newbie efforts turn bright orange

[almcl]

Hi,

Newcomer to StarTools and like many, I guess, have a few questions? My set up is an astro modded Canon 700d fitted to a SW200p reflector. Raw images (CR2) processed in Deep Sky Stacker where I probably have just enough knowledge to really mess things up.

Have had the camera for almost two weeks now and been trying to make my light polluted, no-flats, too-excited-to-take-darks stack, which looks like this on exit from DSS:

look a bit better but whenever I use autodev the image turns bright orange

and I can't get rid of this with Wipe and when I try in the color module the results aren't always an improvement.

Is there anything I should try (apart from moving to a less light polluted location) that might help?

Second, when wiping, strange geometric shapes show up:

and while these sometimes go away, the result isn't quite what was expected:

Advice would be most welcome!

Work flow is load, select 1st data option, bin (50%), autodev with ROI on the nebula, wipe, look in surprise at result.

[admin]

Hmm.... That looks and smells like a bug... What version of StarTools are you running and 32-bit or 64-bit?

AutoDev will in first instance show up the bias, as to tell you "FIX THIS!". After fixing the issues AutoDev shows (e.g. a color bias like this, stacking artifacts, etc.), using it for your global stretch is typically the way to go (I rarely using anything else).

[almcl]

Thanks for the response, I am using the Windows 7 32 bit version of 1.3.5.289.

[admin]

Thanks for the response, I am using the Windows 7 32 bit version of 1.3.5.289.

Thanks! Is there any chance you could share the data with me? I can do a personalised worklfow with that (and test!).
I did notice though your 'Aggressiveness' in the screenshot is 0%, which will not do much but remove an even bias (e.g. it don't detect any gradients).

[almcl]

That's most kind, Ivo. Anyone told you your product support is awesome? Well, it is.

I've put the FITS file here: https://www.dropbox.com/sh/g135fxnu8gur ... yRDoa?dl=0

as this is the first time I've tried to share a Dropbox file, it may not have worked.

The other caveat is that there are three FITS files I might have used to get the result, I think the one there is the one, and it certainly goes bright orange.

As for your advice on the wipe aggressiveness, I think that's what the default is when I click the color cast button. Is there a 'correct' order to use those buttons? I've been clicking Gradient, then color cast, then vignette, but sometimes this just seems to go round in a circle and we end up back where we started. I may have misunderstood the way they should be used so any additional thoughts would be most welcome.

[admin]

Hi,

I had a look at the data. Can you tell me how you pre-process/stack it? The data appears to be stretched (but I could be wrong)

I wasn't able to reproduce the 'bug', but I'm wondering if the small stacking artifacts might have anything to do with it (leaving them in does produce problems and gradient remnants).

Autosave002(2).jpg (95.73 KiB) Viewed 7320 times

This is how I (conservatively) processed your data;

--- Auto Develop
To see what we got. Big red bias, severe oversampling (and tracking error), some small stacking artifacts.
--- Bin
To reduce oversampling (and improve SNR).
Parameter [Scale] set to [(scale/noise reduction 25.00%)/(1600.00%)/(+4.00 bits)]
--- Crop
To get rid of stacking arftifacts.
Parameter [X1] set to [12 pixels]
Parameter [Y1] set to [15 pixels]
Parameter [X2] set to [1286 pixels (-17)]
Parameter [Y2] set to [860 pixels (-14)]
--- Wipe
Vignetting preset. Please start taking flats - they're not optional!
Parameter [Dark Anomaly Filter] set to [4 pixels] in case of any dead pixels or small darker-than-real-background anomalies.
--- Auto Develop
Put ROI over M42.
Parameter [Ignore Fine Detail <] set to [3.6 pixels] to make AutoDev ignore fine noise
--- Deconvolution
Auto mask generation failed, possibly because the data was stretched. Stretched data will not yield mathematically correct deconvolution, but typically yields *something* (though suboptimal and with more artifacts and ringing).
I made my own star mask (Fat Stars preset, Shrink to get rid of noise that was picked up, then Grow x4, Invert).
Parameter [Radius] set to [1.5 pixels]
--- Wavelet Sharpen
Using same mask we made for decon.
Parameter [Small Detail Bias] set to [96 %] to optimize more for smaller detail in lieu of larger detail.
--- Color
Came up with perfect color balance all by itself.
Parameter [Dark Saturation] set to [4.20]
--- Wavelet De-Noise
NOTE: Settings for 1.4.300
Parameter [Grain Size] set to [10.6 pixels]
Parameter [Read Noise Compensation] set to [Off]
Parameter [Smoothness] set to [81 %]

Hope this helps!

[almcl]

Wonderful, Ivo. Many thanks, it certainly moves me forward considerably.

Understand about flats, but haven't quite finished making the light-box yet!

Found another parameter to turn off in DSS and restacked - much better results. Better understanding too, of how to apply the modules.

Also will remember to check polar alignment once on target, check the tightness of all bolts in the mount, use the recommended white balance on modded camera, and try to capture more than three subs!

Still ST has done an awesome job. Thanks!