Feedback appreciated please - Widefield M81 et al

[Dazzyt66]

Hi,

This is my first 'real' attempt at AP. I'm using an ST80 with unmodded Canon 1100d. The image was done over 3 nights using Ekos and is around 500x30s lights, 60xbias 60xflats. The images were stacked and processed in Siril, with final touches in GiMP and Darktable. I think I've done a good job of removing the CA in post processing but would still appreciate some feedback as this is my first real attempt at doing something 'properly' 😂 I know that focus may be an issue as I did it all manually with no mask... Comments on what I can do to further improve appreciated - I'm not going to go down the ED route as I'm generally happy with what I'm able to get as is atm, but would like to be able to get the most out of it.

[AstroNebulee]

Great capture there, especially with manual focus and no batinov mask. I use deep sky stacker to stack my files and photoshop to process so can't advise on your software processing. But you've made a great start and I to am trying with no darks and the final image is a lot better as long as you take lots of light frames. I'd definitely recommend getting a batinov mask makes focusing so much easier and quite cheap to purchase. I'd also say you can crop the frame in a little as they may help you, the stars are a little eggy, so maybe cut the exposure time down or adjust your PA (what mount are you using) but still a fantastic cadprure seeing the arms of the galaxy, just keep plugging away and enjoy what you do because as soon as it becomes a chore you wont enjoy it. clear skies 👍

Edited May 2, 2021 by AstroNebulee

[Dazzyt66]

1 hour ago, AstroNebulee said:

Great capture there, especially with manual focus and no batinov mask. I use deep sky stacker to stack my files and photoshop to process so can't advise on your software processing. But you've made a great start and I to am trying with no darks and the final image is a lot better as long as you take lots of light frames. I'd definitely recommend getting a batinov mask makes focusing so much easier and quite cheap to purchase. I'd also say you can crop the frame in a little as they may help you, the stars are a little eggy, so maybe cut the exposure time down or adjust your PA (what mount are you using) but still a fantastic cadprure seeing the arms of the galaxy, just keep plugging away and enjoy what you do because as soon as it becomes a chore you wont enjoy it. clear skies 👍

Thanks for this. I’m using a HEQ5 semi permanently sited - I adjusted the PA using Ekos, but it may still be a little out so I’ll check it again next time - good tip about lowering exposure time - I may drop it to 20s and combined with a better PA may get rounder stars.  I have now got a Bahtinov mask. I’m definitely enjoying it and was pretty pleased with this image. Still a lot to learn though 🙂

Edited May 2, 2021 by Dazzyt66

[AstroNebulee]

3 hours ago, Dazzyt66 said:

Thanks for this. I’m using a HEQ5 semi permanently sited - I adjusted the PA using Ekos, but it may still be a little out so I’ll check it again next time - good tip about lowering exposure time - I may drop it to 20s and combined with a better PA may get rounder stars.

Hi

I'm not sure as I don't own this mount but once polar aligned it should give you a couple minutes maybe more exposure time and iso of 800. The focal length of your scope is 400mm I believe so should achieve a longer exposure time with the tracking the mount provides. I could be wrong though, someone will correct me if I'm wrong. 

Edited May 2, 2021 by AstroNebulee

[Clarkey]

I would agree an HEQ5 should manage to track reasonably well at 400mm. I have use one at 1000mm for a couple of minutes and that was OK. As you are not using a field flattener (I assume) the stars round the edges will always be stretched. The central stars are not too bad and with perfect polar alignment I see no reason why you could not manage 1 minutes subs. Check the mount though, as my HEQ5 did have quite a bit of backlash from new which I had to adjust out. Also the worm gears were loose and needed tightening. Both of these made a lot of difference, along with the belt modification.

[Dazzyt66]

2 hours ago, Clarkey said:

I would agree an HEQ5 should manage to track reasonably well at 400mm. I have use one at 1000mm for a couple of minutes and that was OK. As you are not using a field flattener (I assume) the stars round the edges will always be stretched. The central stars are not too bad and with perfect polar alignment I see no reason why you could not manage 1 minutes subs. Check the mount though, as my HEQ5 did have quite a bit of backlash from new which I had to adjust out. Also the worm gears were loose and needed tightening. Both of these made a lot of difference, along with the belt modification.

