[RCC] Horsehead Nebula

3.01 akshay87kumar ... · view views · 3 likes · Share link
Request your feedback and critique on on the image below. This is a quick one night integration (5 hours only), will be taking another pass by adding more data and including improved post-processing based on critiques! What I can see very evident is the patchiness of colours - is it the nebula itself, or the background extraction not done too well, or the stretches too agressive? This time, I also upgraded a 0.8x field flattener to WO GT71 APO, and decided to go aggressive on dithering (every 2 frames) to make use of drizzle this time. 180s subs, total ~5 hours in Bortle 8 sky. Workflow is as follows: 1. WBPP with drizzling (2x, 0.8), followed by 200% downsampling. Used the drizzled integrated image going forward (difference was very visible!) [EDIT]: Missed writing about SetiAstro ADBE step right after WBPP 2. Background Neutralization 3. SPCC - I think this is not requiredfor a HOO palette. For traditional colours, I used 'L-Extreme' option in the filter dropdowns for R,G and B. I have seen different views here, so will be good to hear feedback. Even if I use narrowband filters and enter 7nm, the final regression graph i get is mostly similar. So I am unsure which one to use. 4. BlurXterminator at default settings (playing around with trial) - Earlier when I did not have the field flattener, I used to run the starless image through SIRIL deconvolution to remove the elongatedness of stars. The field flattener is pretty awesome here! But the results of BlurXterminator have taken it one level further. 5. Starless and Star marsk using StarNet Traditional Palette: Executed the following on Starless Image 6. NoiseXterminator (70%) 7. SetiAstro statistical stretch (median 18%, unlinked stretch, curves boost almost around mid position of slider) 8. Some masked curves stretches (RGB S curve, R boost, Saturation boost) Starless Image: 9. NB to RGB SetiAstro Script with color boost and star stretch 10. Added the starless and star images (Used straightforward Starless + Star, as there were no bright cores that I had struggled with orion nebula earlier). If you have other recommendations apart frmo usual addition or screening, please do share your thoughts! 11. NoiseXterminator (75%) 12. NoiseXterminator (60%) HOO Palette: This is one of my first images, havent even seen too many videos about this, just playing around for now. 11. Used the outcome of starless + star image, just ran it through Narrowband normalization. Played around with sliders, and realized the defaults just gave the best results for now. Please do share your feedback, will incorporate it along with more data for the final processed image. Edited ... Like

3.01

...

· 3 likes

Request your feedback and critique on on the image below. This is a quick one night integration (5 hours only), will be taking another pass by adding more data and including improved post-processing based on critiques!

What I can see very evident is the patchiness of colours - is it the nebula itself, or the background extraction not done too well, or the stretches too agressive?

This time, I also upgraded a 0.8x field flattener to WO GT71 APO, and decided to go aggressive on dithering (every 2 frames) to make use of drizzle this time. 180s subs, total ~5 hours in Bortle 8 sky.

Workflow is as follows:
1. WBPP with drizzling (2x, 0.8), followed by 200% downsampling. Used the drizzled integrated image going forward (difference was very visible!)
[EDIT]: Missed writing about SetiAstro ADBE step right after WBPP
2. Background Neutralization
3. SPCC - I think this is not requiredfor a HOO palette. For traditional colours, I used 'L-Extreme' option in the filter dropdowns for R,G and B. I have seen different views here, so will be good to hear feedback. Even if I use narrowband filters and enter 7nm, the final regression graph i get is mostly similar. So I am unsure which one to use.
4. BlurXterminator at default settings (playing around with trial) - Earlier when I did not have the field flattener, I used to run the starless image through SIRIL deconvolution to remove the elongatedness of stars. The field flattener is pretty awesome here! But the results of BlurXterminator have taken it one level further.
5. Starless and Star marsk using StarNet

Traditional Palette: Executed the following on Starless Image
6. NoiseXterminator (70%)
7. SetiAstro statistical stretch (median 18%, unlinked stretch, curves boost almost around mid position of slider)
8. Some masked curves stretches (RGB S curve, R boost, Saturation boost)

