1.20
#...
·
 |
|---|
|
Hello I`m having issues with stacking 2 nights of data. Both nights i plate solved so every frame lines up. Quality does vary but for some reason when i stack both nights the end image comes out and the red green and blue bars instead of being solid are segmented and the histogram is only one bar (not the usual 3). I know i should be using WBPP but my laptp just doesn`t have the oomf (power, RAM). I`m using calibration frames. Im just pulling my hair out with why its not working. Does anyone have any ideas? The whole point of plate solving is so you can image one target mulitple nights..... Thanks for reading CS Alan |
7.90
#...
·
|
|---|
| Not sure I follow you on this... What exactly isn't working? You calibrate then what? And, are you using PI? |
|
1.20
#...
·
·
1
like
|
|---|
| When i stack more than 2 nights of data i`m unable to process the end image as there is information missing in the histogram. Usually the end image has 3 peaks (RGB). When i stack this data there is only 1 peak and the streched image has no data. I`m using calibration frames but this doesn`t seem to make any difference. I`ve tried stacking with and without calibration frames but it doesn`t make any difference. |
|
7.90
#...
·
|
|---|
Alan Hancox: If you were to process each night separately would that result in a viable image? You mentioned WBPP which means you have PI but now you use DSS, correct? |
|
1.20
#...
·
|
|---|
|
I do have PI but my laptop isn't powerful enough to use WBPP. It takes an eternity to process anything tbh. I need another one but funds are tight atm. Yes if I stack each night separately it works but not if I combine the subs and only use the best frames. Hth |
|
7.90
#...
·
|
|---|
| Then don't use WBPP which hogs up resources like there is no tomorrow. Do the process one step at the time and albeit slowly it will undoubtedly converge to the final results. I'm used to a very slow, 8GB laptop and it was a pain but eventually it did make it. At any rate you could still use the two night's frames and combine them together in PI using StarAlignement and then add them up in PixelMath. |
0.00
#...
·
|
|---|
| Do the exposure lengths, Gain or ISO settings, any filters used, and image size match up from each night? |
9.58
#...
·
|
|---|
|
Hi Alan, Like many other folks, I've had my share of stacking issues with DSS but, like @andrea tasselli I'm not quite sure to understand what you actually get as output. In particular I'm having trouble picturing what you mean by "end image comes out and the red green and blue bars instead of being solid are segmented". Do you mind uploading a screen shot of your final image (as displayed in DSS) along with the histogram you get. That could really help understand what you get and narrowing down the problem you're having. |
0.00
#...
·
·
1
like
|
|---|
|
Are you using flat and dark frames? If so examine each of those to make sure there isn't anything odd about any of them. I am not sure I was having the same problem you are having but I remember having some odd problem that I could not figure out until examined those and discovered there was a problem with one or more of the frames I was using. |
|
9.58
#...
·
|
|---|
Scott Alber: I would second Scott's suggestion. Some of the most annoying stacking issues I've had were caused by bad flats. But now, if I understand correctly, you did try to stack without calibration frames and got the same problem, correct? |
0.90
#...
·
·
1
like
|
|---|
|
Check your settings in Raw/FITS settings in DSS. I ran into similar issues and figured out the DSS wasn't reading the files correctly. Go the the FITS file tab and click the first box (monchrome 16 bits). Then ensure that you have selected the correct Bayer pattern and bilinear interpolation. That worked for me. |
3.81
#...
·
|
|---|
|
Alan- have you tried to put each night into a new group tab rather than all of the subs in the main group? I have had different outcomes doing that. DSS will register even if each sub isn't perfectly matched to the others, so plate solving won't really impact your stack and registration. And only because I"ve done it by accident, make sure the cal frames are inputted as lights... |
11.33
#...
·
|
|---|
|
I like DSS although I realize that WBPP is probably better. I'm assuming you are stacking OSC data. I would try to put all of the lights into one group and let DSS register everything as opposed to putting the separate night's work into separate tabs. As you mentioned, the images should be very similarly framed since we're using plate solving. You did not say whether or not the restulting stack is a color image or monochrome so, and this might sound overly simplistic but make sure under "options-settings-raw/fits/DPP settings-fits files" that you address the checkbox at the top of that dialogue box appropriately, i.e. specifying if you are processing monochrome images or those created by a color camera. In my paricular situation, I image almost exclusively in monochrome but every now and then, I do a OSC image and I always forget to change this setting. I hope this helps! |
6.76
#...
·
|
|---|
|
It would be helpful to know what kind of data files you have. it sounds as if they are OSC data so need an RGB debayer? From your description it is taking the files as mono, which gives only one bar in the histogram. Go to the left side and choose Raw/FITS DDP settings as below.  It they are raw files, all should have been good. So I assume they are FITS. Go to the FITS Files tab and check the top box. This should ensure you get color, not mono. If it is unchecked FITS files are interpreted as mono, unless the proper keywords is saved in the FITS. Best to make sure and set this manually.  One issue with DSS is that sometimes settings can change, best to check everything before running. And note, DSS version 5 is way better and faster, if you are using v4 I suggest you update to v5. If darks and flats are all the same for both nights, they can go in the same tab, there is no problem with doing this. If you need different darks or flats for each night, put each night in a different tab group. That way each tab can have different darks or flats. Any flats and darks put in one tab will be averaged, DSS doesn't use date information to know what flats/darks go with an image if all are in the same tab. But flats or darks from different nights in different tabs will be used correctly if each nights flats/darks/lights are in the same tab group. Rick |
|
1.20
#...
