Stuck with bad data or just bad editing?

Questions and answers about processing in StarTools and how to accomplish certain tasks.
Post Reply
Apagador
Posts: 6
Joined: Fri Dec 18, 2020 8:14 pm

Stuck with bad data or just bad editing?

Post by Apagador »

I'm just starting out taking pictures of the night sky and I'm a bit stuck. Two nights ago it was clear here, a bit chilly at about -20°C. I tried to image the elephant's trunk nebula using astro modded canon 6D, 150mm f2.8 Sigma lens wide open (I know that this might be a problem, but I'm willing to crop out the edges and I was greedy for light. Of course it also affects the center of the image) on a HEQ-5 (unguided). I took a total of 58 exposures of 185 seconds each (my intervalometer and camera see the time moving at different rates so 3 minutes = 185 seconds). I took 15 dark frames, 30 flats using an aurora flatfield panel and a bunch of bias frames. I stacked them using APP as instructed when stacking for Startools and then tried to make a decent image out of them. I know that I can't expect much out of 3 hours of data, but I was hoping for something useful anyways as a learning tool. I'm sometimes overly optimistic :)

Here is where I'm stuck. I don't know if my data is hopelessly bad or if I'm just hopeless in editing (well this is a given really). All I can get out of the images is a candy colored mess where the stars overwhelm everything else and I don't know if I can even use these images as a learning data for getting to know startools.

Would be really grateful if someone would be willing to take a look at my stacked image and tell me what I'm clearly doing wrong in acquiring data or if there is something worth trying to salvage.

Capture1.jpg
Capture1.jpg (422.36 KiB) Viewed 3838 times

Here is the FITS file from APP: https://drive.google.com/file/d/1ltMBhP ... sp=sharing

I would love to see if someone could show me what's possible with this file, or even just give me some pointers.
Burly
Posts: 387
Joined: Mon Aug 31, 2015 7:06 am
Location: Northamptonshie uk

Re: Stuck with bad data or just bad editing?

Post by Burly »

i would be happy with that data as ever more data the better ,there seems to be an area bottom left of centre that i cant seem to rid the cloudiness this may be caused by flats you would need to stack with out flats to see ,also try with bias frames too maybe ivo will have a better idea , here is my go at your data using 1.7.455 rc3 other than that data seems fine

regards Dave
Attachments
elephants_trunk_for_startools (1).jpg
elephants_trunk_for_startools (1).jpg (33.23 KiB) Viewed 3827 times
Apagador
Posts: 6
Joined: Fri Dec 18, 2020 8:14 pm

Re: Stuck with bad data or just bad editing?

Post by Apagador »

Thank you for the reply. Glad to hear that the data is not worthless and that I'm just bad at editing. That's an area that I can work on even when it's cloudy outside :D

For some reason the image you posted shows up as tiny for me, but I think I can see the cloudy area. I'll try to stack the images without flats when I have the time and see if it's still there.

Thank you again for taking a look at the data and giving me hope that I should now mainly focus on improving my editing and that I don't need to start from scratch on acquiring data and editing at the same time :)
Burly
Posts: 387
Joined: Mon Aug 31, 2015 7:06 am
Location: Northamptonshie uk

Re: Stuck with bad data or just bad editing?

Post by Burly »

yes sorry resized due to file size ,pm me and i will email you the full image

regards Dave
Apagador
Posts: 6
Joined: Fri Dec 18, 2020 8:14 pm

Re: Stuck with bad data or just bad editing?

Post by Apagador »

I don't think I'm allowed to send private messages yet :think: my email address is my alias at gmail :)
Apagador
Posts: 6
Joined: Fri Dec 18, 2020 8:14 pm

Re: Stuck with bad data or just bad editing?

