Deep Sky Stacker removing too much color from my Orion stack [Deep Sky] Processing techniques · Brent Newton · ... · 25 · 1275 · 0

brent1123 2.41
...
Hello everyone, I am having a problem with DSS.

Single 120" Image: https://www.dropbox.com/s/4z58cbukm9c8ag8/IMG_9866.CR2?dl=0

Stacked and Calibrated (dozens of Darks, Flats, and Bias) using 3 hours of exposure (Cropped and Binned in Star Tools): https://www.dropbox.com/s/qe90q737h72kjx9/2a%20-%20NEB%20Rot%200.6%2C%20Bin%2050%25%2C%20Crop%20for%20CS6.tiff?dl=0

As you can see there is a lot of lost color in these images. I have seen similar problems through googling, I am not the only one who has dealt with this, but the only advice I can find is that stretching the data should bring the color back - this has not been my experience. What I end up getting is a lot of background dust around the center nebula, but none of it has color. At best I am stuck with a moderately saturated M42 surrounded by a grey~ish collection of dust, which looks terrible to me.

Are there any good alternatives to DSS I can use for stacking and calibrating my frames? Or perhaps ways to increase color saturation without making the entire image look like a psychedelic mess?
Like
Magellen 8.51
...
Hi,

I do not use DSS a lot, but there are alternatives: Of course PixInsight http://pixinsight.com/, a great program.
Then there is Theli, a wonderful and free stacking software, used by professionals but it depends on Linux.
Or, you could try Fitswork, which is free as well and then there is Regim, which comes with an excellent colour calibration.
I use PI in most cases and a prefer them all over DSS
Like
nekitmm 2.41
...
I used DSS with my DSLR and didn't have this issue. What do you use to stretch your final image? I do not recommend stretching it in DSS, it's terrible at this, better to use something like PS to do it.
Edited ...
Like
brent1123 2.41
...
Nikita Misiura:
I used DSS with my DSLR and didn't have this issue. What do you use to stretch your final image? I do not recommend stretching it in DSS, it's terrible at this, better to use something like PS to do it.

I've been using either Photoshop CS6 or Star Tools to stretch the data, I never touch it in DSS except to glimpse what it could look like, but I always reset the changes before saving, and never apply them
Like
nekitmm 2.41
...
Hmmm... This is very strange. But I am sure that color is there after stacking. I can only suggest you to try to find mistakes in your processing after stacking. What filetype do you use to save stacked image? It should be something like 32bit tiff image.
Like
nekitmm 2.41
...
And what is you camera? Do you use RAW as an nput?
Like
lock042 5.25
...
On OS-X and/or Linux you can give a try to Siril (free and opensource)
Edited ...
Like
ecdon 0.00
...
I am using an unmodded Canon T3i in DSS. I always align the color channels (Align B and G with R) and increase the saturation to 20%
Like
brent1123 2.41
...
I'm not to the point of using separate RGB channels yet, I'm only using color, though I separate it into an artificial luminence version for sharpening.

I did find a solution though, DSS makes an auto save file for every stack,which for some reason contains far more color than any manually saved file. So I just restocked it and grabbed the auto save and renamed it
Like
Judson_Graham 0.00
...
id recommend using adobe camera raw to convert the files to tiffs before you stack in dss. in that conversion you can also apply a lens profile correction which will remove most light falloff and help you with the processing in the end.
then save all the files as tiffs and stack in dss. once it is sacked do not use the autosave tiff.what you want to do is hit save image on the left side menu of dss then check the "save image with adjustments embedded" box. that way dss saves you a median stacked (or whatever stacking method you used) image that looks just like the inputs just with less noise.
this article by user Roger Clark will explain it http://clarkvision.com/articles/astrophotography.image.processing/
Like
brent1123 2.41
...
Read through it and a few of his other articles, could you slightly simplify it for me? What is the advantage of using Tiff over RAW?
Like
Hondo 0.00
...
Judson Graham:
this article by user Roger Clark will explain it http://clarkvision.com/articles/astrophotography.image.processing/

I would have a read of the link below before you put any faith in that website.

http://forum.startools.org/viewtopic.php?f=4&t=912
Like
crazygame12345 0.00
...
Try Bayer plane Extraction (in MDL)。
this Method use R,1/2*(Gr+Gb),B to compose  RAW  color channels with a half resolution .That can make color saturated and high SNR.
Like
bobzeq25 0.00
...
I'm not to the point of using separate RGB channels yet, I'm only using color, though I separate it into an artificial luminence version for sharpening.I did find a solution though, DSS makes an auto save file for every stack,which for some reason contains far more color than any manually saved file. So I just restocked it and grabbed the auto save and renamed it

When you manually save the file do you check the box labeled something like "embed but do not apply settings"?  It's mandatory.  DSS settings are useful for seeing what you've got, terrible for processing.  When you open the stack in your processing program, it should look dismayingly dark.  That's correct.

Stretching always reduces color.  You have to put it back in processing.  Just part of the procedure.
Like
LookoutLane 0.00
...
I have basically experienced the same problem with DSS and am relieved to find out it isn't just me.   I am newbie to astrophotography so assume I am missing something.     My first attempt was last week when I took many exposures of Orion with my Canon T2i at 6400 at 15s, 30s, 60s, 90s and 120s through a 6 in Mak Cas.  After I stacked the best exposures using DSS I experienced the same thing - namely a loss of color.     Had a little better luck using JPGs instead of Raw photos, but still the stacked photos initially looked like they were shot in monochrome.   If I increased saturation enough and tweaked RGB channels I could get color to appear, but none of my stacked pictures looked as good as my best single frames did after a little enhancement.     Obviously something isn't right.
Edited ...
Like
bobzeq25 0.00
...
It's unlikely that DSS is doing this, I used it for two years on a variety of targets with a variety of equipment and a variety of exposures,
never had the problem.