Thanks for this. I’ll defo check the PA again. I’ll also have to read about the mount adjustments, although as I’m in a Bortle 8 zone I don’t think subs longer than 30s or so would bring me much benefit anyway? 🤔

[AstroNebulee]

3 hours ago, Dazzyt66 said:

although as I’m in a Bortle 8 zone I don’t think subs longer than 30s or so would bring me much benefit anyway? 🤔

Eeeeek 😬, you could experiment to see, light pollution filter or light pollution removal tool for your software may be available, I know colormancer is a good free noise reduction software that I use with photoshop, whether it works on gimp is another thing, clear skies and good luck 👍

Edited May 2, 2021 by AstroNebulee

[Laurieast]

I can't see that you say what ISO setting your using? I'm in a Bortle 8 sky, and use ISO 400 or 800 and have got good results with an ST80 at 60s 120s and above, give it a go!

You could try GradientXterminator on that to even out the background.

Edited May 2, 2021 by Laurieast

[Seelive]

On 21/04/2021 at 09:52, Dazzyt66 said:

Hi,

This is my first 'real' attempt at AP. I'm using an ST80 with unmodded Canon 1100d. The image was done over 3 nights using Ekos and is around 500x30s lights, 60xbias 60xflats. The images were stacked and processed in Siril, with final touches in GiMP and Darktable. I think I've done a good job of removing the CA in post processing but would still appreciate some feedback as this is my first real attempt at doing something 'properly' 😂 I know that focus may be an issue as I did it all manually with no mask... Comments on what I can do to further improve appreciated - I'm not going to go down the ED route as I'm generally happy with what I'm able to get as is atm, but would like to be able to get the most out of it.

What stacking method did you use? A lot of the brighter stars appear to have a fainter companion so it's hard to say whether it's a problem with the individual images or with the stacking software or both. If you've not already, try a Kappa-Sigma stacking method to see if that reduces the effect of the 'doubles'?

[Dazzyt66]

2 hours ago, Laurieast said:

I can't see that you say what ISO setting your using? I'm in a Bortle 8 sky, and use ISO 400 or 800 and have got good results with an ST80 at 60s 120s and above, give it a go!

You could try GradientXterminator on that to even out the background.

Cool! I used ISO 1600 as I’d read somewhere that it was the signal to noise ‘optimal’ for my 1100d for Astro. I have used 800 too and (personally) didn’t notice that much difference so I just stuck to the 1600 😀

[rickwayne]

I think your focus is pretty good -- look at the smaller stars, hardly any doughnuts. Likewise your tracking is good enough  to pull out good detail in M82, even if it isn't technically perfect -- I see what they mean about "multiple" stars, in fact I see three lobes to some of the brighter ones.

You may be able to get better star color out of this -- or not. Check out your stacked but unstretched image with a tool that allows you to read pixel values (I don't recall but I think SiRiL will do it). Are the stars saturated before stretching? Are the RGB values different enough for some of them to indicate color? If not, they're just overexposed and that's pretty much that for this try. But if you do have color lurking in there, you can use a star-removal tool to produce a "stars-only" image and a "no stars" image, stretch and saturate them separately,  and then recombine them (Photoshop "screen"  blending mode works well for me, not sure what the equiv is in The GIMP).

I use starnet++ for this, it's a neural-net tool that the author trained to pull stars out of images. My workflow is to run starnet on the image, use "difference" blending mode in Photoshop on the original and the starless image to produce a stars-only layer, then save that and process the starless and stars images separately. You  can run  an aggressive stretch on the nebulosity to bring out details, and use a much lighter touch on stars, and whomp the saturation to egregious levels as well to bring  out the little twinkly  jewel-tones.

Of  course, on  an image like this, where the nebulosity to protect is such a concentrated and defined area, you could just use masking to apply different stretches and saturation to different parts of the image. For something where nebulosity is most of the image and it's speckled with stars all over, that way madness lies. 🙂

Really nice result for a first shot. I mean, I am super-jealous here.

[Dazzyt66]

12 minutes ago, rickwayne said:

I think your focus is pretty good -- look at the smaller stars, hardly any doughnuts. Likewise your tracking is good enough  to pull out good detail in M82, even if it isn't technically perfect -- I see what they mean about "multiple" stars, in fact I see three lobes to some of the brighter ones.

