DSS not stacking... star shapes ??

[Craney]

Hello there,   these are three subs taken of NGC891.

DSS cannot seem to 'see' the stars that I can  !!!   It registers about 5 to 7 on the frames and refuses to stack them.    I have done the usual 'set star detection' to 2% and tried the recomended options variants several times.

These example  subs are 2 mins @ 1600 ( one from the start, middle and end of a 41 exposure run). 

I have loaded into PS as CR2 files ( tweeked the contrast and clarity to bring out the stars a bit more ) and saved them as JPG to try to stack them again    (... still no success).

These below are the processed JPG ones, so they look a bit stretched with artefacts, but I hoped it would at least give it a go to stack them.

Is there something obvious I am missing ??  I know the galaxy is migrating across the frame due to bad balance on the mount that night, but why does DSS not see those lovely stars ... !!!     Is it their intrinsic shape which is the problem ??

Cheers for any hints.

 

 

 

Sean.

 

 

[Atreta]

Did you try to Uncheck the: "use median filter to reduce noise" button? 

 DSS refusing to stack usually happens when the stars are out of focus. 

[bobro]

The jpeg versions have a high sky brightness (plus lots of noise). When I tried to stack them DSS reported too many stars (actually noise) until I reduced the resolution ( so reducing the noise) and increased the star detection threshold to 50 to only detect actual stars. 

A good way to get feedback is to post a converted raw to a jpeg without any processing at all (Windows 10 does this for certain formats). 

[Craney]

Thanks @Atreta,   I will give it a go,  but I have tried a lot of the options and none of them seem to make an great improvement on the scores.  With the focus, if I have a bit of elongation due to drift , it may upset the programme.  ie.  it wants 'pin-points' and not 'eggs'.

 

Thanks for giving it a go @bobro.  I did some tweaking on the original CR2's just to convince myself that the stars were there.   I will convert the original frames to jpg and post them.  Cheers.

[Craney]

So, here are the orinal CR2's  converted to jpg.

 

 

The slightly red hue to the stars is due to me using a UHC filter ( the 3/4 Moon was rising in the East).

As you can see a lot darker and starless.

 

Sean.

[bobro]

Looks like @Atreta was correct. Looking at older posts trailing/focus seems to create this problem with stacking not working. I had the same stacking problem with the latest jpegs above but stacking worked fine if the image resolution was reduced to 25%, so making stars smaller and reducing the effect of trailing.

[dph1nm]

DSS does not like non-round stars. Try the manual star editor (button down the RHS when you display an image in DSS) to see if you can pick any extra up. You only need 8 stars in common between frames to stack. Otherwise you can try using superpixel mode for debayering. This reduces the image size 2x2, but does seem better at picking up distorted stars.

NigelM

[Craney]

Thanks @dph1nm.   Have tried the   "superpixel mode for debayering"  still no  success.

I read about a PS  adjustment that can shrink pixels in the x or y direction. This de-eggs the egg shaped stars..... but knowing my luck will probably crinkle the galaxy....

Ho Hum !!....   Maybe quicker to re-take the subs on the next clear night with a longer exposure...

 

Sean.

[happy-kat]

What is the FWHM range on your lights after you registered them?

[Craney]

I will have a re-re- attempt tomorrow.... and report back.

Sean.

[Craney]

Hi @happy-kat

This is the screen grab of the a non-stacked run.......with the required info...

It is the 'NC's  that worry me.   Hope this might help.

 

Sean.

[happy-kat]

You did select all the frames? They all had a tick in the box before registering?

Just loaded one of mine and all lights stacked and dx,dy and angle are all NC.

[Craney]

4 hours ago, happy-kat said:

You did select all the frames? They all had a tick in the box before registering

Oh yes....  tried top 4, best 50 % etc....      It almost immediately comes up with a box stating that only one of the frames will be stacked and would I like to continue ??

It will stack the (1+1), but thats all.

 

[alacant]

5 hours ago, Craney said:

It is the 'NC's  that worry me.

Hi. If you want to see the NC values replaced with numbers, just hit Compute offsets. It's the next item after Register checked pictures; you don't need to because it will calculate these before stacking anyway. HTH.