Starless Image:
9. NB to RGB SetiAstro Script with color boost and star stretch
10. Added the starless and star images (Used straightforward Starless + Star, as there were no bright cores that I had struggled with orion nebula earlier). If you have other recommendations apart frmo usual addition or screening, please do share your thoughts!
11. NoiseXterminator (75%)
12. NoiseXterminator (60%)

HOO Palette: This is one of my first images, havent even seen too many videos about this, just playing around for now.
11. Used the outcome of starless + star image, just ran it through Narrowband normalization. Played around with sliders, and realized the defaults just gave the best results for now.

Please do share your feedback, will incorporate it along with more data for the final processed image.

Edited ...

0.00 FilippoTib ... · view views · Share link
Hi. Am I wrong or did you not do a gradient removal (ABE/DBE/Graxpert/AutomaticDBE) at the beginning of your processing? How come a double pass of NXT at the end of the workflow? p.s., why an unlinked stretch if you had done color calibration first? By doing this, don't you lose the profile created by SPCC? F.T. Clear skies! Edited ... Like

3.01 akshay87kumar Topic starter ... · view views · 1 like · Share link
Hi. Am I wrong or did you not do a gradient removal (ABE/DBE/Graxpert/AutomaticDBE) at the beginning of your processing? How come a double pass of NXT at the end of the workflow? p.s., why an unlinked stretch if you had done color calibration first? By doing this, don't you lose the profile created by SPCC? F.T. Clear skies! I missed writing. Yes, the step immediately after wbpp is SetiAstro ADBE. Double pass of NXT as there was still noise left which I thought to remove. Regarding unlinked stretch - went with what "looked" better as I wasn't sure if ai wanted the traditional or HOO palette! But valid point, I wouldn't do unlinked along with spcc Edited ... Like

3.01

akshay87kumar

Topic starter

...

· 1 like

· Share link

Hi. Am I wrong or did you not do a gradient removal (ABE/DBE/Graxpert/AutomaticDBE) at the beginning of your processing?

How come a double pass of NXT at the end of the workflow?

p.s., why an unlinked stretch if you had done color calibration first? By doing this, don't you lose the profile created by SPCC?

F.T.
Clear skies!

I missed writing. Yes, the step immediately after wbpp is SetiAstro ADBE. Double pass of NXT as there was still noise left which I thought to remove.

Regarding unlinked stretch - went with what "looked" better as I wasn't sure if ai wanted the traditional or HOO palette! But valid point, I wouldn't do unlinked along with spcc

Edited ...

0.00 FilippoTib ... · view views · 2 likes · Share link
Hi. Am I wrong or did you not do a gradient removal (ABE/DBE/Graxpert/AutomaticDBE) at the beginning of your processing? How come a double pass of NXT at the end of the workflow? p.s., why an unlinked stretch if you had done color calibration first? By doing this, don't you lose the profile created by SPCC? F.T. Clear skies! I missed writing. Yes, the step immediately after wbppn is SetiAstro ADBE. I would tell you to very quickly retry the workflow by changing gradient reduction method (try graxpert or gradient correction in pixinsight). I don't think it's that but it's worth a quick 5 minute test Like

0.00

FilippoTib

...

· 2 likes

· Share link

Hi. Am I wrong or did you not do a gradient removal (ABE/DBE/Graxpert/AutomaticDBE) at the beginning of your processing?

How come a double pass of NXT at the end of the workflow?

p.s., why an unlinked stretch if you had done color calibration first? By doing this, don't you lose the profile created by SPCC?

F.T.
Clear skies!

I missed writing. Yes, the step immediately after wbppn is SetiAstro ADBE.

