StarTools

OK, yes I'm very new at this... With this ST's attempt, toward the end,I wanted to work on the stars a bit. So I did a star mask, and it ended up masking scads of little specs also, (is this noise I did not get rid of?) and when I subsequently rounded my stars the little specs remained in the image.
I have processed about 5 objects so far with ST's, and this is the 1st time I have run into this bad (I'm assuming this is a really bad noise issue) of a result after doing my final denoising. I have been acquiring most of my objects from Fits files I have created after stacking in Nebulosity using only light and dark frames; no bias or flats. This image was done the same. I'm using a Canon 60Da, in this case with a C-8. All my acquisitions are done unguided. In this image I used 31 lights @ 2.5 minutes, ISO 1600 and 39 darks.
So, did I just get away with my 1st 4 images using this method, and I need to add bias and flats, or could i have pushed the denoising further? I don't think this later is the case because, as it is it appeared to me that increasing the read-noise compensation any further would have made the image worse. Or... is there some other mistake I may have made using ST's? Or, do I really have to start doing a better job in acquisition?
Any thoughts or advise will be appreciated.
CS, Tom

Hi Tom,

trobbet wrote:OK, yes I'm very new at this... With this ST's attempt, toward the end,I wanted to work on the stars a bit. So I did a star mask, and it ended up masking scads of little specs also, (is this noise I did not get rid of?) and when I subsequently rounded my stars the little specs remained in the image.

The star mask generator sometimes picks up noise in addition to stars (it sometimes hard for the algorithm to tell the difference). Fortunately there are many ways to get rid of the specks.
One way is to increase the Filter Sensitivity parameter (confusingly, increasing this value makes the algorithm less sensitive).
The other way is to let the default preset do its thing, and then click 'Shrink' and then click 'Grow'. This will shrink away fine noise specks and regrow the bigger specks that were affected by the 'Shrink' operation.
Remember you can always manually touch up any mask, or you can configure the 'Auto' feature to do completely different things - you don't have to settle for the defaults or presets!

Without seeing the data, before and after its hard to make any evaluations with regards to acquisition quality and/or things you might do different in StarTools. Maybe you could share it with me?

Thank you Ivo!

I sent you links to my fit file and my current Tiff files for the image I'm having trouble with in Dropbox. If there is something more that is needed let me know. Once again, I so appreciate your software, and your excellent support!

I will work with the tools you suggested as soon as I get a chance. Probably because of my newness to using imaging software, when I do read the information provided within ST's, I don't really understand the terminology. When I do "get it", I am so impressed with what ST's can do! The above explanation does seem quite clear now, and I believe it will make it so I "get it" once again.

CS, Tom

No problem Tom,

Happy to help where I can!

