Can Startools improve my processing?

[JST200]

Hi,

After a couple years (but few nights) I have managed to get my astrophotography kit to a state that I feel I can take reasonable images. I now want to improve my processing, as I am sure there is more data in my images that I currently am getting. This is where I hope that Startools comes in!

I have download the trial version and spent quite a few hours trying to process some of my images. But I either seem to end up with a black and white image or a red background image with a overexposed DSO. I feel sure it is down to my lack of understanding of Startools best practices, but I have tried following the various videos and tutorials without success.

If someone is willing to show me it is possible to improve my processing, and how they did it, I would be extremely grateful!

To that end, I have uploaded below a couple of images (M51 and m42) I processed, using a very old version of Photoshop and the Astronomy Tools and Annie's Actions add-ins, on dropbox here: http://www.dropbox.com/sh/r1m5kiukajgde ... sEsLa?dl=0
I have also put up .fts files for the images. The .fts files were created using DSS, with (as recommended) "RGB Channels Background Calibration" and "Per Channel Background Calibration" both set to "No".

My m51 image was processed using a DSS .tif file from a stack of 15 x 45s iso 1600 Raw exposures.
And my m42 image was made from blending a DSS .tif file from a stack of 15 x 120s iso 800 and a 2nd .tif file from a stack of 13 x 120s iso 400 Raw images. The iso 400 stack were only used to recover the detail in the very middle of m42. The processed images are also here on Flickr:
m51 - http://www.flickr.com/photos/139333252@ ... ed-public/
m42 - http://www.flickr.com/photos/139333252@ ... ed-public/

I realise you won't be starting from exactly the same base as I did, since I saved my stack to .tif format and I had "RGB Channels Background Calibration" and "Per Channel Background Calibration" both set to "Yes". Though I a not sure how much difference that makes?

If it's of interest; my setup consists of a Skywatcher 150P (focal length 750mm) on an driven EQ3 mount. I use a Canon EOS 600D controlled by APT (Astro Photography Tool). I auto-guide using a Phillips SPC900 and a 80mm (fl 162mm) scope controlled by PHD2 - which can give me well over an hour of accurate guiding - but my images get overexposed after 4mins because of light pollution - possibly why I have such trouble processing the files with Startools? I use Cartes du Ciels, EQMOD and the ASCOM Platform to control the scope.

So... thanks for any help anyone can give!!!

Cheers, Jim

[ecuador]

I am sure others can do much better as I went too quickly, but I think my most useful comments are about your initial data:

-There is some strong vignetting, are you using flat frames? Similar to what is true for darks, flat frames are not optional...
-There is a high noise floor, 15 x 45s is not nearly enough data. You say you can go to 4 minutes, well how about at least 2, along with more exposures? And you can drop your ISO a bit (although close to ISO 1600 is your camera's optimal read noise). If you have a lot of light pollution a CLS-type filter will help with nebulas and an IDAS-LPS or Orion Imaging will help with any target, including galaxies.
-For wide field images (so apart from galaxies where you are interested in a small part of the image) the Skywatcher 0.9x reducer/corrector will fix your stars at the periphery.

Anyway, here is a quick try:

m51-15-x-45s-iso160ST.jpg (336.25 KiB) Viewed 8446 times

It went autodev, crop, bin, wipe, ROI-autodev, contrast, hdr, decon, sharpen, color, life "less4more" and de-noise. The only non-default things I did were the autodev where I selected the galaxy, the color where I made an "auto" mask to select the stars, used that to "sample" and then returned to a full mask to apply the color and finally in the de-noise I had to increase the color detail loss and read noise compensation.

[JST200]

Hi Ecuador,

Thanks for such a swift reply! And thank you for your valuable insights.
You have given me hope that it really is my lack of understanding of Startools that is the problem!

I only used light files for m51. So no darks or flats. Partly because I thought I read somewhere (possibly the APT documentation) that the author thought darks weren't necessary if you had more than about 15 lights and they were taken with dithering enabled. The m42 image did have darks.