I would tell you to very quickly retry the workflow by changing gradient reduction method (try graxpert or gradient correction in pixinsight). I don't think it's that but it's worth a quick 5 minute test

3.58 Mikolaj Wadowski ... · view views · 2 likes · Share link
I think your workflow can be improved in some areas: 1. WBPP with drizzling (2x, 0.8), followed by 200% downsampling. Used the drizzled integrated image going forward (difference was very visible!) This should give the exact same result as drizzling 1x with the same dropshrink. You will save a bit of stacking time if you just drizzle 1x. I missed writing. Yes, the step immediately after wbpp is SetiAstro ADBE. ADBE seems to have killed a lot of signal in the Ha region above the horsehead and introduced patches everywhere. I suggest you do normal DBE instead - it might take a bit to learn but it will pretty much always give better results than current automatic options. Or better yet, use MSGR if you have a reference image available. 3. SPCC - I think this is not requiredfor a HOO palette. For traditional colours, I used 'L-Extreme' option in the filter dropdowns for R,G and B. I have seen different views here, so will be good to hear feedback. Even if I use narrowband filters and enter 7nm, the final regression graph i get is mostly similar. So I am unsure which one to use. If you're going to change the color balance manually later anyway, SPCC is sort of useless. You might as well do manual color balance right away. What your workflow is missing is subtracting Ha signal leaks from your green and blue channels. The color matrix in OSC cameras lets some red through green and blue filters, leading to the Ha signal contaminating your Oiii signal and making the Ha orange. Using DBXtract right after color calibration takes care of this issue in a simple, one-click way. 5. Starless and Star marsk using StarNet For a lot of people, this is a personal preference but I find that noise reduction before star extraction gives better results. But perhaps noisex is different, I don't use it. 6. NoiseXterminator (70%) 11. NoiseXterminator (75%) 12. NoiseXterminator (60%) That's way too much noise reduction in my opinion. It made your final image look a bit plastic-y to me. And I think you'd be better off with just one pass of denoising at the beginning with maybe some slight additional denoise as the final step. Plus, I really recommend trying out different denoising tools, deepsnr or graxpert in particular. From comparisons I have seen, even graxpert outperforms noisex nowadays. Deepsnr especially is the one to try - it works like magic if you satisfy the conditions for it to work. 10. Added the starless and star images (Used straightforward Starless + Star, as there were no bright cores that I had struggled with orion nebula earlier). If you have other recommendations apart frmo usual addition or screening, please do share your thoughts! Screening often gives me more natural-looking results, you just need to stretch the stars a bit more. You can also try the pixelmath formula from Charles Hagen's tutorial. Hopefully all this helps! CS, Mikolaj Like

3.58

Mikolaj Wadowski

...

· 2 likes

· Share link

I think your workflow can be improved in some areas:

1. WBPP with drizzling (2x, 0.8), followed by 200% downsampling. Used the drizzled integrated image going forward (difference was very visible!)

This should give the exact same result as drizzling 1x with the same dropshrink. You will save a bit of stacking time if you just drizzle 1x.

I missed writing. Yes, the step immediately after wbpp is SetiAstro ADBE.

ADBE seems to have killed a lot of signal in the Ha region above the horsehead and introduced patches everywhere. I suggest you do normal DBE instead - it might take a bit to learn but it will pretty much always give better results than current automatic options. Or better yet, use MSGR if you have a reference image available.

3. SPCC - I think this is not requiredfor a HOO palette. For traditional colours, I used 'L-Extreme' option in the filter dropdowns for R,G and B. I have seen different views here, so will be good to hear feedback. Even if I use narrowband filters and enter 7nm, the final regression graph i get is mostly similar. So I am unsure which one to use.

If you're going to change the color balance manually later anyway, SPCC is sort of useless. You might as well do manual color balance right away.

What your workflow is missing is subtracting Ha signal leaks from your green and blue channels. The color matrix in OSC cameras lets some red through green and blue filters, leading to the Ha signal contaminating your Oiii signal and making the Ha orange. Using DBXtract right after color calibration takes care of this issue in a simple, one-click way.

5. Starless and Star marsk using StarNet

For a lot of people, this is a personal preference but I find that noise reduction before star extraction gives better results. But perhaps noisex is different, I don't use it.

