which is typical for all four filters. The camera is a QHY600M.

Several attempts, rechecking the settings, produced the same results. I received no error or diagnostic messages.

Master flats, darks and biases all appear OK. WBPP used my 10 sec dark_flats for L,R,B and selected the bias master for the green data. “Optimize Master Darks” was checked.

I used the same darks and biases two weeks ago on another set of data and did not see any problems. There have been no changes to the optics since this last set of images.

Calibrated images appear OK. I am able to generate the master L,R,G,B outside of WBPP using the PI Process ImageIntegration with the WBPP generated lightfiles_c_cc_r.xisf.

If anyone else seen or experienced this behavior I would appreciate any comments you may have.

Frank
 

Attachments

  • masterLight_BIN-2_EXPOSURE-240.00s_FILTER-LUM_Mono.jpg
    masterLight_BIN-2_EXPOSURE-240.00s_FILTER-LUM_Mono.jpg
    636.1 KB · Views: 81
try loading your calibrated and registered subs into ImageIntegration and leaving the normalization mode set to the default (rather than LocalNormalization) and see what happens.

rob
 
Thanks for your reply.
If I understand you, this is the way I originally ran Imageintegration outside of WBPP. The results were OK.
Settings were: Average, Additive, SNR Estimate
Pixel rejection(1): ESD, Scale plus zero offsets
Pixel (2): Default settings

I just ran with the defaults only. The resulting master looked about the same as the above.

Running with the defaults and setting the two Normalization settings to "No Normalization" did result in in a cleaner looking image. The "amp glow"
in the two right hand corners was eliminated and the left hand corners looked much better. The INF disappeared as well but it was at low signal and I did not think that I could use anyway.
Frank
 
well there are certain situations that seem to confuse the new LN (which i think juan is fixing in the upcoming release) and it seemed to me a little like you might be suffering from that. i believe WBPP uses LN by default, so i was suggesting trying a less fancy form of normalization to see if that might have been the problem. additionally if the subs look OK to the eye under blink, then whatever is changing them must be 'behind the scenes', like LN (LN creates 'sidecar' files containing the normalization information by default, and does not change the actual image.)

however it doesn't seem to have been the problem if Additive normalization didn't change anything.

are you sure all the subs look OK? do all of them have the huge black border on the left?

rob
 
This has nothing to do with the problem we have recently solved, which has been commented in this thread. Black borders cannot be a problem (if they are black or have very low values). As always, we need the data set to know what happens, otherwise we cannot help.
 
My subs all look good. I have 25 for each filter.

I ran WBPP again with "Local Normalization" unchecked for the light frames. This cleaned up the large black border areas on the integrated image.
Ran ImageIntegration again with: "Additive with Scaling", and "No Normalization" in Pixel Rejection (1). The results look like the new WBPP image.

As mentioned above, I ran ImageIntegration with the default settings except for "No Normalization" selected at both options. The resulting integrated image looks cleaner (less noisy) but appears to my eye have lost some signal in the galaxy arms and in the diffraction spikes.

I think that there is a good solution here. I just need to play with the data and clean up some satellite trails.
Thanks for all of your help,
Frank
 
Frank, just curious but have you had the camera driver remove the overscan area before saving the file? I had a similar problem until I did that and removing overscan solved it. With the overscan still in the image, the images came out ok if I skipped LN. Since you had success previously, this is probably not the problem, but I wanted to mention it.
Larry
 
Hi Larry,
I did notice that the problem areas on this set of images were adjacent to the overscan areas. The camera is relatively new and I am still working on the settings that will be best for my imaging situation. I think I will follow your example and at least look into the pros and cons of the overscan. Do you see any issues other than that this would be another state variable to keep track of?
Frank
 
I ran WBPP again with "Local Normalization" unchecked for the light frames.

Local normalization is a must if you want to get the most out of your data. I strongly recommend that you solve this problem on your raw frames, since it cannot happen if your data is correctly calibrated.
 
I believe we have a solution to my Master lights calibration problem. Larry suggested that I set the ignore overscan in the camera driver. I had planned to do but it was too late for the present set of data. It turns out that you can do essentially the same thing in WBPP after the fact. Worked like a charm.
Thanks to All,
Frank
 
I realize I am late to this discussion but I have had the exact same issue when trying to use LN in WBPP with quality subs and calibration images from the QHY 600C. If I use LN there is an unusable area of data along the bottom and left side of the image, similar to the image above but in color. If I don't use LN, the integrated image looks fine. I have tried to disable the overscan area on the QHY 600 but the check box does not seem to activate (that is, in SGP, in my camera settings, there is a check box for disabling the overscan area but it doesn't register as being activated and images still have the overscan area). I read with interest the comment above that I can eliminate the overscan area after the fact in WBPP. Could someone educate me on how to do this? I would like to incorporate LN in my processing but with this camera it's hopeless. I also use a ZWO 2600 OSC and have no issues with using LN. Please Advise.
Thank you,
Dan Dall'Olmo
 
Hi Dan,
I ended using the "Remove Overscan" found on the QHYCCD Driver Setup utility. This assured that all of my calibration frames were the same size as my Lights and that I don't need to think about this when I get to processing. I am using the mono camera.

I also tried the procedure suggested by thutchison and found that it worked well. As I recall, I had to adjust for the fact that I was binning 2X2.
ImageCalibration and Overscan area procedure
Frank
 
Hi Dan,

if you want to use the available frames that include the Overscan area, you can use Overscan calibration in WBPP with the original frames.

If you don't want to use Overscan calibration for future projects, you should enable cropping in the camera driver. SGP will use the ASCOM camera driver, and you can configure this option in SGP. Of course, light and calibration frames must be captured with the same setting in the camera driver regarding Overscan.

I summarized my knowledge of Overscan and Overscan correction in my guide, https://pixinsight.com/forum/index....processing-of-raw-data-with-pixinsight.11547/ , section 7.1.

Bernd
 
Back
Top