Post by Apagador »

Took a closer look at my light frames and the lighter colored area in the lower left is visible in 36 of the light frames and then just completely disappears. Now I'm thinking that it's a reflection from some light that was shining on my lens hood? The lens hood on my Sigma is far from being completely black. It's more of a dark grey color so I'm guessing it could easily cause problems if some light is shining on it. This would mean that my flat frames are probably ok?
Burly
Posts: 387
Joined: Mon Aug 31, 2015 7:06 am
Location: Northamptonshie uk

Re: Stuck with bad data or just bad editing?

Post by Burly »

That’s good you’ve found the cause , maybe try using some self adhesive flocking material on inside of hood , here in the uk you can buy a roll from Wilkinson’s , I’m sure you’ll find it on eBay , also be mindful of dew forming on lens so if cloudiness happens later on in images then possibly dew starting to form .
User avatar
admin
Site Admin
Posts: 3367
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: Stuck with bad data or just bad editing?

Post by admin »

Hi,

You probably already saw my replies on CN, but I'll post them here as well, as it may be useful for others.
If you put the single sub you uploaded through dcraw like so;

dcraw -T -4 -o 0 -r 1 1 1 1 IMG_4108.CR2

You will obtain what a stacker does/uses internally (provided it is configured correctly).

In terms of luminance, the resulting frame looks virtually identical to the stack - plenty of stars visible (flat frame correction not withstanding of course).

So many stars visible in your linear data may or may not be a problem in some instances, depending on whether your camera's sensor response is truly linear, even when approaching saturation (it can affect deconvolution's effectiveness for example). Right now though, you probably have bigger fish to fry than worrying about that.

In terms of coloring, however, things look a bit different;
IMG_4108.jpg
IMG_4108.jpg (105.94 KiB) Viewed 3790 times
This is pretty close to the expected color signature for a modded camera or OSC.

In contrast, your current full stack appears color balanced somehow, with the strong red response quite evident;
post-194246-0-93942400-1608425994.jpg
post-194246-0-93942400-1608425994.jpg (140.55 KiB) Viewed 3790 times
Ergo, something is meddling with your coloring, which - at the end of the day - is a fairly minor problem and comes down to just changing a setting somewhere. Otherwise, comparing the single frame and the stack, shows good calibration and a nice, flat field. You use the right ISO for your camera and doing a lot right.

The lens you are using is probably the weakest link in your optical train right now. Lenses meant of terrestrial use, exhibit chromatic aberration that is (depending on your standards of course) often not acceptable for AP goals. Chromatic aberration will manifest itself in blue or purple rings of color. They can be processed out however. Worse, though, some/most stackers will have trouble aligning stars and color channels properly, resulting in color fringing of stars (as is the case here), while some stackers can wholesale introduce aberrant color information themselves (as is the case here as well) if not configured properly;
Selection_406.jpg
Selection_406.jpg (3.33 KiB) Viewed 3790 times
(an overexposing core should obviously not have a color like this)

In such case, color preservation through something like StarTools' "color constancy" mode will readily reveal the aberrant color information. Again, there are several techniques to hide the problem (using the Highlight Repair feature in the Color module is a good start). You can also decide to not use the Color Constancy mode (again, this is hiding the problem and not a "solution" as such). Stopping down your lens can also help mitigate some of the worst chromatic aberration. EDIT: FWIW, AutoDev counters stellar bloating, so using it, particularly in your case is highly(!) recommended.

Chromatic aberration also bloats stars in the luminance domain, causing stars to quickly take over from any larger scale nebulosity. In widefields like these, that quickly becomes a serious problem. You can use ST's Shrink module or PI's morphological transform for stellar profile manipulation, and/or ST's Super Structure module or PI's wavelet transform to manipulate the larger scale structures vs small stuff like the stars, in order to mitigate the problems with the stars.

To top off the cascade of "problems", is the fact that you're not guiding, which causes stars, which already have obnoxiously large stellar profiles due to the chromatic aberration, to "smear around".

All said and done, you're well on your way - it's all tweaks to your acquisition and software settings, and learning how to work around any left-over issues inherent to your gear. Let us know how you go!