6. NoiseXterminator (70%)
11. NoiseXterminator (75%)
12. NoiseXterminator (60%)

That's way too much noise reduction in my opinion. It made your final image look a bit plastic-y to me. And I think you'd be better off with just one pass of denoising at the beginning with maybe some slight additional denoise as the final step.
Plus, I really recommend trying out different denoising tools, deepsnr or graxpert in particular. From comparisons I have seen, even graxpert outperforms noisex nowadays. Deepsnr especially is the one to try - it works like magic if you satisfy the conditions for it to work.

10. Added the starless and star images (Used straightforward Starless + Star, as there were no bright cores that I had struggled with orion nebula earlier). If you have other recommendations apart frmo usual addition or screening, please do share your thoughts!

Screening often gives me more natural-looking results, you just need to stretch the stars a bit more. You can also try the pixelmath formula from Charles Hagen's tutorial.

Hopefully all this helps!

CS,
Mikolaj

3.01 akshay87kumar Topic starter ... · view views · Share link
Mikolaj Wadowski: I think your workflow Thanks Mikolaj for the detailed review. Will go through the resources you shared and incorporate them in next pass. Appreciate your help! Mikolaj Wadowski: What your workflow is missing is subtracting Ha signal leaks from your green and blue channels I am hearing/reading this for the first time. Will go through details and check. Like

8.79 Jonny Bravo ... · view views · 2 likes · Share link
I don't see any mention of calibration frames, and that patterning is not part of the emission, but rather looks like failed flat calibration. For example, and I know this isn't the same target, but take a look at what happens if I stack Ha data from my 294MM Pro without using calibration frames: Obviously there's amp glow, but pay attention to the brighter areas near the corners and under the target. Here's what a typical Ha flat looks like with that sensor: Weird, right? The important thing to look at is the pattern around the edges and in the center and compare to the uncalibrated data. Looks pretty similar. OK, now let's look at that same data, but this time calibrated: The pattern is gone, as is the amp glow. So, my suggestion is to calibrate your data. Worry about post-processing after you figure calibration out . Like

8.79

Jonny Bravo

...

· 2 likes

· Share link

I don't see any mention of calibration frames, and that patterning is not part of the emission, but rather looks like failed flat calibration. For example, and I know this isn't the same target, but take a look at what happens if I stack Ha data from my 294MM Pro without using calibration frames:

Obviously there's amp glow, but pay attention to the brighter areas near the corners and under the target. Here's what a typical Ha flat looks like with that sensor:

Weird, right? The important thing to look at is the pattern around the edges and in the center and compare to the uncalibrated data. Looks pretty similar. OK, now let's look at that same data, but this time calibrated:

The pattern is gone, as is the amp glow.

So, my suggestion is to calibrate your data. Worry about post-processing after you figure calibration out

3.01 akshay87kumar Topic starter ... · view views · Share link
Jonny Bravo: I don't see any mention of calibration frames I did calibrate the frames - used ~60 each of bias, darks (180s) and flats (0.4s with same filter). I agree, my L-extreme filter also gives me weird vignette patterns, very similar to the one you had posted. The only one challenge I am facing with imaging is that I am forced to image at 2 deg C. My ASI294MC Pro will always frost on the sensor - have tried all techniques including replacing dessicants, sealing tightly, adding a dew heater strip, very gradual cooling, etc. It will always, without fail, frost and I am constrained to image at 2 deg C to avoid this issue. More details about the frosting issue in this thread. Edited ... Like

3.01

akshay87kumar

Topic starter

...

· Share link

Jonny Bravo:
I don't see any mention of calibration frames

I did calibrate the frames - used ~60 each of bias, darks (180s) and flats (0.4s with same filter). I agree, my L-extreme filter also gives me weird vignette patterns, very similar to the one you had posted.
The only one challenge I am facing with imaging is that I am forced to image at 2 deg C. My ASI294MC Pro will always frost on the sensor - have tried all techniques including replacing dessicants, sealing tightly, adding a dew heater strip, very gradual cooling, etc. It will always, without fail, frost and I am constrained to image at 2 deg C to avoid this issue. More details about the frosting issue in this thread.