Here's what I came up with using the data you provided;
Processed as follows.
File loaded [/media/WD/home/irwjager/Downloads/M-77 Stacked 1600 C-8 2.5 min..fit].
--- Auto Develop
As always, I do an AutoDev (default settings) to see what we got.
We can see stacking artifacts. Severe oversampling (e.g. detail is 'smeared out' over multiple pixels).
We can also see some tracking error (slight star elongation across the image in the same direction).
--- Bin
First thing I do is bin the image to 35%, so that 1 unit of detail better corresponds to 1 pixel.
This way we convert the useless oversampling into noise reduction (by pooling multiple pixels into one new reading) without losing visible detail.
Parameter [Scale] set to [(scale/noise reduction 35.38%)/(798.89%)/(+3.00 bits)]
--- Crop
Next, I crop away the stacking artifacts.
Parameter [X1] set to [206 pixels]
Parameter [Y1] set to [155 pixels]
Parameter [X2] set to [1779 pixels (-269)]
Parameter [Y2] set to [1124 pixels (-284)]
--- Crop
It turned out that during Wiping, more stacking artifacts were visible at the top. Wipe will back off and leave some light pollution where this is the case, as it will regard the darker stacking artifact as 'the new background' and it is very careful to never remove a higher pixel value than the background permits.
Parameter [X1] set to [0 pixels]
Parameter [Y1] set to [51 pixels]
Parameter [X2] set to [1573 pixels (-0)]
Parameter [Y2] set to [919 pixels (-50)]
--- Crop
More stacking artifacts were visible at the left edge.
Parameter [X1] set to [16 pixels]
Parameter [Y1] set to [1 pixels]
Parameter [X2] set to [1569 pixels (-4)]
Parameter [Y2] set to [868 pixels (-0)]
--- Wipe
I used the Vignetting preset and wiped with these parameters.
Parameter [Temporary AutoDev] set to [Yes] (so we can see what Wipe is doing).
Parameter [Dark Anomaly Filter] set to [5 pixels] (just in case there are any dark/dead pixels in the image that could potentially confuse Wipe as to how dark the real interstellar background is).
I also created a mask with M77 'cut out' (using the Lassoo brush in the mask editor - you should end up with everything colored 'green', except for the area where M77 is located). The reason I do this is bcause I don't want Wipe to touch the fainter halo around M77 - we don't want it to think it's localized light pollution! (The higher you set Aggressiveness, the quicker Wipe decides that localized halos and such might constitute light pollution)
--- Auto Develop
Final stretch of the Wiped image.
Parameter [Ignore Fine Detail <] set to [3.3 pixels] (this stops AutoDev optimizing the stretch for the finer noise in the image - we don't really care about seeing that!).
I specified a ROI (click & drag) around M77.
--- Deconvolution
I applied some deconvolution (click AutoMask, as else, Decon will think that we want to use the current active mask that we created for Wipe).
Parameter [Radius] set to [2.7 pixels]
Parameter [Iterations] set to [9]
Parameter [Regularization] set to [1.11 (smoother, less detail)]
--- Wavelet Sharpen
Applied some Wavelet Sharpening.
Parameter [Intelligent Enhance] set to [No]
Parameter [Scale 1] set to [10 %]
Parameter [Scale 2] set to [10 %]
Parameter [Amount] set to [184 %]
Parameter [Small Detail Bias] set to [87 %]
--- HDR
Reveal preset.
--- Life
Isolate preset to push back noise and keep M77. Be sure to 'Clear, Invert' the mask (so that the whole image is selected).
--- Color
Final color calibration.
What we're looking for is a good continuum of star temperatures (red->orange->yellow->white->blue). Typically, a galaxy has a yellow core (older stars) and a purplish/blue ring (younger stars and HII regions)
However, here we see some issues with the data. It appears the color channels either contain false color data, or the color data has been misaligned (notice how blue always sits higher than red/orange?). This is something to sort out in Nebulosity (or whatever other program you used to stack).
I settled for the following parameters;
Parameter [Dark Saturation] set to [4.40]
Parameter [Bright Saturation] set to [Full]
Parameter [Saturation Amount] set to [236 %]
Parameter [Green Bias Reduce] set to [1.20]
Parameter [Red Bias Reduce] set to [1.70]
--- Wavelet De-Noise
FInal denoising, switching Tracking off;
Parameter [Color Detail Loss] set to [15 %]
Parameter [Brightness Detail Loss] set to [14 %]
Parameter [Read Noise Compensation] set to [10.72 %]

That's it! From here we could Repair (round) the stars if you feel so inclined. This will also help with the color misalignment.

Hope this helps!

Ivo,

Thanks very much!

Its funny..., I worked on this image some more last night, and was having trouble at first. After shirking the specs, I couldn't get the stars to grow right. Sometimes, only a few of the stars would grow, or I shrunk everything too much (I later realized), until I was able to get the hang of the right combination of shrinking and growing, and the right filter sensitivity. In other words, it took a little practice, but sticking with it, paid off. Of course my final image isn't as good as yours, but I was pretty pleased with it. And now with your example, I know I will learn a lot more.

I continue to be impressed with what one can do with these less than perfect, unguided images (only stacked with Darks), using StarTools!

CS, Tom

"-- Color
Final color calibration.
What we're looking for is a good continuum of star temperatures (red->orange->yellow->white->blue). Typically, a galaxy has a yellow core (older stars) and a purplish/blue ring (younger stars and HII regions)
However, here we see some issues with the data. It appears the color channels either contain false color data, or the color data has been misaligned . (notice how blue always sits higher than red/orange?)This is something to sort out in Nebulosity (or whatever other program you used to stack).
I settled for the following parameters;
Parameter [Dark Saturation] set to [4.40]
Parameter [Bright Saturation] set to [Full]
Parameter [Saturation Amount] set to [236 %]
Parameter [Green Bias Reduce] set to [1.20]
Parameter [Red Bias Reduce] set to [1.70]
--- Wavelet De-Noise
FInal denoising, switching Tracking off;
Parameter [Color Detail Loss] set to [15 %]
Parameter [Brightness Detail Loss] set to [14 %]
Parameter [Read Noise Compensation] set to [10.72 %]

That's it! From here we could Repair (round) the stars if you feel so inclined. This will also help with the color misalignment.