·
|
|---|
Thanks for all your assistance so far. I am indeed working with a OSC a QHY 168C. Here is a screen shot of my issue.... |
|
6.76
#...
·
|
|---|
|
DSS is getting the Bayer matrix okay, but the problem is it is not showing any signal in the output. Can you show what all you input groups look like. For fits file input it should look something like this, after registration. After stacking the dx, dy, and angle should be filled with values. Can you please post this file summary after stacking would be best, for any tab groups you used. No need to include all the lights, but please make sure you include the offset/dark and flat as well in the file list. Rick  |
|
1.20
#...
·
|
|---|
Hope this helps? I`m working on a laptop so screen resolution is a bit tricky |
|
1.20
#...
·
|
|---|
Flats or some of them.... |
|
9.58
#...
·
·
1
like
|
|---|
|
I think the screenshot showing the Light will be more useful than the flats. and I think that what Rick asked. A couple of comments from your first screenshot. First, Something I see at the very top) is the "(1 frame)". There definitively a problem here, this is reflecting the total number of light frames which have successfully been stacked. So it does look like DSS has "stacked" only one frame. However, the puzzling thing is that the time indicated is 5h57, which doesn't seem to correspond a single frame. I've tried out of curiosity to "stack" one frame and DSS doesn't seem to mind. Second, and somehow related to the previous comment. The histogram does NO contain any data. In the little test I've just mentioned, I've "stacked" a 3s exposure (meaning that it's pretty dark with very little info). STILL, I get a broad histogram in the shadows. Based on these two points, It looks like DSS is "stacking" only a single frame of 5h57 which contains no data. |
1.91
#...
·
|
|---|
|
You might want to try adjusting thre Star Detection Threshold, which can be accessed via Settings - Register Settings - Star Detection Threshold. I use 8% but you'll need to experiment - too low and it won't find enough stars to regisiter and align. Graham |
|
0.00
#...
·
|
|---|
| This will happen if your light frames are moved into a different folder after you've loaded them into DSS or the drive has briefly lost connection. |
|
1.20
#...
·
|
|---|
| What do you mean the drive has briefly lost connection? I don`t understand... |
|
6.76
#...
·
|
|---|
Alan Hancox: Alan, sorry, but it is mostly the lights I am looking for. I want to see if the images are registered properly and that the offsets are computed correctly. If you can, right after stacking, take a look at the file list. You access it after stacking by clicking in the upper left box, anywhere should work. You don't need to click on any of the links. If you click on the link, a window will pop up. Close it, and the file list will be there. You need to do this right after the stack, as computed offset info (dx,dy) is not saved. Rick  |
|
1.20
#...
·
|
|---|
Hope this helps. Strangely i tried to ajust the threshold level to around 500 which gave me 58 frames to stack and it worked. I then tried exactly the same setting as before and it didn`t work. I really don`t understand what the issue is.... |
|
6.76
#...
·
|
|---|
Alan Hancox: You mean 500 for the quality threshold? I tend not to use that, it is really hard to judge what value to use and you may throw out a lot of good frames or include many bad frames. What I do is look at the score list as you have above. Look at the worse score images, zoom in and see if they look bad or not. Also look at the FWHM, if there are a few that high compared to the others uncheck that file. similarly I look at the worst background, if they are much worse than the rest of the files I uncheck them (I actually just delete them if I can see they are bad). I don't trust just setting an arbitrary quality number. Unfortunately, the columns on the right are cut off. Can you narrow these visible columns so all the columns are visible, stars, background, FWHM, etc. I am curious how many stars you detected, 1000 is a very high score suggesting you are getting a lot of stars. Ideally about 50 to 100 stars is the target. More than that will cause problems. So if it takes noise as stars, thenall the offsets shown here can be off. I assume the 342(3), 342(4) are imaged sequential? If so, for the most part they look okay, in that there are not big shifts from consecutive frames. If you are guiding then they all should be nearly identical from the same night. There is a big jump between 3 and 5, and another big jump from 8 to 10. Did you dither? If these are all sequential and guided without dithering than it may be your star threshold is too low, letting noise be used to calculate the offsets. So if some frames are offset from where they should be, then anything in your image tends to get wiped out, as they are out of register. Also, what kind of averaging did you do? If you are not doing a simple average try that first to make sure that is not part of the problem. Rejection methods only work well when the images are similar, particularly bad if background is changing a lot, can cause all sorts of problems. Do you know how to adjust the star detection? Under register click the advanced tab, and move the slider to adjust. Click compute the number of detected stars. This is an estimate based on one of the image files, but it gives you an idea. The selection threshold here is a % of full well. So for example, if my background % is say 0.5%, then the threshold is set above the background, generally I run 2 to 4% depending on how many stars in the area. If you background is 5%, then the detection must be well above that other wise you will get noise counted as stars. Do you know what threshold was set for and again, if I can see the stars column, I would like to see what that is showing.  I have another suggestion, if you can take say 5 light frames from each night, and one dark. one flat and one flat dark and reproduce the problem, then perhaps you could load that to dropbox or google drive, so I can go in and try to see what is wrong. You could load everything, but that is a lot of time and effort. Since the problem occurs when you have both nights, there may be a minimum set of frames that still causes you trouble, that I could troubleshoot with my DSS. Rick |