Edited ...

8.79 Jonny Bravo ... · view views · Share link
The flat calibration clearly failed. I figured you were using some flavor of 294 based on the pattern evident in the images. There are a metric ton of posts regarding the calibration woes of a 294MC with an L-eXtreme filter. Some suggestions to try: 1. Don't use bias frames. Use darks for your flats instead (you'll see the terms dark flats and flat darks for these). 2. Take longer flats. Typical recommendation is to go for at least 2-3 seconds. The 294 can be quirky and doesn't behave nicely at short exposures. 3. Use balanced natural white light, not an iPad or similar. 4. Try for a lower ADU on the flat. Sometimes one of the channels can clip 5. Look at the white balance settings in the camera (ZWO app should show you). Some ZWO cameras have blue set to like 95 for one reason or another. Make sure the R, G and B channels are set equally. I can certainly sympathize with the difficulty of calibrating a 294. I've spent literal days of my life in the garage doing nothing but taking flats . Like

8.79

Jonny Bravo

...

· Share link

The flat calibration clearly failed. I figured you were using some flavor of 294 based on the pattern evident in the images.

There are a metric ton of posts regarding the calibration woes of a 294MC with an L-eXtreme filter. Some suggestions to try:

1. Don't use bias frames. Use darks for your flats instead (you'll see the terms dark flats and flat darks for these).
2. Take longer flats. Typical recommendation is to go for at least 2-3 seconds. The 294 can be quirky and doesn't behave nicely at short exposures.
3. Use balanced natural white light, not an iPad or similar.
4. Try for a lower ADU on the flat. Sometimes one of the channels can clip
5. Look at the white balance settings in the camera (ZWO app should show you). Some ZWO cameras have blue set to like 95 for one reason or another. Make sure the R, G and B channels are set equally.

I can certainly sympathize with the difficulty of calibrating a 294. I've spent literal days of my life in the garage doing nothing but taking flats smile

3.01 akshay87kumar Topic starter ... · view views · Share link
Jonny Bravo: The flat calibration clearly failed. My suspect is more on a failed background extraction than a failed flats. Will work out a better version and post once updated. Like

8.79 Jonny Bravo ... · view views · Share link
It is not a failed background extraction. That worked fine and is exposing the underlying issue - the failed flat calibration. Like

3.58 Mikolaj Wadowski ... · view views · Share link
It's probably a mix of both. The patches could be from flat calibration error but there are also signs of poor background extraction, especially visible in the bright Ha above the horsehead. It's very unlikely that it's from flats under/overcorrecting. Like

8.79 Jonny Bravo ... · view views · 1 like · Share link
I would argue it's definitely patterning from failed calibration. Using my own data from the screenshot earlier in the thread, I decided to stack just the lights and the darks - no flats. I used WBPP with local normalization, cosmetic correction, etc. The patterning was clearly evident. I used standard ABE and it stood out even more. However, I ran it through GraXpert at its default setting and the result was actually extremely good. Here's with ABE and default settings: Pattern still very clearly visible. Here's with GraXpert also at its default settings: That's not bad at all! Now, let's look at the properly calibrated master. No form of background extraction done at all: Now, with GraXpert applied in the same fashion as on the no-flat master: I honestly didn't expect GraXpert (or any other background extraction method) to work so well on the non-flat-calibrated data. I'm looking at both side by side on my screen and there's very little noticeable difference. A tiny bit more contrast in some areas, but certainly an impressive result. @akshay87kumar maybe download and try GraXpert on your data . Like

8.79

Jonny Bravo

...

· 1 like

· Share link

I would argue it's definitely patterning from failed calibration. Using my own data from the screenshot earlier in the thread, I decided to stack just the lights and the darks - no flats. I used WBPP with local normalization, cosmetic correction, etc. The patterning was clearly evident. I used standard ABE and it stood out even more. However, I ran it through GraXpert at its default setting and the result was actually extremely good.

Here's with ABE and default settings:

Pattern still very clearly visible. Here's with GraXpert also at its default settings:

That's not bad at all!