When you stretch data you inevitably lose color, which you have to restore.  A 20% increase in saturation is nothing, I used between
500 and 1000% in StarTools.

@Scott.  Do you ensure DSS is not stretching the image, as Brent does?  That's absolutely required, DSS does a terrible job of stretching.
The way I do it is to stack.  DSS shows you a stretched image so you can see what you've got.  There are sliders to change the stretch,
ignore them.  I click save image to file on the left.  In the save box I'd be sure "settings not applied but embedded" was checked.

When you open that file in your processing program it will look ridiculously dark.  Like "where's my image?".  That means you
did it correctly.  You now stretch it in your processing program.

Pretty much every experienced imager does it that way.  Because:

DSS does a lousy stretch.

As you develop in processing, you will want to perform certain operations on the linear (unstretched) stack.  A prime example is
gradient reduction, which really only works on linear data.  It reduces light pollution by looking for  a characteristic pattern.
Stretching destroys the pattern.
Edited ...
Like
Erlend_Langsrud 0.90
...
After using this software for some time, I am now convinced that this is due to a profound error in DSS.

If you stack, say 20-40 subs, it is possible to bring the color back by increasing the saturation.

This week, I stacked 188 raw subs from a dslr. (Yes, i use raw, format and yes, there is color in each sub.). The color was almost completely gone. I could increase the saturation and bring out some kind of color, but the color data was basically washed out beyond rescue. It would be more easy to understand is the image was a perfect grayscale, but there is some trace of color left. The difference between the color channels is in the order of 1% at a given point in the image.

Sometimes I suspect that the debayering is appiled after stacking, or something stupid like that.

I will try to use the image as a luminance channel and stack the .jpgs to produce a color layer.
Like
silkpericles 0.00
...
For a long time used DSS, I get the impression that you could be applying some unwanted adjustment to the final image, I can not access the link if I saw could tell you with more certainty which may be the source of the problem, try to change the stacking settings , such as the background calibration for each channel can help you get the desired result, but as in all the best is always the activates and error. As advice I will tell you that you should never save the image already stacked that post processes in DSS, I always got better results post-processing the AutoSave raw in another program, I use Pixinsight, but to see it with more color and not so pale you should upload the saturation in DSS up to 10-20.

Best regards.

Alberto.
Like
Jir11 0.00
...
Hi Guys,

Quick question, I use DSS and really have no issues with it. Just wanted to know that after stacking and such I normally retrieve the autosave file and then use that file in PS to do my stretching. I've noticed some reference to "saving the file" manually. Is this the same as the autofile that DSS generates? Is there a difference or advantage between the two methods? Thx.

Jirair
Like
PEJU 0.00
...
Try APP (Astro Pixel Processor)  very competent all in one
Not so different from dss to start with.  Gives hi-qual integrations
Like
PEJU 0.00
...
Try APP (Astro Pixel Processor)  very competent all in one
Not so different from dss to start with.  Gives hi-qual integrations
Like
Hondo 0.00
...
Jirair Afarian:
I've noticed some reference to "saving the file" manually. Is this the same as the autofile that DSS generates? Is there a difference or advantage between the two methods?

I use the autosave file, for me a fits file, and use Star Tools for processing.  I never had a problem with DSS stacking and registering my files and I shoot in mono.  I read a long time ago that you should only use the autosave file and that is what I have been doing for several years now.  I do not know if there is any difference in using a manually saved file but the creator of Star Tools also recommends using the autosave file.  Bottom line is to use what ever works for you.  Post processing is an art and everyone has an unique way of producing images they are happy with.

Scott
Like
bobzeq25 0.00
...
The important thing is to stop DSS from processing the stack.  There's a buried setting somewhere that does that with autosave.  But the surest way to do it is:

Stack.  Ignore the sliders,  Click save picture to file on the left.  When the save box appears, be sure
"settings embedded but not applied" is checked.

Open the file you saved in your processing program.  It will look very dark.  That is correct.
Edited ...
Like
Jir11 0.00
...
Scott:
Jirair Afarian:
I've noticed some reference to "saving the file" manually. Is this the same as the autofile that DSS generates? Is there a difference or advantage between the two methods?
I use the autosave file, for me a fits file, and use Star Tools for processing.  I never had a problem with DSS stacking and registering my files and I shoot in mono.  I read a long time ago that you should only use the autosave file and that is what I have been doing for several years now.  I do not know if there is any difference in using a manually saved file but the creator of Star Tools also recommends using the autosave file.  Bottom line is to use what ever works for you.  Post processing is an art and everyone has an unique way of producing images they are happy with.

Scott

Exactly what I've been doing I use the autosave 32bit tiff file and process it in photoshop. I convert the file to a 16 bit exposure and gamma file then proceed from there. But as you see from the above poster some people believe that DSS processes the file before autosaving. I have not seen or read any evidence that it actually does.

I do notice that after converting, most of my files show a histogram with the RGB channels slightly skewed and I have to bring them together in "Levels" before stretching the image using "Curves".  The red is usually at the far left of the histogram then further to the right the green then slightly further more the blue channel is presented in the histogram. All three channels are generally to the left of the histogram however. Am I going about it correctly?

Jirair
Edited ...
Like
bobzeq25 0.00
...
{deleted}
Edited ...
Like
 
Register or login to create to post a reply.