Looking for advice on image processing [Deep Sky] Processing techniques · Andy Wray · ... · 19 · 738 · 9

andymw 11.01
...
· 
I'm really struggling to acquire and process decent images.  I know it's not an equipment issue and that it's my expertise at processing the issues downstream.  Below is my latest example where I've captured 5 hours of LRGB on M106.  I'd welcome any recommendations you may have (however critical) on how I could handle this better.

To my eye:  the colour calibration looks wrong, the stars are bloated and elongated, it generally looks unnatural and blotchy.  Apart from that it seems just fine ;) 

Image1150pc.png



M106 after 5 hrs integration
Edited ...
Like
URmyArtemis 0.00
...
· 
I would like to first ask a question: how many subs do you have for each LRGB and how did you do the RGB combination in pix?
Like
andreatax 7.22
...
· 
·  2 likes
Well, it is hard to comment without having an idea of what the source material is like. 5 hours isn't much to start with, at that image scale anyway. A few things of note:

1. Your background is far from being uniform so you should start with that.

2. You stars look like bloated triangles (with "hair" on one side, too, if you see what I mean), you should look into collimation and positioning of the CC and camera with respect with each other.

3. The whole image looks mushy because you went in heavy with NR. The rest descends from there. Less flattening of the image would allow you to bring out a bit more colour and noise is still the hallmark of a real image.
Like
kuechlew 7.75
...
· 
Better than anything I ever produced on my own Andy ...

I don't understand the colour management issues. You are imaging with a mono camera in LRGB, so you should have full control over the colours. Just mix your R, G and B channels to your taste. With a bit more colour in the stars and in the galaxies and some deconvolution this will be  a very nice image in my eyes.

Only slight criticism: it's a bit visually unbalanced to my taste. While the galaxy is placed in the golden ratio there is something missing in the upper right part of the image to counterbalance the "left heavy" image. There is just nothing of visual interest in the right half of the image. Did you artificially remove the spikes of your Newtonian in the image? Interestingly some spikes in the stars on the right side could just provide the missing visual weight.

I would crop something like this:
image.png

or like this (but losing the small critter on the bottom is a bit of a shame ...):
image.png

It would take a bit more time than I invested to figure out the best crop ...

Clear skies
Wolfgang
Like
OregonAstronomer 2.81
...
· 
Andy,

You are being way too hard on yourself. You have an image there that most of us would be proud to display on our wall or website! You are 90% of the way there. The last 10% is tweaks like what Andrea mentioned above. Sometimes those tweaks, though, will consume most of your post-processing time.

Do you use Photoshop? Neil Carboni has some excellent PS actions that I find incredibly helpful. For your image above I would use the star reducing action and the Local Contrast Enhancement action to bring out detail in the center of the galaxy. I'm sure other programs have there own versions of these actions.

Congrats on a great image!

Arnie
Like
andymw 11.01
...
· 
·  1 like
Thanks for the feedback folks.  I've had a go at re-processing it as follows:

* Not been as heavy-handed with the noise reduction
* Removed the frames from a full-moon night as they had really bad gradients that were affecting my background
* I couldn't undo the collimation issues with the stars, so used a little bit of convolution to smudge out the spikes
* Tried to get some colour back in the stars by not blowing them out
* Pushed the saturation on the galaxy

FWIW:  I just processed RGB this time:  70x90s Red, 44x90s Green, 55x90s Blue (approx 4 1/4 hrs overall integration time).  Gain 139, Offset 10.  ASI1600MM Pro

Image22_50pc.png
Edited ...
Like
andymw 11.01
...
· 
I would like to first ask a question: how many subs do you have for each LRGB and how did you do the RGB combination in pix?

See my updated picture above this post for number of subs.  I would usually do:

* DynamicCrop on all channels
* DynamicBackgroundExtraction on all channels
* A linear fit on the Red, Green and Blue channels
* Use ChannelCombination to produce an RGB composite and then stretch the RGB using a variety of the stretching tools
* I would then work on the Luminance channel to get that where I like it and then use LRGBCombination to add the luminance back into the RGB image

For my new one just above this post I took a different approach by stretching the individual colour channels first before doing the channelcombination.  I also haven't had time to do the luminance work yet, so it is just the RGB channels.
Edited ...
Like
andreatax 7.22
...
· 
Well done, Andy.
Like
andymw 11.01
...
· 
That's so beautiful


