Thanks for all of the replies. There is quite a lot to go through here so I'll expand on my initial post above. Have 10.5 hours on M109, shot over three nights. I initially stacked all of the data together, using groups on WBPP but noticed the calibration problems in the left image above. I stacked the data again separately by session to investigate the problem and discovered that the problems are coming from the data from night one.
Kevin Boucher:
Hi Craig,
I'm in an airport waiting to come home after my solar eclipse trip to Austin but I agree with one of the other replies that it looks like the image on the left did not calibrate at all. So my question would be did you only create calibration frames for one of the two nights and if you did did you set up your grouping keywords so that it would know to find the calibration frames and apply them for both nights?
As I don't have Pixinsight in front of me I can't give you any detailed instructions to check this out... but that's what I would pursue at this time. You should see this in your WBPP log.
Kevin
I used a master dark and master flat dark from my library but did take flats after each session and I did use grouping in WBPP to make sure the correct flats were applied to each night's data.
Dan Kokinda:
Craig Dixon:
Okay, thanks. So it sounds like there is nothing to worry about but I need to figure out why the flats for the first image didn’t work properly then
I think you need to start there. You can't make a comparison at all unless you work out any known discrepancies, which you know of at least this one.
Also, is it possible there were slight differences in the atmosphere, such as thin high clouds on the first night that may have diffused things? Parasitic lights as someone else suggested? But, as Daniele mentioned, because the images are different, STF will behave differently, possibly magnifying and discrepancies.
The seeing wasn't great for each of the nights I imaged this target but I open all light frames in ASI FITS Viewer and manually delete any subs that have clouds or poor stars. This way, every sub that goes into WBPP has been manually visually inspected. Some frames will naturally have better signal than others due to passing loud and seeing but the worst frames are manually removed and my understanding is that WBPP then assigns weights to each sub to account for this.
Adam Block:
Ok... so I have an educated guess that assuming facts not yet in evidence.
Is the masterflat that you used on BOTH evenings...was it constructed prior to the image on the right or from the session of the image on the right (which is better calibrated)?
1. If yes, then your optics are changing... you are physically changing the conditions during acquisition so the flats will need to be taken matching each session. THIS IS TYPICAL unless you really touch nothing with a permanent setup.
2. If no, and you are using an OLD master flat and it correctly calibrated NEW data but not the older data... this is very hard to explain unless you really made adjustments in some non-expected way.
-Adam
I took flat 20 frames for each session using an LED flat panel and the auto flats feature in ASIAir. My setup is in an observatory so after night one, I removed the dew shield, put the LED panel on top of the scope, took my flats, replaced the dew shield and then moved the scope to it's parked position and closed the roof. The next night, I opened the roof, turned the power on and ran the same plan from ASIAir again. None of the imaging train was changed so it shouldn't be necessary to do new flats between sessions but I do this as a matter of caution. The only change in optics would be the very slight change in focus between sessions as well as throughout the imaging sessions (Using a ZWO EAF set to refocus every hour or ever 2 degrees Celsius change in temperature). My understanding is that very slight changes in focus shouldn't cause problems with flats.
I'm not the big expert in WBPP, but did you use local normalization, and if so - did you check the local normalization master?
Local normalisation was used in WBPP but I'm not sure how to check the master file. The log doesn't show any errors here. When I use WBPP, I load up all the lights, flats, master dark and master flat dark, then group the images by night using the date. I select the "Maximum quality with no compromised" preset, check the diagnostics button and then run the stack.
Dustin Gazz:
It looks like 1 was taken before the meridian and the other after. If you did not rotate your camera or open your image train, use the same flats for both.
Both imaging sessions were started at roughly the same time and included a meridian flip about 2 hours in. A lot of pre-flip subs might have been deleted from that stack due to cloud so that might be why WBPP has outputted the file 180 degrees flipped but WBPP handles the orientation of pre and post-flipped images well. Either way, the rotation of the camera relative to the scope isn't changed so the flats would still be in the same orientation in relation to any dust/vignetting.
Mike Sample:
Since your sessions are under the same conditions, I would not stack the two separately. The statistics work better when you feed it more frames. WBPP's weighting would not be as effective when stacking sessions separately.
Personally, I never rely on STF other than judging what workflow is appropriate. In your case, I would run DynamicCrop, DBE, GradXpert, ImageSolver, SPCC, BlurXterminator, and NoiseXterminator. After that, I start stretching with HistogramTransfer stopping before saturating very many stars. I would then separate the stars from the target and further stretch the target with GeneralizedHyperbolicStretch. I use CurveTransformation to further stretch the stars.
I would not be concern with either result you posted, but you should be stacking all the lights in one WBPP pass.
I've only stacked the two nights separately to troubleshoot and to find find out which data set is causing the problems. The intention is to figure this issue out and stack all three nights together. I agrees with your workflow but at the moment, I'm just relying on an STF to check the data for issues before I move on. applying an STF has allowed me to see some calibration problems, which need fixing first.