I intend to take far more images in future. To some degree I was self limited by not wanting to subject my "photography" camera to the rigours of astrophotography. This is now solved as my family have just bought me, a as a birthday present, a second-hand EOS 1100D specifically for the purpose. Also, on the night I took m42 the moon was up, so I was limited to about 45 secs before the images were starting to overexpose. So in the future I'll try longer and more images on darker nights.

Investing in a the light pollution filters you suggest will have to be a long term goal. My aim is to see how far I can push without spending a fortune (each filter would cost more then the EOS 1100D!). Hence my use of the Philipps SPC900 (which a friend gave me as he had no use for it) as a guide camera. But it is something I will bear in mind. Though I do wonder if the current trend, near my house, to replace the old sodium street lights with modern LED based ones means the filters will be less effective?

Finally, I'll try following your process tomorrow to see if I can get a similar result. I can then work from there. Just one question. Did you select "Not bayered or is white balanced" or "bayered not white balanced"?

Anyway thanks, again, very much.

Cheers, Jim

[ecuador]

If you start as recommended, i.e. directly from the DSS stack, you always select Linear, was debayered, was not white-balanced, which (strangely IMHO) is the second button.
I guess the more light frames you get, the less the benefit of the dark frames, but a DSLR like yours (or mine, or most) has significant thermal noise, so I do see a benefit of dark frames even with dozens of frames. There are some people who claim that "relatively modern" (usually defined as post 2008-2009) Canon DSLRs have something called "dark current suppression" and they don't need darks because they suppress noise, but both my 600D and 550D exhibit plenty of noise which increases with exposure time and temperature.
You don't have to get darks on each session, you can build a "dark library". Just download the free program called Dark Master, you point it to a folder with light frames, to another folder with your dark library and it finds the darks that have the same exposure and sensor temperature as your lights and produces a dss file.
Yes, LEDs are not good for filtering, so if the entire city changes to those and light pollution on your sky is from white light then the broadband filters won't help. The narrowband (Ha, S-II, O-III etc) will still help of course with nebulae.
I guess the biggest improvement you can get is if you mod your camera to be sensitive to the red Ha/SII that many nebulae emit.

[JST200]

Thanks for the pointer to Dark Master; another piece of software to add to my growing list!

I agree about the Darks. I also tried setting EOS Custom Function Long exp. noise reduction to on. But all that did was effectively double the set exposure time (50% light 50% dark), so I might as well take my own and form the library as you suggest.

I still haven't managed to replicate what you achieved. I'm getting closer, but I have yet to convince myself that it achieves better results than my old process. I'll keep trying though.

Thanks for your help.

Cheers, Jim

[ecuador]

Canon's long exposure noise reduction is useful if you shoot something like a nightscape which you won't stack in DSS. But for regular astrophotos it is obviously not a good idea (you lose precious time, only stack with 1 dark, cannot do correct calibration with flats later etc).
In my experience I can get a result that I like with StarTools faster than other software. Also, in some images that are very hard to deal with noise-wise with other software, it can give some great results. The one drawback for me is that I sometimes get rather "pastel" looking colors, which I can't seem to tweak exactly as I like in the color module.

[Burly]

]

If you start as recommended, i.e. directly from the DSS stack, you always select Linear, was debayered, was not white-balanced, which (strangely IMHO) is the second button.
I guess the more light frames you get, the less the benefit of the dark frames, but a DSLR like yours (or mine, or most) has significant thermal noise, so I do see a benefit of dark frames even with dozens of frames. There are some people who claim that "relatively modern" (usually defined as post 2008-2009) Canon DSLRs have something called "dark current suppression" and they don't need darks because they suppress noise, but both my 600D and 550D exhibit plenty of noise which increases with exposure time and temperature.
You don't have to get darks on each session, you can build a "dark library". Just download the free program called Dark Master, you point it to a folder with light frames, to another folder with your dark library and it finds the darks that have the same exposure and sensor temperature as your lights and produces a dss file.
Yes, LEDs are not good for filtering, so if the entire city changes to those and light pollution on your sky is from white light then the broadband filters won't help. The narrowband (Ha, S-II, O-III etc) will still help of course with nebulae.
I guess the biggest improvement you can get is if you mod your camera to be sensitive to the red Ha/SII that many nebulae emit.

If using DSS then option one ,unless using pixinsight