Hope this helps!
Ivo Jager"




Ivo,

So, I have been using my original data and your very helpful example in order to learn StarTools better. I went back to Nebulosity and did some reading, and then some experimenting. I don't really understand what you mean by: "notice how the blue always sits higher than the red/orange". Nonetheless, I did some playing around and found a tool in Nebulosity called "colors in motion", and thought that might help. It took a while to figure out how to use it as I couldn't find an outline in the literature. But, I was able to come up with an alternate stack; although i don't know if it made any difference as far as color goes. However, this new stack appears significantly less noisey, even though the RAW data was the same.

I'm including a new image I created with this new stack. Although, I largely copied Ivo's example, somehow I missed something as it seems the fainter detail surrounding the Galaxy's core in Ivo's image is more distinct than in mine. Also Ivo, I'm having trouble seeing what setting Parameter's 1&2 to 10% as opposed to 100% in Wavelet Sharpen is doing. I fail to see any difference, but I copied you anyhow. Using this new stack I felt I didn't need any Read Noise Compensation in the final de-noise.

I did end up working a bit on the stars.

And, in light of my earlier post today, for whatever reason i was able to save this image the very 1st time! Go figure....?????

Here's my log:

File loaded [/Users/trobbe/Documents/Neb k1 folder/M-77 color motion 2.fit].
---
--- Bin
Parameter [Scale] set to [(scale/noise reduction 35.38%)/(798.89%)/(+3.00 bits)]
--- Auto Develop
Parameter [Ignore Fine Detail <] set to [Off]
Parameter [Outside ROI Influence] set to [15 %]
--- Crop
Parameter [X1] set to [290 pixels]
Parameter [Y1] set to [224 pixels]
Parameter [X2] set to [1705 pixels (-345)]
Parameter [Y2] set to [1108 pixels (-301)]
--- Wipe
Parameter [Mode] set to [Correct Color & Brightness]
Parameter [UNKNOWN] set to [Yes]
Parameter [Precision] set to [256 x 256 pixels]
Parameter [Dark Anomaly Filter] set to [5 pixels]
Parameter [Drop Off Point] set to [0 %]
Parameter [Corner Aggressiveness] set to [100 %]
Parameter [Aggressiveness] set to [75 %]
Undo.
--- Wipe
Parameter [Mode] set to [Correct Color & Brightness]
Parameter [UNKNOWN] set to [Yes]
Parameter [Precision] set to [256 x 256 pixels]
Parameter [Dark Anomaly Filter] set to [5 pixels]
Parameter [Drop Off Point] set to [0 %]
Parameter [Corner Aggressiveness] set to [100 %]
Parameter [Aggressiveness] set to [75 %]
--- Auto Develop
Parameter [Ignore Fine Detail <] set to [3.3 pixels]
Parameter [Outside ROI Influence] set to [15 %]
--- Deconvolution
Parameter [Image Type] set to [Deep Space]
Parameter [Mask Behavior] set to [De-ring Mask Gaps, Hide Result]
Parameter [Radius] set to [2.9 pixels]
Parameter [Iterations] set to [7]
Parameter [Regularization] set to [1.06 (smoother, less detail)]
Parameter [Mask Fuzz] set to [8.0 pixels]
--- Wavelet Sharpen
Parameter [Intelligent Enhance] set to [Yes]
Parameter [Scale 1] set to [10 %]
Parameter [Scale 2] set to [10 %]
Parameter [Scale 3] set to [100 %]
Parameter [Scale 4] set to [100 %]
Parameter [Scale 5] set to [100 %]
Parameter [Mask Fuzz] set to [8.0 pixels]
Parameter [Amount] set to [185 %]
Parameter [Small Detail Bias] set to [87 %]
--- HDR
Parameter [Small Detail Precision] set to [Max]
Parameter [Channels] set to [Brightness Only]
Parameter [Algorithm] set to [Reveal DSO Core]
Parameter [Dark/Bright Response] set to [Full]
Parameter [Detail Size Range] set to [64 pixels]
Parameter [Noise Suppression] set to [Off]
--- Life
Parameter [Detail Preservation] set to [Linear Brightness Mask]
