NGC 2011 (LMC) processing issues

Hi all,

I have been imaging the LMC, specifically NGC 2011 and 2014 for the last 4 nights, and have collected about 12 hours of data. I am using a ZWO 1600MM with ZWO Ha and Oiii filters (low quality i understand). They are mounted on the Sharpstar AL-90R and NEQ6 mount.

When I started processing the images, I thought that I would have some ease with it considering I have 12 hours of data, however it seems to still have a lot of noise that is buried within the data. All the subs have been taken in new moon, about Bortle 4/5 so not bad light pollution and they were all 600s subs. Here is a link to the master lights from Pixinsight if anyone wants to have a play. Below are some screenshots of how bad the noise is. I have used NoiseXterminator and after stretching, it is still terrible amounts of noise. In stacking I have used darks, flats and bias.

Am I doing something wrong, or can I do something better? Help would be much appreciated as it has happened to a couple of my other images as well.

- Close shot of STF and Noise Exterminator Applied

- Close shot without NoiseXterminator (STF still applied)

- After stretching using Histogram Transformation and Curves Transformation

Please let me know if I can provide more screenshots for more examples.

Looks about normal for HOO combination if you ask me. I wouldn't be worried. 12 hours isn't a ton of time for narrowband. Assuming that's 6 hours per channel, 10 minute subs, 36 subs per. If it makes you feel better, I'm in the middle of about 70 hours of HOO data and it still has the same noise profile

I see. I’m used to seeing images with so much smoother data and when I look closely, they aren’t crunchy like mine is. Is that a bi-product of having an effect older camera or is there some technique that I’m missing?

Lachlan Wilson:
I see. I’m used to seeing images with so much smoother data and when I look closely, they aren’t crunchy like mine is. Is that a bi-product of having an effect older camera or is there some technique that I’m missing?

Small aperture and yesterday's sensor generation does that. You should be really on a 3"/px to be on an efficient data gather profile, even more so with NB filters.

andrea tasselli:
Lachlan Wilson:
I see. I’m used to seeing images with so much smoother data and when I look closely, they aren’t crunchy like mine is. Is that a bi-product of having an effect older camera or is there some technique that I’m missing?

Small aperture and yesterday's sensor generation does that. You should be really on a 3"/px to be on an efficient data gather profile, even more so with NB filters.

What do you mean 3”/px? Is that a hard number or should I be below that? My resolution is 1.31”/px in which I try my absolute best to keep my guiding below (99.5% of the time it is). Could you expand on that for me?

Lachlan Wilson:
What do you mean 3”/px? Is that a hard number or should I be below that? My resolution is 1.31”/px in which I try my absolute best to keep my guiding below (99.5% of the time it is). Could you expand on that for me?

What I am saying is that you would need to be at least twice your current pixel scale, which is far to expensive to keep for such a small aperture and type of sensor. Your sweet spot is close to 3"/px and in fact it should your nominal one. Since it is probably unachieveable (not sure how much you can reduce your focal length to) I'd stick with binning x2 at the very least. That should give you a signficant improvement with photon noise, which is what you see here.

That's the above (binning done in software) end result. Not optimised by any means but enough to show the potential (btw, you have some issue with your optics, PSF is slightly triangular).

andrea tasselli:
That's the above (binning done in software) end result. Not optimised by any means but enough to show the potential (btw, you have some issue with your optics, PSF is slightly triangular).

Thanks for trying! Looks very good. I shall have a try of using 2x2 binning next clear night (could be a while). I assume I can't intergrate the data with the existing 1x1 data? And to take it up to full res, I just need to drizzle it when stacking?

In terms of the optics, I'm pretty sure that it is from the tracking which wasnt perfect. I am ironing out a couple of issues with the tracking and guiding so that is likely the problem.

Thanks

Lachlan Wilson:
Thanks for trying! Looks very good. I shall have a try of using 2x2 binning next clear night (could be a while). I assume I can't intergrate the data with the existing 1x1 data? And to take it up to full res, I just need to drizzle it when stacking?