You may be able to get better star color out of this -- or not. Check out your stacked but unstretched image with a tool that allows you to read pixel values (I don't recall but I think SiRiL will do it). Are the stars saturated before stretching? Are the RGB values different enough for some of them to indicate color? If not, they're just overexposed and that's pretty much that for this try. But if you do have color lurking in there, you can use a star-removal tool to produce a "stars-only" image and a "no stars" image, stretch and saturate them separately,  and then recombine them (Photoshop "screen"  blending mode works well for me, not sure what the equiv is in The GIMP).

I use starnet++ for this, it's a neural-net tool that the author trained to pull stars out of images. My workflow is to run starnet on the image, use "difference" blending mode in Photoshop on the original and the starless image to produce a stars-only layer, then save that and process the starless and stars images separately. You  can run  an aggressive stretch on the nebulosity to bring out details, and use a much lighter touch on stars, and whomp the saturation to egregious levels as well to bring  out the little twinkly  jewel-tones.

Of  course, on  an image like this, where the nebulosity to protect is such a concentrated and defined area, you could just use masking to apply different stretches and saturation to different parts of the image. For something where nebulosity is most of the image and it's speckled with stars all over, that way madness lies. 🙂

Really nice result for a first shot. I mean, I am super-jealous here.

Really great feedback, thank you. I’m pretty sure there were star colours in there (various shades of oranges and blues) but I lost all that when I did the stretching to bring the galaxies out. So, I’ll try the suggested tweaks above and see what I come up with - I may have to re-read it a few times though cos some of it sounded like a different language! 😂

[Dazzyt66]

3 hours ago, Seelive said:

What stacking method did you use? A lot of the brighter stars appear to have a fainter companion so it's hard to say whether it's a problem with the individual images or with the stacking software or both. If you've not already, try a Kappa-Sigma stacking method to see if that reduces the effect of the 'doubles'?

I think I may have used Kappa-Sigma but if not I’ll give it a try. Thanks 😊

[Seelive]

1 hour ago, Dazzyt66 said:

... I lost all that when I did the stretching to bring the galaxies out...

'Normal' stretching generally tends to loose colour. I believe Siril has an Arcsinh stretching function which will help retain the colour whilst still bringing out the fainter details.

[rickwayne]

Sorry to be idiolinguistic! Happy to explain any obscure bits.

[Dazzyt66]

1 minute ago, rickwayne said:

Sorry to be idiolinguistic! Happy to explain any obscure bits.

It’s ok, I can work my way through it! 😂 Being new to image processing though, stuff like stretching, masks and layers still has me thinking! 🤔

[rickwayne]

Well, of course. It's a very deep discipline (ow).

May I recommend my personal bible on it? The Deep-Sky Imaging Primer by  Charles Bracken is an eminently readable and fascinating...I was going to say "introduction",  which it certainly is, but his gift is that, while requiring no prior knowledge, he pulls you along so effortlessly that you're amazingly well-grounded before you even realize it. I was reading away on various fora and it was only after I read the book that I realized I'd been a victim of Rumsfeldian "unknown unknowns" -- the bits and pieces you pick up  here and on other online fora are no substitute for a well-rounded, well-planned pedagogy.

Steve Richards's Making Every Photon Count also comes highly recommended, though I haven't read that one.

[Dazzyt66]

Made a few tweaks using some of the suggestions above (still reading up on others to try). I'm very happy with this now! Kappa Sigma stacking and Sinh stretching initially made quite a huge difference I think:

 

[AstroMuni]

26 minutes ago, Dazzyt66 said:

Made a few tweaks using some of the suggestions above (still reading up on others to try). I'm very happy with this now! Kappa Sigma stacking and Sinh stretching initially made quite a huge difference I think:

Looking good. 🙂 Did you use the Photometric calibration? Reason for asking is that some of the stars look a bit too blue.

[Dazzyt66]

48 minutes ago, AstroMuni said:

Looking good. 🙂 Did you use the Photometric calibration? Reason for asking is that some of the stars look a bit too blue.

Thanks! Yeah, I used the photometric cal in Siril - I too think they are a bit too blue but I guess I’ll get better at sorting that out as I go 🙂