Parameter [Compositing Algorithm] set to [Multiply, Gamma Correct]
Parameter [Inherit Brightness, Color] set to [Off]
Parameter [Output Glow Only] set to [No]
Parameter [Airy Disk Sampling] set to [128 x 128 pixels]
Parameter [Airy Disk Radius] set to [8 pixels]
Parameter [Glow Threshold] set to [0 %]
Parameter [Detail Preservation Radius] set to [20.0 pixels]
Parameter [Saturation] set to [100 %]
Parameter [Strength] set to [100 %]
Parameter [Mask Fuzz] set to [1.0 pixels]
--- Color
Parameter [Cap Green] set to [No]
Parameter [Bias Slider Mode] set to [Sliders Reduce Color Bias]
Parameter [Style] set to [Scientific (Color Constancy)]
Parameter [LRGB Method Emulation] set to [Straight CIELab Luminance Retention]
Parameter [Dark Saturation] set to [4.40]
Parameter [Bright Saturation] set to [Full]
Parameter [Saturation Amount] set to [235 %]
Parameter [Blue Bias Reduce] set to [1.00]
Parameter [Green Bias Reduce] set to [1.65]
Parameter [Red Bias Reduce] set to [1.40]
Parameter [Mask Fuzz] set to [1.0 pixels]
--- Wavelet De-Noise
Parameter [Scale 1] set to [90 %]
Parameter [Scale 2] set to [90 %]
Parameter [Scale 3] set to [90 %]
Parameter [Scale 4] set to [90 %]
Parameter [Scale 5] set to [0 %]
Parameter [Mask Fuzz] set to [1.0 pixels]
Parameter [Scale Correlation] set to [3]
Parameter [Color Detail Loss] set to [15 %]
Parameter [Brightness Detail Loss] set to [14 %]
Parameter [Redistribution Kernel] set to [4.5 pixels]
Parameter [Read Noise Compensation] set to [Off]
Parameter [Smoothness] set to [49 %]
--- Repair
Parameter [Radial Samples] set to [32]
Parameter [Sub Sampling] set to [4x]
Parameter [Algorithm] set to [Redistribute, Core Is Avg Location]
Parameter [Grow Mask] set to [0 pixels]
--- Layer
Parameter [Layer Mode] set to [Lighten]
Parameter [Cap Mode] set to [Clip]
Parameter [Brightness Mask Mode] set to [Off]
Parameter [Filter Type] set to [Gaussian (Fg)]
Parameter [Blend Amount] set to [100 %]
Parameter [Mask Fuzz] set to [1.0 pixels]
Parameter [Filter Kernel Radius] set to [3.6 pixels]
Parameter [Offset X] set to [0.0 pixels]
Parameter [Offset Y] set to [0.0 pixels]
Parameter [Brightness Mask Power] set to [1.00]
File saved [/Users/trobbe/Documents/Neb k1 folder/M-77 2-9-14 color motion.tiff].

I continue to love your software! Thanks!!!

Tom

Love this rendition! Very good colors too.

trobbet wrote:I don't really understand what you mean by: "notice how the blue always sits higher than the red/orange".

Sorry that wasn't very clear. I meant that there seems to be blue fringing at the top stars and red/orange fringing at the bottom of stars. This seems to indicate color channel misalignment.

Also Ivo, I'm having trouble seeing what setting Parameter's 1&2 to 10% as opposed to 100% in Wavelet Sharpen is doing. I fail to see any difference, but I copied you anyhow.

Wavelet sharpening is like a graphic equalizer (remember those from the 80s? ) for images instead of sound.
Instead of manipulating the prevalence of high (let's say a flute) and low frequencies (let's say a bass) in sound, you're manipulating high (small detail) and low frequencies (large detail) in your image.
Noise in images tends to be the most prevalent in the high frequencies (fine detail). This is why boosting the amplitude of the fine detail (which is the frequency - aka detail size - which scale 1 and, to a lesser extent, scale 2 govern) will also boost the visible noise (along with any fine detail). Because we can see the data is already quite noisy, I refrained from boosting the fine detail much, as to no exacerbate the noise.

StarTools Wavelet sharpening is a bit more sophisticated than most other Wavelet Sharpening algorithms, but that's a story for another post...