With CMOS is probably best to either a) bin at the source and forget about drizzling unless you get an average PSF of 2px or less or b) bin after pre-processing on case by case basis. Of course you can mix and match after the lights have been stacked.

andrea tasselli:
Lachlan Wilson:
Thanks for trying! Looks very good. I shall have a try of using 2x2 binning next clear night (could be a while). I assume I can't intergrate the data with the existing 1x1 data? And to take it up to full res, I just need to drizzle it when stacking?

With CMOS is probably best to either a) bin at the source and forget about drizzling unless you get an average PSF of 2px or less or b) bin after pre-processing on case by case basis. Of course you can mix and match after the lights have been stacked.

Would software binning have the same effect? That is, calibrating all the subs in WBPP, using the IntegerResample on every sub and then stack them. Would that produce the same results or would I need to actually do hardware binning?

Is the 3"/px a set standard? If I go and use a different scope, how would the different focal length effect it?

Lachlan Wilson:
Would software binning have the same effect? That is, calibrating all the subs in WBPP, using the IntegerResample on every sub and then stack them. Would that produce the same results or would I need to actually do hardware binning?

For most CMOS sensors, including yours, there is no real hardware binning so doing after the pixels are read in camera or after they are transmitted to the host computer does not make much difference at all (except for potential noise issue on the USB cabling).

Lachlan Wilson:
Is the 3"/px a set standard? If I go and use a different scope, how would the different focal length effect it?

No, it isn't, but it is my best guess on where you should land in terms of efficiency given your aperture and sensor QE. Different telescopes/lenses would yield different numbers but as my test image shows there is no loss in actual resolution and a gain of a factor of 2 in noise. Which means 4 times more effective exposure. In other words you would have to capture for 4x as long (assuming equal conditions) to have the same results.

andrea tasselli:
For most CMOS sensors, including yours, there is no real hardware binning so doing after the pixels are read in camera or after they are transmitted to the host computer does not make much difference at all (except for potential noise issue on the USB cabling).

No, it isn't, but it is my best guess on where you should land in terms of efficiency given your aperture and sensor QE. Different telescopes/lenses would yield different numbers but as my test image shows there is no loss in actual resolution and a gain of a factor of 2 in noise. Which means 4 times more effective exposure. In other words you would have to capture for 4x as long (assuming equal conditions) to have the same results.

Thanks for that info. Is there anywhere I can find out the ideal for different cameras? I am going to be using some really large telescopes soon and maybe some different cameras so am keen to find the actual numbers for those.

thanks!

Lachlan Wilson:
Thanks for that info. Is there anywhere I can find out the ideal for different cameras? I am going to be using some really large telescopes soon and maybe some different cameras so am keen to find the actual numbers for those.

I'm afraid the question needs qualifiers. Are you optimizing for resolution or SNR? If it is the former then 1/3 of your average (across the board) seeing expressed as the median FHWM value of the worst filter (normally B or OIII) would do the trick. For the latter there is no hard and fast rule but, as an example, you and I shot at about the same focal length (603 for you and 575 for me) with essentially the same pixel size. I am at 1.35"/pixel and average 3.3" FWHM (panchromatic as I use an OSC camera), or expressed in pixels around 2.3 px. You have a scale of 1.31"/px and have an average FWHM (from the OIII shot) of 3.5px, that is 4.6". Efficiency would require an effective pixel size for your typical integrated seeing (that is the one taken over an extended amount of time, in your case 600s I believe) about the same as mine, smaller if anything. Even looking from the pointing of optimizing image scale with regard seeing you should be shooting @ 1.53"/px, not 1.31. That alone is a net loss of 36% of incoming light.

Filters also are detrimental to SNR, by and large, but most often they cannot be avoided.

andrea tasselli:
Efficiency would require an effective pixel size for your typical integrated seeing (that is the one taken over an extended amount of time, in your case 600s I believe) about the same as mine, smaller if anything.

Filters also are detrimental to SNR, by and large, but most often they cannot be avoided.

Not quite sure I understand this part. I am definitely looking to priorities SNR over resolution.

