Arrrrggghhhh... DSS not stacking.

[StargeezerTim]

Not for the first time DSS is causing me problems. I got 32 subs like this one.

They are taken with the Star 71 on the 100D and I used a CLS filter.

I have stacked now 20+ times and keep getting results similar to this:-

 

I have tried all the stacking methods, tried with the Median filter on and off, with the hot pixel settings on and off, without Flats, without Bias (dithered so no darks used).

Anyone have any ideas of what to do? I've also tried different star threshold  settings and tried getting rid of all DSS files in the folder and staring again. Also tried stacking just 10 subs...If I had hair, I would be pulling it out!

 

[wxsatuser]

Looks the the bottom half is stacking on something other than stars, may be a hot pixel.

Try reducing the number of stars it's stacking on, have noise reduction checked as well.

[tooth_dr]

2 hours ago, StargeezerTim said:

Not for the first time DSS is causing me problems. I got 32 subs like this one.

They are taken with the Star 71 on the 100D and I used a CLS filter.

I have stacked now 20+ times and keep getting results similar to this:-

 

I have tried all the stacking methods, tried with the Median filter on and off, with the hot pixel settings on and off, without Flats, without Bias (dithered so no darks used).

Anyone have any ideas of what to do? I've also tried different star threshold  settings and tried getting rid of all DSS files in the folder and staring again. Also tried stacking just 10 subs...If I had hair, I would be pulling it out!

 

Happened to me too Tim, same image too! Median filter worked for me but I guess no luck here. Why don’t you give APP free trial a whirl at stacking them if you are out of ideas?

 

[happy-kat]

I had weird problems earlier in the week turned out I had two odd frames causing the issue, removed them all was fine.

[Shibby]

Another thing to try is picking a different reference frame - a good scoring one near the middle of your dataset.

[StargeezerTim]

3 hours ago, tooth_dr said:

Happened to me too Tim, same image too! Median filter worked for me but I guess no luck here. Why don’t you give APP free trial a whirl at stacking them if you are out of ideas?

 

Must be a shy bit of sky that doesn't like being photographed! So frustrating. Was up to 5.30am trying to sort it and started again soon as I got up! Need to get a life 

[carastro]

There was someone else on this forum with the same problem (unless it was you first time around).  If not you might be worth searching for it and see whether there was a conclusion by the end of the thread.

Carole 

[carastro]

Oh dear, just found it, it was you.

Have you tried using a different stacking programme?

Carole 

[StargeezerTim]

Thanks for your help everyone. I did manage to get an image, minus flats, with the Median stacking method. When I added flats, it reverted to type, when I took em out, it still played up! So I think the programme is unstable. I managed to apply the flat in PS using 'apply image', but discovered whilst doing it that the pixel image size of the stacks (with bias) and the flat differed slightly for some reason. Wonder if this is the thing that caused all the probs....

[Martin Meredith]

Looking at the bottom part of the display, I wonder which registration transformation you are using? I can't see how you'd get those curved edges with either a translate only or translate + rotate model (or even translate+rotate+scale). For a more general (unconstrained) model that tries to correct for lens distortions I can see that the warping of individual subs might end up looking curved though.

Martin

[Chris-A]

I've seen this a few times, it was due to having too low a threshold on the star selection, ie. too many stars used for alignment. I usually adjust the slider and aim for approx 50 stars.

[tooth_dr]

14 minutes ago, Chris-A said:

I've seen this a few times, it was due to having too low a threshold on the star selection, ie. too many stars used for alignment. I usually adjust the slider and aim for approx 50 stars.

You are definitely right. In fact I had to use a median filter as there was so much noise in the subs it couldn’t get low enough. 

[StargeezerTim]

3 hours ago, tooth_dr said:

You are definitely right. In fact I had to use a median filter as there was so much noise in the subs it couldn’t get low enough. 

I don't think its just that... I adjusted sensitivity to the point where is found no stars, moved back one percent (the highest % and lowest stars) and it still played up. Also, exactly the same settings on different occasions gives inconsistent results... think I'll reinstall and see if that helps...

[tooth_dr]

10 minutes ago, StargeezerTim said:

I don't think its just that... I adjusted sensitivity to the point where is found no stars, moved back one percent (the highest % and lowest stars) and it still played up. Also, exactly the same settings on different occasions gives inconsistent results... think I'll reinstall and see if that helps...

did you try that median filter?  I agree that DSS is very inconsistent

[StargeezerTim]

Yes i did. I ended up going about it quite methodically, stacking 5 lights, when i stacked five i added until i stacked all lights. Then i added bias and all working well. At this the settings were median stacking, median filter, no hot pixels. For the final try I added the master flat and back to square one! I then removed the flat but it wouldn't work, despite working just before! It seems to work less consistently on brighter and less contrasty subs. I don't recall ever having a problem when using a uhc filter. 

Cest la vie...  

[Thalestris24]

Hi

I would try kappa-sigma stacking. DSS works pretty well - most of the time but I think a rogue sub or problem flat can throw it. Don't forget - if you change flats/calibration frames you have to re-register everything. Hope you can sort it.

Louise