Now, let's look at the properly calibrated master. No form of background extraction done at all:

Now, with GraXpert applied in the same fashion as on the no-flat master:

I honestly didn't expect GraXpert (or any other background extraction method) to work so well on the non-flat-calibrated data. I'm looking at both side by side on my screen and there's very little noticeable difference. A tiny bit more contrast in some areas, but certainly an impressive result.

@akshay87kumar maybe download and try GraXpert on your data

3.01 akshay87kumar Topic starter ... · view views · Share link
Thanks, will dig back into the lights, flats and calibrated frames to see if I am able to observe such patterns. And ofcourse, explore the other tools too. Like

3.01 akshay87kumar Topic starter ... · view views · 1 like · Share link
Jonny Bravo: I would argue it's definitely patterning from failed calibration. You were so correct! I revisited WBPP (it actually restores to what settings were used last), and I can somehow see that the flats were not correctly addressed. I was actually trying to create master frames for my darks, flats and bias, and I had shoved all of them in together (just to save time). Didnt realize that the light frames had not picked up the flats correctly (cross visible against the flats). Usually PixInsight used to correctly pick up the bias, darks, flats and lights. But looks like this time it was a miss, and few lessons learnt - (i) every single step needs to be thoroughly checked. (ii) Do not change too many variables - i updated a reducer, did drizling, attempted to create master frames for all past data, etc. (iii) Try same step (e.g. background extraction) with multiple tools Posting the difference in integrated image with and without the flats (only background neutralization and statistical stretch done to export into jpg) After using the flats correctly When the flats were missed - clearly shows the vignetting in the border region The master flat for L-extreme filter This is where I totally missed linking the flats. Usually Pixinsight does this automatically, but looks like it broke when I attempted to create master data for all historical calibration frames taken so far. Like

3.01

akshay87kumar

Topic starter

...

· 1 like

· Share link

Jonny Bravo:
I would argue it's definitely patterning from failed calibration.

You were so correct! I revisited WBPP (it actually restores to what settings were used last), and I can somehow see that the flats were not correctly addressed. I was actually trying to create master frames for my darks, flats and bias, and I had shoved all of them in together (just to save time). Didnt realize that the light frames had not picked up the flats correctly (cross visible against the flats).
Usually PixInsight used to correctly pick up the bias, darks, flats and lights. But looks like this time it was a miss, and few lessons learnt - (i) every single step needs to be thoroughly checked. (ii) Do not change too many variables - i updated a reducer, did drizling, attempted to create master frames for all past data, etc. (iii) Try same step (e.g. background extraction) with multiple tools

Posting the difference in integrated image with and without the flats (only background neutralization and statistical stretch done to export into jpg)

After using the flats correctly

When the flats were missed - clearly shows the vignetting in the border region

The master flat for L-extreme filter

This is where I totally missed linking the flats. Usually Pixinsight does this automatically, but looks like it broke when I attempted to create master data for all historical calibration frames taken so far.

3.01 akshay87kumar Topic starter ... · view views · Share link
I spent some more time gathering data for L-extreme filter, as well as without the filter for RGB stars and hoping for additional nebula details. It is my first time imaging in broadband (Bortle 8-9). I have been able to get proper integration with the L-extreme filter, but it seems like the flats taken without the filter continue to give me grief. I have ensured to make no mess-up (e.g. not linking flats in WBPP), but my image continues to show a very wierd vigenette, which I would generally see only on images taken with L-extreme. What could be going wrong with the flats? Here is my workflow last night: 1. I took about 1.5 hrs of image without the L-extreme filter (no filter effectively). Noted the focus was at position 6743 2. Then I moved to L-extreme filter, noted that the focus was at position 6977 3. Took a series of flats using morning sky - Got the L-extreme flats as they were already in place. Then I moved the filter to 6743 and took seies of flats for no-filter. 4. To be on a safer side (I know flats always haunt me!), I also took another series of flats using white screen on my tab. Still safer side, I took series of dark flats using the white screen too. The L-extreme image comes out well. Vignettes are clearly cancelled out and the image is clean. It is the image without filter that continues to haunt me! Posting images taken without the filter below, and request your feedback on what is troubling here again. Interestingly, I did not see any noticeable difference in quality of image filter whether I used the flats generated using sky, or the white screen. Similarly, it did not make material difference if I took the flats or dark flats. Integrated Image (without filter) after calibration with flats - Only STF applied Integrated Image (without filter) without calbiration for flats - Only STF applied A typical flat frame (without filter, ensuring focus is at the same positon I took the light frames): viewed in ASIFits Viewer Like