Even looking from the pointing of optimizing image scale with regard seeing you should be shooting @ 1.53"/px, not 1.31. That alone is a net loss of 36% of incoming light

Why is it 1.53"/px? How is it net loss of light?

And on another note, I have seen thisimage from James Baguley on Instagram, of the Horsehead nebula. I completely understand that is a different target with more nebulae, and that it is a different setup to mine, but I am interested in how one gets such a smooth and clean image. Especially considering that it is the same amount of integration time as me. It is a target that I am looking to redo soon but I am worried because the last 2 times I've tried, it always comes out super duper noisy. I have even had a couple thoughts of is it the camera and is it just getting to old?

Thanks

Lachlan Wilson:
Not quite sure I understand this part. I am definitely looking to priorities SNR over resolution.

Then try to achieve a FWHM of 2-2.2 pixels.

Lachlan Wilson:
Why is it 1.53"/px? How is it net loss of light?

4.6"/3=1.53", remember 3 pixels per FWHM (in the example 4.6")? Well, isn't actually a loss as if it went into a black hole, only it is spread out more than it should and thus reduces the SNR.

Lachlan Wilson:
And on another note, I have seen this image from James Baguley on Instagram, of the Horsehead nebula. I completely understand that is a different target with more nebulae, and that it is a different setup to mine, but I am interested in how one gets such a smooth and clean image. Especially considering that it is the same amount of integration time as me. It is a target that I am looking to redo soon but I am worried because the last 2 times I've tried, it always comes out super duper noisy. I have even had a couple thoughts of is it the camera and is it just getting to old?

A LOT of things can be different, starting from local seeing and darkness. Plus, it may have a much bigger scope than yours. That may go to a great length in explaining it. Only compare images taken side by side with the same aperture and image scale.

Your camera isn't particularly good by current standards, starting from its sensitivity. Modern cameras are on an entirely different league I'm afraid. Aging, I don't think is really a concern when it is just few years old. If it were 10 years old it would be another story...

andrea tasselli:
Then try to achieve a FWHM of 2-2.2 pixels.

A LOT of things can be different, starting from local seeing and darkness. Plus, it may have a much bigger scope than yours. That may go to a great length in explaining it. Only compare images taken side by side with the same aperture and image scale.

Your camera isn't particularly good by current standards, starting from its sensitivity. Modern cameras are on an entirely different league I'm afraid. Aging, I don't think is really a concern when it is just few years old. If it were 10 years old it would be another story...

How do I try and achieve that FWHM? Is there some changes to my setup that will help me achieve this?

Yes i figured there were many different factors. It is a 10" scope that was used so much larger than mine, and slightly darker skies than mine.

Yeah my camera I think would possibly be getting close to 10 years old which is unfortunate. I think i am the 3rd owner with each before me having it for at least 4 or 5 years. Any way I can check how old from a serial number?

Lachlan Wilson:
How do I try and achieve that FWHM? Is there some changes to my setup that will help me achieve this?

You can only change your pixel scale or get a reducer, if there is one to be had. Assuming the latter isn't available that means that, unless you buy another camera, you have to bin your camera and be done with it. I am assuming you have been religious in getting the sharpest focus there is and the scope has no aberrations, that is up to you.

Lachlan Wilson:
Yeah my camera I think would possibly be getting close to 10 years old which is unfortunate. I think i am the 3rd owner with each before me having it for at least 4 or 5 years. Any way I can check how old from a serial number?

They can by looking into the firmware data. Not sure users can though. Best contact ZWO and see what they say.

Thanks for that info. There isn't a dedicated reducer for the scope as I beleive it is one of very few made of the brand. I have an EAF so focus has been automated for about a year and a half now.

And I shall look into it. Anything I can do to help the sensisitvity of the camera bar getting a new, more modern one?

Lachlan Wilson:
And I shall look into it. Anything I can do to help the sensisitvity of the camera bar getting a new, more modern one?

Bin it (as in binning the sensor, not chucking it into the trash!). That's the way forward.

Ok will do next imaging night. Thanks for all the help. A lot of info dumped here that I gotta process over the next couple of sessions.

Related discussions