Thanks Russ, much appreciated.
I've been using gimp but it can only handle 8 bit tiffs and doesn't handle CR2 files. It was a case of lose info in the finished stack file or convert prior to stacking. Gimp works well and is easy to use, I just couldn't find a permanent fix for handling the input of files.
Now it's too easy, stack the raw files and straight into photoshop.
No worries glad I could help out. I wasn't aware GIMP was 8bit format only.
Quote:
Originally Posted by chiaroscuro
Sorry to revive this thread, but I've run out of ideas.
After spending the weekend going through every sub, picking the best by looking at the image and the histogram, I then used the Canon EOS software Digital Photo Professional to tweak the RGB histogram which improved the image quality substantially. I kept a copy of unmodified subs too. All were ISO 1600 at 30 secs, with matching darks and bias frames.
Then I stacked the unmodified subs - poor result seen below, and stacked the modified subs - little improvement in image. Neither could I do much to improve them in Startools or nebulosity.
The same problems seem to be there, the main ones being a green cast, poor colour and contrast overall, which I can't seem to improve with processing.
Both stacking sessions took about 10 hours each, using median sigma clipping on the subs and bias frames as recommended by DSS and youtube tutorials.
I still though get a left sided RGB histogram, and the diagonal dotted line is always diagonal! Never a sigmoid curve as it is in any other images I've seen online of the final stacked file.
Below is the final stacked image of both sets of subs, and one sample image of the modded subs I used, which looks so much better than the stacked images!
I'm lost.
Luke
Hi Luke I wouldn't alter the subs before stacking, stack as is, then import into another program where you can adjust the histogram and curves to tweak the photo (I'm not sure you found a Mac solution yet?). DSS is very hard to control the image after stacking because it is in 32bit format and smashes your computer's memory. If you can upload a stacked version somewhere I will have a go at processing it for you (as I mentioned earlier I wouldn't bother with bias frames with a dslr).
Hi Luke I wouldn't alter the subs before stacking, stack as is, then import into another program where you can adjust the histogram and curves to tweak the photo (I'm not sure you found a Mac solution yet?). DSS is very hard to control the image after stacking because it is in 32bit format and smashes your computer's memory. If you can upload a stacked version somewhere I will have a go at processing it for you (as I mentioned earlier I wouldn't bother with bias frames with a dslr).
Hi Russell,
Thanks for your help: heres a dropbox link to the stacked, unprocessed fits file. I've increased my RAM to 16GB today, so I'll try once more overnight to stack them, but I'd love to see if the problem is the data, the software or me.
Luke, if your images are taking a long time to stack - they shouldn't in any reasonably recent PC - then adjust the Star Detection threshold upwards, i.e. higher number. Keep raising it until you get about 200 stars should be safe enough. You should then find your stacking takes a lot less time...
Thanks for your help: heres a dropbox link to the stacked, unprocessed fits file. I've increased my RAM to 16GB today, so I'll try once more overnight to stack them, but I'd love to see if the problem is the data, the software or me.
I have downloaded your file and used it in PS. I have also read Marc's comments and my image turned out very similar to his. The image is capable of finer detail but there are a few things you need to sort out first.
Marc is correct, the problem is the varying orientations you have stacked and not your processing or options selected in DSS. It appears when you captured the data the camera position was altered approximately 5-6 times. While in theory DSS can stitch these together it is really meant only for minor changes. there is quite a lot of overlap and it appears that DSS isn't capable of working adequately with these parameters or degree of field rotation.
I suggest you only select the subs that are all on the same alignment and work with that in the first instance. It maybe possible to still use all the subs you have selected but you would have to stack each orientation separately to create the 6 or so images of orientation and then stitch together after processing first and then putting together in mosaic software. but I wouldn't worry about that at the moment and instead concentrate on the Great Orion Nebula central stack for the best result.
i also agree with the comment below, DSS shouldn't take that long, on my ancient cheapo laptop it would take 50 mins at the absolute max - i don't know how it did it to be honest.
good luck and let us know how you get on
cheers
russ
ps - as mentioned in my other posts - i also think you should not include your bias frames as they maybe adding to the problem here, in any case it would be good to remove that unknown from the equation.
