Here's the fits file. I want to see if I get a good blue or purple color on the galaxy ngc7331
https://www.mediafire.com/file/aacomkfk9wxmdq0/ngc7331_GraXpert.fits/file
https://www.mediafire.com/file/aacomkfk9wxmdq0/ngc7331_GraXpert.fits/file
NGC7331: is a blue color for the galaxy possible with my data?
Hi, I played around a bit with it, but it doesn't seem like there is color in your data. It is pretty noisy overall. To improve, you probably need more integration time overall, and longer individual subs.
Helpful Concise
I gave it a try; not much to work with unfortunately; you can do surely better than this, but not by a lot - I think.
You can get color from this easily by simply color correcting. It will give you the specific colors you are looking for, but I would prefer files that haven't been through GraXpert. I find that program destructive at best. As mentioned there are issues with noise and artifacts in there, I'm guessing the integration is low, so by bringing out the spiral arms we also get a lot of the noise coming out with them. Is this a fully calibrated master from ASIStudio?
To the left is the original with STF applied, to the right is a copy with calibrated colors (also with STF applied)
Pushing the image further than it should be will only result in strange colors and even more artifacts. As you can see the color noise in the image also has a blue-ish color and the spiral arms are barely separated from the noise. I think it just needs more data, and possibly some better calibration to deal with whatever is causing these gradients. Are you dithering?
To the left is the original with STF applied, to the right is a copy with calibrated colors (also with STF applied)
Pushing the image further than it should be will only result in strange colors and even more artifacts. As you can see the color noise in the image also has a blue-ish color and the spiral arms are barely separated from the noise. I think it just needs more data, and possibly some better calibration to deal with whatever is causing these gradients. Are you dithering?
Helpful Insightful Engaging Supportive
Of course you can, but you should upload the master file without GraX.
You can get color from this easily by simply color correcting. It will give you the specific colors you are looking for, but I would prefer files that haven't been through GraXpert. I find that program destructive at best. As mentioned there are issues with noise and artifacts in there, I'm guessing the integration is low, so by bringing out the spiral arms we also get a lot of the noise coming out with them. Is this a fully calibrated master from ASIStudio?
To the left is the original with STF applied, to the right is a copy with calibrated colors (also with STF applied)
I didn't use dithering. Here's the unedited fit file ( i took flats, darks, bias for calibration).
https://www.mediafire.com/file/wxqlqhp5x5476yo/ngc7331.fit/file
I like how it looks like in image20 in your post. Can you color correct in pixinsight? or is that the same as photometric color calibration or spectro color collabration? I'm on alt-az and my laptop battery ain't that good. To get better pics I'll have to get better equipment.
Ferran Bosch:
Of course you can, but you should upload the master file without GraX.
Great work here's the unedited version if you want to do anything with it.
https://www.mediafire.com/file/wxqlqhp5x5476yo/ngc7331.fit/file
i like how you can see the arms from the background.
Could you get the colors to look something like this or more specifically keep the core a more white/yellow color while the arms stay more blue. Also was the color of the galaxy caused by increasing saturation.
I didn't use dithering. Here's the unedited fit file ( i took flats, darks, bias for calibration).
https://www.mediafire.com/file/wxqlqhp5x5476yo/ngc7331.fit/file
I like how it looks like in image20 in your post. Can you color correct in pixinsight? or is that the same as photometric color calibration or spectro color collabration? I'm on alt-az and my laptop battery ain't that good. To get better pics I'll have to get better equipment.
I'll have a look at the unedited file as well
Yes that's the same image as the above, but stretched/non-linear. Normally I would apply SPCC, you can also manually adjust the histogram or apply scripts like "Autocolor" that calculates the RGB values. I find that the Autocolor script works fairly well quite often.
The issue with the arms is that they are barely any stronger in signal than the noise surrounding the galaxy. You can be "creative" with masks but likely it will look very unnatural and not reflect the actual arms very well IMO. Being careful with color/saturation is also key as you will pull some noise from being to aggressive, as well as get unnatural colors like the purple. It's easy to manually put colors in that don't necessarily belong there. I prefer to keep it straight though. Accepting how far the data can be pushed and then processing it as such is the best approach I think.
Helpful Insightful Respectful Engaging
Alright, after reviewing the unedited file it seems like there's a few prominent issues. If flats were applied they appear to not work, so something is off there. The rest may be collimation, focus, issues with tracking, backfocus or dew. Kind of hard to tell if it's one or several of those issues.
In order to combat this I applied a very crude method that I really only used as a very last resort, creating a synthetic flat and then subtracting it from the image. This destroys everything that isn't masked off, so there are drawbacks:
Resulting in this (as you can see the masking was far from perfect, so there's still some stuff left there):
There's still a lot of noise there and it's not optimal. So still, if over-stretching, it will become prominent. Downsampling and clipping the black point doesn't really hide anything, but makes it a bit less extreme. But this is just an example of how you could nuke the image if there is a crisis.
The edit above looks over-processed to me, a lot of color noise and general noise being brought out, clipped black point and an overall less than good finish I think.
If you go a bit more careful on the stretching, while monitoring the noise levels you get a much smoother result but you sacrifice some brightness to the object. As mentioned the spiral arms lack a lot of SNR to bring them out nicely. That's my opinion at least. Here is a couple of more moderately stretched edits.
V1:
V2:
I didn't apply any MLT to the V1/V2 labelled images, but I did to the image labelled "over_processed". Besides that they both had the same amount of noise reduction
Also, keep in mind that the image you referred to also has a lot of noise/color noise as well as dead pixels. So it's far from perfect. Still, it was taken with a monochrome camera and apparently this gigantic telescope which likely has insane photon/resolution capture capabilities compared to whatever most of us mere mortals have at hand. So even though he had a short amount of total integration it is not comparable.
