Wide angle Milkyway DSS settings - Problems

[johneta]

PROBLEM RESOLVED FURTHER DOWN THIS PAGE - IN SHORT DISABLE EVERYTHING IN COSMETIC TAB IN STACKING PARAMETERS.

 

Can anyone help me with settings for DSS on wide angle images.

I have stacked with DSS successfully with scope focal lengths of 800-1000mm for a while now.

But I want to do a few wide angle milky way landscapes, but I find that DSS stacks them poorly.
Maybe I've got the settings wrong.
Does anyone have suggestions.
I have included below a manual stack via Photoshop of 10 images, then a stack of 4 images on DSS.  (No significant difference with a 12 image DSS stack I also did)
Then I show a way zoomed in crop of each image.

Images taken with a Canon 1200d (Rebel T5i) and a stock 18-55.  lens set to 25mm.  Single subs are reasonably good quality.

You can see that the DSS looks OKish at a distance but as you zoom in it seems to make a total mess of the stars.
They are all mushed together and look a bit cartoon like. The PS sack still retains the stars as reasonably sharp and as individual stars, and looks similar to the single subs I had.

Anyone had success with DSS and wide shots? An example of a good quality stack from DSS is here-   https://youtu.be/yACQ3x5dTko   on an Astrobackyard tutorial. You can see his stack is quite smooth as he zooms in, - unlike mine.

Thanks John

Edited November 22, 2019 by johneta

[happy-kat]

Hi

After registering review and make one file the reference frame. I would try stacking using kappa sigma clipping on the light frames. See how that comes out.

[johneta]

OK thanks. I'll try that and report back.

[johneta]

I went to try out your suggestion but went down another path.

I usually never look at the subs in the image subs list on DSS - as I have inspected them closely on APT or windows Photo Viewer before I load them.
I decided to have a close look at the images in the DSS image list. I found that the individual images already looked smudged and messy before I even started stacking!!
I usually load RAW's  (Canon CR2 files) to DSS thinking they will be better quality. But maybe the RAW converter in DSS is poor quality.

I cant believe I haven't noticed this before.  So I converted the images in Photoshop from CR2 to 16bit Tiff before loading them into DSS.
And they looked virtually identical to the original RAW images! (looking at them in DSS)

So, I stacked the 4 x CR2's , then I stacked the same 4 images that had been pre-converted to Tiffs before loading them into DSS.

The Tiffs came out a lot better. This really surprises me.  I haven't noticed other people mention this before.
I'm either missing something simple or this is a really important issue that people should be aware of with DSS.

Below is:  First the Tiff stack, then the CR2 stack. Then very cropped in images of them both,-First Tiff stack, then CR2 stack.
Although pixel scale is poor (stars consist of only a few pixels due to wide angle lens) you can see there is a lot sharper detail in the Tiff images.

[johneta]

Update.

I checked the options, Raw/FITS DDP Settings on DSS and found that made a difference to how DSS converted RAW's to FITS (DSS uses FITS while stacking I think??)

I tried the settings out under 'Bayer Matrix Transformation' section and they did make a difference.

But none of them were as good as converting RAW to Tiff in Photoshop first.
So, I'd still recommend this.

Anyone else got an opinion on this.

 

[happy-kat]

What version of DSS are you using?

[johneta]

DSS 4.1.1

[happy-kat]

I am surprised by the difference in your images and feel there is something missing in explanation.

This a CR2 Canon 1100d on the left and the same file in DSS 4.2.3 the stars are the same in clarity given each is guesstimating the RAW view

Edited November 19, 2019 by happy-kat

[johneta]

Yeah, not sure what to say.  I have always been a little dissapointed using DSS and have noticed this lack of clarity, but do my best to work around it.
Also I'm Having a wee bit of trouble comparing the 2 images since they look a bit different from each other.

I trialed APP and Pixinsight a while ago and noticed an overall improvement in the stacking clarity amongst other things. I intend to buy Pixinsight at some stage.

So I would like to understand what is going on in DSS because I haven't seen other people discuss it 🤔

[johneta]

Also, what type of lens/ Telescope is that photo taken with?
I checked and the effect is much less pronounced on telescope type focal lengths (around 800mm in my case).

Still there, but is worse on wide angle shots say 20mm.

[happy-kat]

That particular cr2 was from the 1100d and a 24mm lens.

Just as a trial try a stack using Sequator (it's free).

[happy-kat]

Could you share the four cr2 files then I could have a go at stacking them as an idea if you like

[johneta]

Sure. Will do tonight

 

[johneta]

Thanks for your interest, I appreciate it 🙂

Hey whats the best way to share the images with you. Is there a mechanism on SGL?
Also I'm not sure how much use it would be as the images I look at in the DSS image list before they are stacked have already been deteriorated just by bringing them into DSS.

So its understandable that the stack output would also be poor.

I will do a stack in Sequator as you suggested-but I also did a proper stack in DSS of 50 of these images (Converting them to Tiff first in PS)and a process in PS and I'm happy with the result. It's as I would expect.

[johneta]

I did a stack in Sequator and it looks fine. A little different than the stack I did in PS but similar and good quality as you would expect.

[johneta]

Problem solved !

I had ->Cosmetic Tab  -  Detect and Clean Remaining Hot pixels ticked and set to 1px and 50% threshold.

I think I have always had this ticked 🙄. I think I thought it made very little difference ages ago and I have just stuck with it on.

So I un-ticked it and the stack just looks normal, None of that fuzzy stuff. I stumbled across this on another thread.
I'm surprised it makes such a big difference!

Oh well I would recommend to always disable everything in the cosmetic tab.

[happy-kat]

That's great you have that sorted now, thanks for sharing your findings.

[johneta]

Thanks again for your interest. It helps to keeping pushing on to find an answer.

Hey I also had a look at your Barn Door thread you made a while ago. Looks cool. Ive got one of those too (wee bit different).

Its amazing what you can shoot with those.

[happy-kat]

Cheers.

I process using DSS and StarTools, the latter best results are from least set DSS and I hadn't considered the other cosmetic options as I don't use them (through I'm going to double check now).

The barn door I'm very pleased with it was enjoyable to construct and has worked very well, I'm really pleased with it's accuracy when using the 40mm lens, I'm hoping when eventually I add the Arduino board and motor I'll get similar use but with longer lens even if it's just the 85mm. Why not share an image of your barn door in the DIY section, I think they're great.

[johneta]

Yeah I may do. Mines got a curved threaded rod and a $2 stepper motor and Arduino driving it.

Im just starting to use it again, hoping to get some wide landscapes with around 18mm - 25mm lens

[dph1nm]

On 22/11/2019 at 06:42, johneta said:

I had ->Cosmetic Tab  -  Detect and Clean Remaining Hot pixels ticked and set to 1px and 50% threshold.

This is because stars are essentially single pixels at this resolution - so DSS tends to think they are hot pixels! It is possible to get caught by this on shots at much longer focal length as well, if the seeing is good. I have had the central pixels of stars removed at 1200mm focal length! For this reason I am very wary of the cosmetic cleaning, but if you turn it off all together I find that some real hot pixels get through.

NigelM

[johneta]

Oh thanks. Nice explanation. That makes sense re wide and long FL. I will let those hot pixels through in preference to the other artifacts.

Regards