Luke, if your images are taking a long time to stack - they shouldn't in any reasonably recent PC - then adjust the Star Detection threshold upwards, i.e. higher number. Keep raising it until you get about 200 stars should be safe enough. You should then find your stacking takes a lot less time...
Hi Dunk & Russell,
The thing that seems to change the duration of the stacking is using the median sigma clipping rather than just average or median processing in the settings. And then if you tick the Drizzle option, that adds a huge amount of time to the processing. I've now got 16GB of RAM, and I started a stack last night at 10 pm, with median sigma clipping and drizzle, just to see if it would be better. The stack still has over an hour to run.:eye pop: As suggested, I've not got the bias frames in this stack, so we'll see if that helps.
If I just use the "average" or "median" options in the settings box, its only an hour long too. I just thought the drizzle option would help with the stacking artefact.
I've spent this morning updating DSS to version 3.3.4 through wineskin (a complex process too) and the yahoo forum on DSS says that this version is configured for the Canon 70D, so I assume that earlier versions are not compatible with this camera, and maybe that is the source of my difficulties.
Its a long and winding road, this astrophotography thing - I'm learning a lot, including control of frustration. Thanks again for the tips.
The thing that seems to change the duration of the stacking is using the median sigma clipping rather than just average or median processing in the settings. And then if you tick the Drizzle option, that adds a huge amount of time to the processing. I've now got 16GB of RAM, and I started a stack last night at 10 pm, with median sigma clipping and drizzle, just to see if it would be better. The stack still has over an hour to run.:eye pop: As suggested, I've not got the bias frames in this stack, so we'll see if that helps.
If I just use the "average" or "median" options in the settings box, its only an hour long too. I just thought the drizzle option would help with the stacking artefact.
I've spent this morning updating DSS to version 3.3.4 through wineskin (a complex process too) and the yahoo forum on DSS says that this version is configured for the Canon 70D, so I assume that earlier versions are not compatible with this camera, and maybe that is the source of my difficulties.
Its a long and winding road, this astrophotography thing - I'm learning a lot, including control of frustration. Thanks again for the tips.
Luke
Hi Luke,
Is the new stack just the central region? Or all the field orientations? If its all the field orientations I'm sure you will have the same problem.
Drizzle essentially rescales and upsizes the image. You don't need to do this - it essentially is making the image 2-3x bigger or something (so instead of a 100mb image it will be 200 or 300mb). sometimes it can be ok if you select a part of an image if it is a small target.
Yeah definitely avoid Drizzle for now, it essentially interpolates (in a more clever way) between pixels, but quadruples your image size which, perhaps unsurprisingly, vastly increases the computing requirements to run successfully. Be comfortable walking first before trying to run
I don't notice any appreciable difference in stacking time using sigma clipping over median, btw.
Is the new stack just the central region? Or all the field orientations? If its all the field orientations I'm sure you will have the same problem.
Drizzle essentially rescales and upsizes the image. You don't need to do this - it essentially is making the image 2-3x bigger or something (so instead of a 100mb image it will be 200 or 300mb). sometimes it can be ok if you select a part of an image if it is a small target.
Cheers,
Russell
I didn't crop before stacking - Can I just crop it after stacking, or do you think I should just stack the images in the same orientation only?
Yeah definitely avoid Drizzle for now, it essentially interpolates (in a more clever way) between pixels, but quadruples your image size which, perhaps unsurprisingly, vastly increases the computing requirements to run successfully. Be comfortable walking first before trying to run
I don't notice any appreciable difference in stacking time using sigma clipping over median, btw.
Yep, you're right. I've just been trying everything to try to improve the images, because it seems the basic data is OK, and in doing so, I've complicated matters.
With the sigma clipping, do you set it at 2.00 with 5 iterations? This seems to be the recommended setting, but its the 5 iterations that seems to add a lot of processing time.
I didn't crop before stacking - Can I just crop it after stacking, or do you think I should just stack the images in the same orientation only?
I would do each orientation separately rather than cropping it will go easier on your computer too. you may be able to create a mosaic afterwards. I would choose either the blue or the red rectangles (see attached).
Quote:
Originally Posted by chiaroscuro
Yep, you're right. I've just been trying everything to try to improve the images, because it seems the basic data is OK, and in doing so, I've complicated matters.
With the sigma clipping, do you set it at 2.00 with 5 iterations? This seems to be the recommended setting, but its the 5 iterations that seems to add a lot of processing time.
I use 2.00 and 5 iterations. it will likely run much faster if there a single orientation.
I would do each orientation separately rather than cropping it will go easier on your computer too. you may be able to create a mosaic afterwards. I would choose either the blue or the red rectangles (see attached).
I use 2.00 and 5 iterations. it will likely run much faster if there a single orientation.
good luck
Fantastic, thanks heaps - I'll do that and post the results.
there may be more tutorials available than other programs anyway its another option for you to consider.
also have you tried stacking without the bias files? I was fairly sure DSLR imaging doesn't require bias frames.
cheers
it is NOT free. That download is legally available to only those who have a registered copy of CS2. Adobe provided it because they took down the registration servers that worked with the older CS2 version.
Please do not spread misinformation or encourage software piracy.
it is NOT free. That download is legally available to only those who have a registered copy of CS2. Adobe provided it because they took down the registration servers that worked with the older CS2 version.
Please do not spread misinformation or encourage software piracy.
Dave
Thanks so much for your concern Dave.
Yes, only use the old CS2 version if you did infact purchase it -as the link explains.
After loading the latest version of DSS, stacking only frames oriented the same way, and simplifying the settings, I think the results are significantly better. Below I've attached screen shots of the unprocessed stack, then the with some tweaking of RGB, luminance and saturation in DSS only. But the histograms look more sensible, and the green cast has gone.
I suspect that there were a couple of problems:
- differently oriented frames causing stacking artefact and disrupting the processing. Startools help says that stacking artefact severely disrupts the processing of the images, so maybe that happens in DSS too.
-older DSS version which may not be fully compatible with a Canon 70D. In my reading, I saw that there were plug-ins that need to be installed for the specific camera type.
Anyway, thats not the end of my problems. Startools crashes every time I try to use it, as does Nebulosity. I'm going to try some processing in GIMP.
Thanks heaps to Russ, Marc, Dunk, Raymo and everyone else who made suggestions. It was extremely useful.
Last edited by chiaroscuro; 02-02-2015 at 09:06 AM.
After loading the latest version of DSS, stacking only frames oriented the same way, and simplifying the settings, I think the results are significantly better. Below I've attached screen shots of the unprocessed stack, then the with some tweaking of RGB, luminance and saturation in DSS only. But the histograms look more sensible, and the green cast has gone.
I suspect that there were a couple of problems:
- differently oriented frames causing stacking artefact and disrupting the processing. Startools help says that stacking artefact severely disrupts the processing of the images, so maybe that happens in DSS too.
-older DSS version which may not be fully compatible with a Canon 70D. In my reading, I saw that there were plug-ins that need to be installed for the specific camera type.
Anyway, thats not the end of my problems. Startools crashes every time I try to use it, as does Nebulosity. I'm going to try some processing in GIMP.
Thanks heaps to Russ, Marc, Raymo and everyone else who made suggestions. It was extremely useful.
looking much much better there now Luke - definitely headed in the right direction now.
Nice one mate, glad to hear you're finding a combination/process that makes sense to you, don't be discouraged by the frustration, it just takes time and lots of energy