TL;DR Data.
And message 2;
A quick process in StarTools with mostly defaults where possible, taking into account all the observations of my previous post;
Image
Workflow as follows;

--- Imported with first option (linear) as this dataset appears to be already color balanced.

--- Auto Develop

To see what we got. See previous post for some observations.
--- Bin

This makes the tracking error and color fringing a little less obvious. It improves signal as a bonus, but really, it is quite strong as-is.
Parameter [Scale] set to [scale 50.00% / +2.00 bits / +1.41x SNR improvement]
Image size is 2744 x 1834

--- Crop

Get rid of some stacking artifacts.
Parameter [X1] set to [11 pixels]
Parameter [Y1] set to [16 pixels]
Parameter [X2] set to [2736 pixels (-8)]
Parameter [Y2] set to [1819 pixels (-15)]
Image size is 2725 x 1803
--- Wipe

Defaults.
Parameter [Dark Anomaly Filter] set to [4 pixels] just in case - the image is very busy and I am finding it hard to see any dark pixel clumps that don't belong...
--- Auto Develop

Defaults. No RoI.
--- Color

Legacy preset, which mimics the way old/conventional software desaturates highlights as chrominance information is stretched along with the luminance.

Default color balance was used. What we're ideally after - even with a modded camera - is a good representation of all stellar temperatures. Red->orange->yellow->white->blue. That is how you know you got a good color balance. All stars should have a color when you zoom in. When using the Legacy preset, the color will be subtle due to the desaturation but it should nevertheless be detectable. The H-alpha is obviously red, tending to pink due to other bluer emissions and/or reflections that almost always are also present. Green is a very rare color in outer space (notable exception O-III dominant areas such as M42's core), so if you are absolutely sure your color balance is otherwise good, then any leftover green can be booted using the Cap Green parameter.
Parameter [Cap Green] set to [100 %] to get rid of any spurious green
Parameter [Highlight Repair] set to [Off], as the oldschool desaturation trick/limitation was enough to hide the issues with chrominance information in the highlights.
--- Shrink

As said in my previous post, taming the star field can really help with widefields like these. Complete removal of some (or all) stars is IMHO not OK if you're doing documentary photography. Some even find shrinking them objectionable, but IMO it can really help images like these, particularly if you avoid complete removal of the dimmer ones, and - of course - avoid artefacts.

Auto Mask.

Parameter [Halo Extend] set to [0 pixels] (no need to grow mask which is already very busy).
Parameter [De-ringing] set to [Off] (only really needed when decon was used)
--- Super Structure

Saturate preset (you can try the others too obviously).

The Ha nebulosity is the star of the show here, and this is an easy way of emphasizing it without performing (frowned upon) manual selective processing. As said, you can also use the Entropy module here.

Parameter [Airy Disk Radius] set to [11 %] to match field of view (smaller stars)
Parameter [Saturation] set to [100 %] to moderate the effect a little.


No noise reduction applied. Tweak the defaults to taste obviously - this is just a quick run-through.


Again, it's preferable if you can fix the acquisition/optics and stacking issues (tracking, chromatic aberration, color balancing) so there is less to work around and worry about, but you're well on your way! :thumbsup:
Ivo Jager
StarTools creator and astronomy enthusiast
Apagador
Posts: 6
Joined: Fri Dec 18, 2020 8:14 pm

Re: Stuck with bad data or just bad editing?

Post by Apagador »

And here is the single subexposure that Ivo is using in the previous post. If someone is tracing his steps :)

https://drive.google.com/file/d/13tmBUP ... sp=sharing
Apagador
Posts: 6
Joined: Fri Dec 18, 2020 8:14 pm

Re: Stuck with bad data or just bad editing?

Post by Apagador »

And thank you Burly for the tip about flocking material. Will have to get some of that to fight with the reflections. I used a dew heater on the lens so that wasn't a problem, but there was a layer of ice on everything else by the time I was done imaging :)
Post Reply