[Deleted] as I responded to wrong post
Edited ...
Like
URmyArtemis 0.00
...
· 
Hi Andy,
I have a different method in taking pictures.
I usually would do 10 subs for each RGB and then L. I usually take as many L as I could. L contribute to the main details of a picture. I know that do a LRB subs at certain ratio will work, But probably normally people take more L then RGB subs. The subs ratio for LRGB is like 3(or4):1:1:1.
Like
andymw 11.01
...
· 
·  1 like
OK, re-collimated my OTA and took another 1.5 hrs exposure.  This time I didn't fiddle with the stars as they have nowhere near the amount of "hairs" as they had before.  I still need to get my secondary re-centered on my focussing tube which I will do tomorrow whilst the skies are cloudy.

Image291b50pc.png
Edited ...
Like
AwesomeAstro 2.39
...
· 
It's really not bad.

Amounts of noise reduction, color balance, and contrast are usually matters of preference, and I've found this to be a challenging target to get exactly right.

The only thing worth mentioning that's equipment/data collection related, would be the triangular stars. This was mentioned above, however in my opinion, this doesn't look like coma, and is likely not astigmatism either. Triangular stars are usually produced from "pinched optics" in which you have uneven pressure on one of the optical surfaces (often the primary mirror). Check the entire scope and make sure nothing is "caught up" or has an opportunity to rub on something; search for this term and you'll find image that have stars just like that...
Like
andymw 11.01
...
· 
The only thing worth mentioning that's equipment/data collection related, would be the triangular stars.


I'm a bit confused about the triangular stars thing  (in relation to my most recent image).  Below is a mosaic produced by the abberation inspector script on the above image.  To me, the stars may be a bit oval, but are not triangular.  What am I missing?
abberations.png
Like
AwesomeAstro 2.39
...
· 
It's a subtle thing, but to me there's clearly a triagonal nature to your star aberrations. Take the panel just above the middle one, for example:

aberations raw2.jpg
I'm definitely seeing a three-sided symmetry here, particularly with the edges being straight, and all stars showing the same pattern, which is helpful in identifying the aberration:

aberations_annotated2.jpg

Course, if you don't really notice it, then you don't have to go too far into finding solutions!
Like
AwesomeAstro 2.39
...
· 
·  1 like
Just to add a bit more on this (it got me thinking), I remember I actually have experienced this sort of aberration before on my SCT (I've never used a newt or a refractor). For me, it would go away after 30-60 minutes or so, meaning it was likely temperature fluctuations that cause uneven strain on the optics, but it eventually "settles" in. 

Once you're able to see what we're referring to in your image, and if you can identify it/recognize it, check out your light frames in chronological order; do they get better, and less triangular, as the night goes on? You'll have to inspect them very closely/zoomed in. If so, this probably means it's thermal, and you should let the scope acclimate earlier to help this. If it's the same throughout, I personally suspect some sort of mechanical strain on the optics.
Like
andymw 11.01
...
· 
check out your light frames in chronological order; do they get better, and less triangular, as the night goes on?


Thank you.  I do see it now.  I think part of it is thermal as the stars do improve on individual subs as the night and scope cool down.
Like
AwesomeAstro 2.39
...
· 
Nice! Seems like you're onto something. I'd let it acclimate the necessary amount of time to help with this or take the scope outside before dark.
Like
andymw 11.01
...
· 
Nice! Seems like you're onto something. I'd let it acclimate the necessary amount of time to help with this or take the scope outside before dark.

I always take it out during daylight, however it's been about 20 degrees C plus recently during the early evening and the scope was way over 25 degrees with the direct sunlight, so probably retaining a lot of heat.  In some ways I would be better taking it out at the last minute during late Spring/Summer.
Like
AwesomeAstro 2.39
...
· 
Andy Wray:
Nice! Seems like you're onto something. I'd let it acclimate the necessary amount of time to help with this or take the scope outside before dark.

I always take it out during daylight, however it's been about 20 degrees C plus recently during the early evening and the scope was way over 25 degrees with the direct sunlight, so probably retaining a lot of heat.  In some ways I would be better taking it out at the last minute during late Spring/Summer.

True yeah don't let it overheat. Sheds/garages are possible solutions to this too!
Like
 
Register or login to create to post a reply.