Startools

[Graeme1858]

A question for the Startools experts.

When I use Startools for processing the fits file created by DSS it takes forever to do anything! I just opened an 843Mb file and that took 5 minutes. The first thing I did was crop the image, to speed things up, after 7 minutes of waiting for that to happen I shut it down.

I'm working on my laptop, it has an i7 processor at 2.5GHz and 16Gb of RAM. Is my laptop spec insufficient? The fits file too big? Or something else?

Regards

Graeme

[Juno16]

What! 843 mb! Dang, that is gigantic!

The fits file created in dss with my settings are about 280mb and I thought that was large.

Are you using drizzle in Dss? If so, try without it. Just an idea.

[bobharmony]

That is a huge file! My experience with DSS stacks that put out fits files are about 200M. If your imaging scale is a lot smaller than your seeing, I would suggest that the first action you take in StarTools is to bin the file. That will make subsequent processes run faster. If your image scale is much smaller that your seeing, it will not reduce the amount of detail you can pull out of the image, and may make the Deconvolution module work better. My image scale is about 1 arcsecond per pixel, and my usually seeing limit is 2 to 4 arcseconds, so binning is a no brainer for me.

And as Jim says, don't drizzle a full image!

Bob

[JayTee]

Why is your output image from DSS in the .fits format? I always output to a .tif file. Just as useable and a lot smaller!

Cheers,
JT

[bobharmony]

Why is your output image from DSS in the .fits format? I always output to a .tif file. Just as useable and a lot smaller!

Cheers,
JT

JT - StarTools will only open a TIFF file if it is not compressed. It will open FITS files as they come out of DSS without and interim processing needed.

Bob

[KingClinton]

Always bin first!

Once DSS has finished stacking and has saved your file, click on the "Save picture to file" under the processing tab.
Select the output as TIFF(16 bit/ch) and check the box for "Embed adjustments in the saved image but do not apply them".

Startools will now happily open that file, remember to bin and then you ahould be fine.
Let startools select the binning amount, I usually open the binning module, let star tools think about it and when done click keep.

[Graeme1858]

Ok, I'll try again!

1. Turn off drizzle
2. Bin first
3. Save As Tiff

I'll be back soon.

Thanks for your help everyone.

Regards

Graeme

[Graeme1858]

I followed the advice above and used Jim's workflow from this post: viewtopic.php?f=66&t=8250&p=71292#p71292

Here's the result:



Maybe it would work better if I had some flat frames.

Regards

Graeme

[KingClinton]

Once you have binned do a minor crop to get rid of any stacking artifacts.
I usually do a wipe next and then start to develop.

Everything in your image can be processed out.
Flat and dark frames make a big difference if you have them.

[Juno16]

Hey Graeme,

But, how was your file size in dss?

Easch image has to be approached a bit differently in Startools. Like Clinton said, bin and the crop to remove stacking artifacts. I always use Autodev to see the artifacts. When you do the second Autodev, you select global restretch and the original Autodev goes away.
Three things that might help for sure. In Wipe, try the Vignetting preset. When you do the second Autodev, use the mouse and select just a small slice of the galaxy (or whatever your target is. Play with that, it makes a huge difference! After selecting a slice (region of interest - ROI) play with the Ignore Fine Detail slider. It really does well removing background noise.

Good luck and lets see what you come up with!

[Graeme1858]

The image is composed of 10 x 27Mb lights, 10 x darks and 20 x bias. I took the drizzle off and the compiled file reduced to 150Mb.

I suppose I just need to keep playing with it!

Regards

Graeme

[Gordon]

Hey Graeme,
Why don't you attach the stacked image and let us have a shot at it...... We can include the steps we used to process the final results.

[Graeme1858]

Hey Graeme,
Why don't you attach the stacked image and let us have a shot at it...... We can include the steps we used to process the final results.

Hello Gordon

Sorry I missed your reply!

Stacked fits for your consideration.

https://1drv.ms/u/s!AtzUKEWck9gpveIjtuo ... w?e=mJDCxB

I am having a go at this now with Startools to see what I get. It would be great to see what someone who knows what they're doing can do!

Regards

Graeme

[Gordon]

Hi Graeme,

Sorry I got busy today and didn't have much time. I did this in StarTools version 1.6 in about 15 minutes. The about 5 minutes in Photoshop just to do some final touches. I will go over details tomorrow when I have more time.....
You might want to check because it doesn't appear that the image you attached and the image you were attempting to process are the same one.....

[Graeme1858]

Ah, yes, not sure how that happened! The image I processed in Startools and posted was M106. The link I posted was to the data I posted was M63. Not sure how that happened but M63 is the one I spent time capturing 25 images. M106 was just a quick grab single frame.

Thanks for doing that but no need to spend any more time on it. But if you could list the Star tools steps you carried out that would be great.

Regards

Graeme

[startoolsastro]

Hi Graeme,

I had a quick look at your M63 dataset, but it appears it was not calibrated with flats. Your dataset would also greatly benefit from dithering between frames, as there is a fair amount of banding and walking noise visible.

Flats are really not optional - your dataset shows a great number of dust dounts, as well as vignetting. This makes it impossible to tease out faint detail for algorithms and humans alike. Lastly there are also some changes to the DSS settings you can make that should help achieve a more optimal dataset Have a look here for the most optimal DSS settings for StarTools, and for some important tips on how to create great datasets.

The thing is, that with a clean dataset, you will notice that workflows are infinitely more replicable. With a compromised dataset, however, you will spend most of your time trying to work around easily avoidable issues. This means that you will not really learn something that will be useful going forward, which would be a shame. Good (not necessarily meaning noise-free or deep), well-calibrated datasets will make your life so much easier and will yield so much better results!

Hope this helps and wishing you good health & clear skies!

[Graeme1858]

I had a quick look at your M63 dataset, but it appears it was not calibrated with flats. Your dataset would also greatly benefit from dithering between frames, as there is a fair amount of banding and walking noise visible.

Hello Ivo

Thanks for taking the time to have a look.

My flats problem is now sorted as of last night! I made a light box a few weeks ago but it didn't work on the night I captured the M63 images. It's working fine now. (dc supply issue!) I'm looking forward to using it each time I'm out capturing now. I was guiding with PHD and I thought it was dithering, I'll look into that.

I'll have a look at your DSS settings for StarTools suggestions, thanks for that. I generally just go with the DSS recommendations.

I've just processed a stack (flats, darks and bias and all!) of images of Comet C 2019 Y4 Atlas from last night. The comet option 3 did quite a good job. Then I stretched it and played with the colour balance in Gimp. (Here viewtopic.php?p=72786#p72786) I'll have a look at your link and then give it a go in StarTools.

Thanks for your help.

Regards

Graeme

[Juno16]

Lastly there are also some changes to the DSS settings you can make that should help achieve a more optimal dataset Have a look here for the most optimal DSS settings for StarTools, and for some important tips on how to create great datasets.

Hope this helps and wishing you good health & clear skies!

Thanks a bunch Ivo for this link!