what am I doing wrong with DSS?

[StuartT]

I just had a go at registering and stacking a bunch of images I took last night of M13 with a 20cm SCT f/6.3

70 light frames (15 sec at ISO1600)
15 dark frames
18 bias frames
30 flat frames

I had to set the star detection threshold right at the bottom end (2%) to find enough stars (it suddenly went from finding none to finding 450)

But the final autosave.TIF looks purple for some reason. Here's a screenshot of it in Photoshop. I also attach one of the light frames so you can see I did actually capture something

 

Edited May 31, 2021 by StuartT

[geordie85]

You'll just need to align the histogram in Photoshop

[StuartT]

13 minutes ago, geordie85 said:

You'll just need to align the histogram in Photoshop

do you mean the levels histogram? It looks like this at the moment. What should I fiddle with?

and where's all that purple come from?

Edited May 31, 2021 by StuartT

[Thorney]

What I normally do is In the levels histogram you can select the colour channels individually instead of RGB. So go into the red channel and move the left hand slider over to just before the data then do the same with green and blue. The data should be roughly together and will balance the colour out. 

[skyhog]

What was the capture file format?, I'm assuming raw? 

[geordie85]

1 hour ago, StuartT said:

do you mean the levels histogram? It looks like this at the moment. What should I fiddle with?

and where's all that purple come from?

In the channel drop down box change that to red, move the left hand slider to a tad before the black. Do the same for green and blue too so they all at the same point.

The purple is probably coming from the red and blue channels being too far over to the right. Once they're aligned you should have the correct colour balance.

[vlaiv]

DSS has background calibration feature and depending on what you select for it - it will produce different background.

In any case, having purple background is no issue if you set black point properly for your image.

You can do that by selecting each of R, G and B channels respectively and setting left levels slider just at the base of histogram on left side, like this:

[StuartT]

 

Thanks everyone. I'll try this and report back.

44 minutes ago, skyhog said:

What was the capture file format?, I'm assuming raw? 

Yes, RAW images 

[StuartT]

Ok, so this is how it looks after adjusting the left hand arrow to the start of the data in each channel separately.

Definitely better, but I am not sure if I have captured quite enough light in my original light frames, maybe?

I limited my exposures to 15 sec as I am only using an EQ6R Pro with no guiding and was concerned about getting trails. Could I afford to take longer exposures do you think?

[vlaiv]

2 minutes ago, StuartT said:

Ok, so this is how it looks after adjusting the left hand arrow to the start of the data in each channel separately.

Definitely better, but I am not sure if I have captured quite enough light in my original light frames, maybe?

I limited my exposures to 15 sec as I am only using an EQ6R Pro with no guiding and was concerned about getting trails. Could I afford to take longer exposures do you think?

It looks like you have some passing high altitude clouds in one or several of your subs. You should inspect them. Another possibility is that DSS messed up calibration somehow (it sometimes does that - and that it clipped your red channel - leaving patches of red light around the frame - this happened to me several times).

You should also try stretching your data to see what is really there. When you stack - you are creating high dynamic range data unlike single sub - and you should stretch it to show all the details.

As far as using longer exposure on your EQ6R Pro - that depends on the mount. Not all mounts behave the same although they are built to same specs - manufacturing errors create variance between specimens. Maybe your mount is good enough that it can handle 30s exposures with only minimal frame rejection - maybe it can do only 15s. You won't know until you try.

Just be aware - longer you go - you'll need to discard larger percentage of frames - but image in given time (total time of subs you kept) will be better.

[StuartT]

Thanks Vlaiv

Here it is after some further fiddling with levels and gamma. But I will do as you suggest and check the lights for clouds

(for some reason, Photoshop won't let me save the TIFF as a JPG, so in order to post here I had to screen grab it with snipping tool!)

 

Edited May 31, 2021 by StuartT

[StuartT]

I just discovered that I had a full set of JPG copies in my lights folder which I uploaded to DSS!! My Canon DSLR shoots a JPG copy of every RAW for some reason and it looks like I forgot to delete them! I'm going to start the whole process again 

[vlaiv]

2 minutes ago, StuartT said:

I just discovered that I had a full set of JPG copies in my lights folder which I uploaded to DSS!! My Canon DSLR shoots a JPG copy of every RAW for some reason and it looks like I forgot to delete them! I'm going to start the whole process again 

That will mess up a bit things in DSS.

You have option to tell your Canon what to save. It is somewhere in menus, on my Canon, I have three different options:

- Jpeg only

- Jpeg+Raw

- Raw only

It can be either like this:

Or like this:

Where you select quality / size and if you want to shoot it for each raw and jpeg

[StuartT]

Ok, so I deleted all those pesky JPGs and re-ran DSS. Now the autosave.TIF is no longer purple. But it looks like there is nothing to move the left hand arrow to the data. It's already there. So what do I do now?

 

 

[vlaiv]

12 minutes ago, StuartT said:

It's already there. So what do I do now?

Do simple three step stretch of your data

1. Step 1 - use right level slider and bring it down until you start to saturate parts of your target - then back it off a bit. You want to move it down as much as you can without saturating core of your target. Apply change

2. Second step - use middle slider in levels and bring it left / down again until you are happy with how your target is exposed. In this step don't worry if background gets too bright. Just don't push your data too hard for noise to be obvious. Again apply change

3. Third step - move left slider / black point again to the foot of histogram. This time histogram will be broader and this will be easy to do.

 

[StuartT]

Thanks Vlaiv. Here is how it turned out