3.01

akshay87kumar

Topic starter

...

· Share link

I spent some more time gathering data for L-extreme filter, as well as without the filter for RGB stars and hoping for additional nebula details. It is my first time imaging in broadband (Bortle 8-9). I have been able to get proper integration with the L-extreme filter, but it seems like the flats taken without the filter continue to give me grief.

I have ensured to make no mess-up (e.g. not linking flats in WBPP), but my image continues to show a very wierd vigenette, which I would generally see only on images taken with L-extreme. What could be going wrong with the flats?

Here is my workflow last night:
1. I took about 1.5 hrs of image without the L-extreme filter (no filter effectively). Noted the focus was at position 6743
2. Then I moved to L-extreme filter, noted that the focus was at position 6977
3. Took a series of flats using morning sky - Got the L-extreme flats as they were already in place. Then I moved the filter to 6743 and took seies of flats for no-filter.
4. To be on a safer side (I know flats always haunt me!), I also took another series of flats using white screen on my tab. Still safer side, I took series of dark flats using the white screen too.

The L-extreme image comes out well. Vignettes are clearly cancelled out and the image is clean. It is the image without filter that continues to haunt me! Posting images taken without the filter below, and request your feedback on what is troubling here again.

Interestingly, I did not see any noticeable difference in quality of image filter whether I used the flats generated using sky, or the white screen. Similarly, it did not make material difference if I took the flats or dark flats.

Integrated Image (without filter) after calibration with flats - Only STF applied

Integrated Image (without filter) without calbiration for flats - Only STF applied

A typical flat frame (without filter, ensuring focus is at the same positon I took the light frames): viewed in ASIFits Viewer

3.58 Mikolaj Wadowski ... · view views · Share link
Still safer side, I took series of dark flats using the white screen too. I hope you didn't take dark flats by shining a bright light down your scope... Anyway, your flats don't display any significant amp glow. Try calibrating with just dark frames and flats, no bias or dark flats. Unless the IMX492 did something weird that should properly calibrate your subs. Like

3.01 akshay87kumar Topic starter ... · view views · Share link
Mikolaj Wadowski: I hope you didn't take dark flats by shining a bright light down your scope.. No, i took a dark frame (no light source) for the same duration as my flats. Regardless whether I use bias or dark flats, the image remains the same. I see where the confusion came from - I too the dark flats not using white screen, but using the same duration as the white screen! Edited ... Like

	Distinct awarded users	Total awarded images
Image of the day
Top picks
Top pick nominations

	Image of the day	Top pick	Top pick nominations	Total submitted
Deep sky
Solar system
Extremely wide field
Star trails
Northern lights
Noctilucent clouds
Landscape
Artificial satellite

	Image of the day	Top pick	Top pick nominations	Total submitted
Backyard
Traveller
Own remote observatory
Amateur hosting facility
Public amaeteur data
Professional, scientific grade data
Mix of multiple sources
Other
Unknown

AstroBin

AstroBin

Image Index:

Contribution Index (beta):

[RCC] Horsehead Nebula Requests for constructive critique · akshay87kumar · ... · 17 · 739 · 16

AstroBin

AstroBin

Image Index:

Contribution Index (beta):

[RCC] Horsehead Nebula Requests for constructive critique · akshay87kumar · ... · 17 · 739 · 16

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(

Sorry

Share this forum post

Post editing not available

Sorry

Sorry

Sorry :-(