M27, AP1100 and a F/10 Celestron Edge


Tom Blahovici
 

Well I finally got an image from my Celestron 8" Edge on the AP1100.  My usual scope is an FSQ106 but I decided to try something with a longer focal length for those small fuzzies.
The Celestron 8 requires lots of effort to get a reasonable image.  I first tried with their 0.7 reducer.  I could never get anything decent from it.  Most likely the reducer itself.  There have been a number of them with less than stellar correction.  So I gave up and went to F/10.
Given that I live in the suburbs of Montreal, F/10 narrowband needs lots of data.  Try as I might I couldn't muster more than 12 subs or 20 minutes on each of Ha,S2 and O3 filters.
Well my usual workflow proved to be a disaster.  Lots of noise and I could never get a decent image.  This was made worse by my new filter, a O3 3nm filter from Chroma which is totally spectacular next to my Baader filters. The issue is not the Chroma but rather the baader filters which gave halos on almost all the stars.  The O3 has none (!) and has way less noise as well.  So the result was a image with a yellow cast of noise and yellow halos around the stars.
 What to do?
Well I noticed that on YouTube someone used a synthetic luminance channel which when then combined with SHO gave a "LRGB i.e. LSHO " composition.  So I took my best channel, O3 and used it as luminance.
Voila!  
Notice I said nothing about the mount... it just disappears into the background.  RMS tracking of 0.4 ArcSec at 2032mm focal length.


Roland Christen
 

That's an interesting rendition that I have not seen before.

Rolando

-----Original Message-----
From: Tom Blahovici <tom.va2fsq@...>
To: main@ap-gto.groups.io
Sent: Mon, Sep 26, 2022 7:40 pm
Subject: [ap-gto] M27, AP1100 and a F/10 Celestron Edge

Well I finally got an image from my Celestron 8" Edge on the AP1100.  My usual scope is an FSQ106 but I decided to try something with a longer focal length for those small fuzzies.
The Celestron 8 requires lots of effort to get a reasonable image.  I first tried with their 0.7 reducer.  I could never get anything decent from it.  Most likely the reducer itself.  There have been a number of them with less than stellar correction.  So I gave up and went to F/10.
Given that I live in the suburbs of Montreal, F/10 narrowband needs lots of data.  Try as I might I couldn't muster more than 12 subs or 20 minutes on each of Ha,S2 and O3 filters.
Well my usual workflow proved to be a disaster.  Lots of noise and I could never get a decent image.  This was made worse by my new filter, a O3 3nm filter from Chroma which is totally spectacular next to my Baader filters. The issue is not the Chroma but rather the baader filters which gave halos on almost all the stars.  The O3 has none (!) and has way less noise as well.  So the result was a image with a yellow cast of noise and yellow halos around the stars.
 What to do?
Well I noticed that on YouTube someone used a synthetic luminance channel which when then combined with SHO gave a "LRGB i.e. LSHO " composition.  So I took my best channel, O3 and used it as luminance.
Voila!  
Notice I said nothing about the mount... it just disappears into the background.  RMS tracking of 0.4 ArcSec at 2032mm focal length.


--
Roland Christen
Astro-Physics


Tom Blahovici
 

I am wondering if the different makes of filters had anything to do with this.
I used Ha and S2 filters (8nM) from Baader and my new O3(3nM) from Chroma$$$.

The O3 filter gave images with little noise and not a lot of stars.  The Baader filters had much more noise and there were many many more stars in the field of view.  The result was that when the three images were channel combined, most of the stars were yellow green as well as the noise.  Try as I may, I could not tame this at all.
When I used the O3 data as a Luminance layer with a typical LRGB type combination, all the color casts were gone as well as the noise.  It was like night and day.
Tom


Roland Christen
 

The "noise" is probably the background sky light, which a 3nm filter removes quite effectively but an 8nm filter lets some of it thru. The 8nm also lets much more starlight thru versus the 3.

Roland

-----Original Message-----
From: Tom Blahovici <tom.va2fsq@...>
To: main@ap-gto.groups.io
Sent: Tue, Sep 27, 2022 1:27 pm
Subject: Re: [ap-gto] M27, AP1100 and a F/10 Celestron Edge

I am wondering if the different makes of filters had anything to do with this.
I used Ha and S2 filters (8nM) from Baader and my new O3(3nM) from Chroma$$$.

The O3 filter gave images with little noise and not a lot of stars.  The Baader filters had much more noise and there were many many more stars in the field of view.  The result was that when the three images were channel combined, most of the stars were yellow green as well as the noise.  Try as I may, I could not tame this at all.
When I used the O3 data as a Luminance layer with a typical LRGB type combination, all the color casts were gone as well as the noise.  It was like night and day.
Tom

--
Roland Christen
Astro-Physics


Chris White
 

On Tue, Sep 27, 2022 at 02:27 PM, Tom Blahovici wrote:
I am wondering if the different makes of filters had anything to do with this.
I used Ha and S2 filters (8nM) from Baader and my new O3(3nM) from Chroma$$$.

The O3 filter gave images with little noise and not a lot of stars.  The Baader filters had much more noise and there were many many more stars in the field of view.  The result was that when the three images were channel combined, most of the stars were yellow green as well as the noise.  Try as I may, I could not tame this at all.
When I used the O3 data as a Luminance layer with a typical LRGB type combination, all the color casts were gone as well as the noise.  It was like night and day.
Tom
Tom,

Nice work.  I like the blue.  

How are you balancing your channels?  Are you using Pixinsight or something else?


Tom Blahovici
 

Hi
I am using PixInsight.  Here is the basic workflow I used:
Calibration and Integration of the SHO channels.
Linear Fit of the channels to the O3 channel.
Histogram transformation from the unlinked Screen transfer function.  Looked horrible.  Very unbalanced.

Next, preparation of the Luminance channel which is the O3 channel which had the lowest noise and best signal levels.
TGV Denoise and MMT noise reduction on the O3 channel
Histogram transformation from the unlinked Screen transfer function.
Local histogram equalization with a mask protecting the background allowing manipulation of the nebula.
LRGB combination.
At that point everything was well balanced, almost no noise and very close to what you see.
Tom


Chris White
 

Tom,

Personally, I am not a fan of linear fit tool. It's not really calibrating channels so much as matching a pixel signal profile. This can result in some strange noise issues and in some ways matching a stronger to a weaker signal channel doesn't make a lot of sense in the scheme of SNR. 

I recommend you give photometric color calibration a try. It has narrowband calibration capability where you input your bandwidth and bandpoint.  Essentially, you create an SHO as an RGB image and then run PCC tool after defining your parameters.  Then you cam choose to process this calibrated image as a straight SHO or split the channels again if you prefer custom blending with pixelmath. I'd be curious if this process gives you a better experience for non-linear processing. 


Tom Blahovici
 

Hi I typically use PCC on RGB channels. I was not aware of the narrowband capabilities. I'll give it a try. Thanks


Frank Widmann
 

Nice image. Russ Croman’s NoiseXterminator is a reasonably priced plug in for PI that works in both linear and non-linear modes. His StarXterminator is also very useful. Once I have an RGB image I use StarXterminator to remove the stars and place them in a separate star image. That allows me to work on the nebulosity without corrupting the stars. I use use PhotometricColorCalibration to get reasonable star colors, and after additional tweaking I add the star image back in using PixelMath. PI has a useful script called CorrectMagentaStars that can help get the magenta tinge out of nebulosity. I go non-linear with my color masters and do a lot of processing before I combine them. PI’s Statistics tool allows me to compare the average brightness of the images. I balance them using HT or ExponentialTransformation.

Frank