In order to combat this I applied a very crude method that I really only used as a very last resort, creating a synthetic flat and then subtracting it from the image. This destroys everything that isn't masked off, so there are drawbacks:
Resulting in this (as you can see the masking was far from perfect, so there's still some stuff left there):
There's still a lot of noise there and it's not optimal. So still, if over-stretching, it will become prominent. Downsampling and clipping the black point doesn't really hide anything, but makes it a bit less extreme. But this is just an example of how you could nuke the image if there is a crisis.
The edit above looks over-processed to me, a lot of color noise and general noise being brought out, clipped black point and an overall less than good finish I think.
If you go a bit more careful on the stretching, while monitoring the noise levels you get a much smoother result but you sacrifice some brightness to the object. As mentioned the spiral arms lack a lot of SNR to bring them out nicely. That's my opinion at least. Here is a couple of more moderately stretched edits.
V1:
V2:
I didn't apply any MLT to the V1/V2 labelled images, but I did to the image labelled "over_processed". Besides that they both had the same amount of noise reduction
Also, keep in mind that the image you referred to also has a lot of noise/color noise as well as dead pixels. So it's far from perfect. Still, it was taken with a monochrome camera and apparently this gigantic telescope which likely has insane photon/resolution capture capabilities compared to whatever most of us mere mortals have at hand. So even though he had a short amount of total integration it is not comparable.
Helpful Engaging Supportive
Alright, after reviewing the unedited file it seems like there's a few prominent issues. If flats were applied they appear to not work, so something is off there. The rest may be collimation, focus, issues with tracking, backfocus or dew. Kind of hard to tell if it's one or several of those issues.
In order to combat this I applied a very crude method that I really only used as a very last resort, creating a synthetic flat and then subtracting it from the image. This destroys everything that isn't masked off, so there are drawbacks:
Resulting in this (as you can see the masking was far from perfect, so there's still some stuff left there):
There's still a lot of noise there and it's not optimal. So still, if over-stretching, it will become prominent. Downsampling and clipping the black point doesn't really hide anything, but makes it a bit less extreme. But this is just an example of how you could nuke the image if there is a crisis.
The edit above looks over-processed to me, a lot of color noise and general noise being brought out, clipped black point and an overall less than good finish I think.
If you go a bit more careful on the stretching, while monitoring the noise levels you get a much smoother result but you sacrifice some brightness to the object. As mentioned the spiral arms lack a lot of SNR to bring them out nicely. That's my opinion at least. Here is a couple of more moderately stretched edits.
V1:
V2:
I didn't apply any MLT to the V1/V2 labelled images, but I did to the image labelled "over_processed". Besides that they both had the same amount of noise reduction
Also, keep in mind that the image you referred to also has a lot of noise/color noise as well as dead pixels. So it's far from perfect. Still, it was taken with a monochrome camera and apparently this gigantic telescope which likely has insane photon/resolution capture capabilities compared to whatever most of us mere mortals have at hand. So even though he had a short amount of total integration it is not comparable.
Thanks for the info. Yeah I had a dew issue and I took the image over multiple nights giving me a lot of stacking artifacts. I think my focus was good but perhaps I can look at that more. I like how you used a synthetic flat. I'll look into that more. I like how it looks in the over processed version the best. Will you post your final edit of that here?
Jan Erik Vallestad:
In order to combat this I applied a very crude method that I really only used as a very last resort, creating a synthetic flat and then subtracting it from the image. This destroys everything that isn't masked off, so there are drawback
How did you create the synthetic flat? with DBE?
Ferran Bosch:
How did you create the synthetic flat? with DBE?
Let me just re-iterate that I dislike this procedure, I cannot say enough how destructive it is. So I strongly suggest not to use it - but:
First make a copy and mask off the galaxies using the CloneStamp tool with reduced opacity and some increase in softness:
Getting this 100% accurately can be hard as there are so many different gradients and patterns, but after this make sure to use MMT in order to blur the cloned image:
Then apply some pixelmath to create the cleaned image:
As I stated above the only things left are the ones masked in the "flat" image. Any hint of IFN, dust, nebula, small galaxies (or anything else) left unmasked will disappear. So the image looks very unnatural - but as a last resort if all else fails and you are unable to calibrate it properly due to lack of flats etc. it can be applied.
I would however strongly suggest that OP, in stead of working around bad frames, cull them in stead. The overall result will look better with only good frames in it. The main galaxy will suffer as a result of including badly tracked/focused frames as well as frames with dew/flats issues.
Helpful Insightful
Thanks for the info. Yeah I had a dew issue and I took the image over multiple nights giving me a lot of stacking artifacts. I think my focus was good but perhaps I can look at that more. I like how you used a synthetic flat. I'll look into that more. I like how it looks in the over processed version the best. Will you post your final edit of that here?
I see, the multiple nights/field rotation issue is possible to work around by cropping. The frames with dew or other issues I would discard to get the best possible image. Most likely the galaxy suffers far more than it gains from including those. Stacking it in PI (as an example because that is my tool of choice) would discard those frames during stacking automatically.
I didn't actually save the images, but I'll have another go at it for you. It's not identical I'm afraid as I just threw a lot of different processes at this image yesterday in an attempt to overdo it. But I made another attempt to push it beyond my personal barriers here
Besides some of the core it's mainly noise though.
Jpg:
