StarTools 1.8.512 public alpha/preview 4 now out

General discussion about StarTools.
happy-kat
Posts: 318
Joined: Sun Feb 01, 2015 11:31 am

Re: StarTools 1.8.511 public alpha/preview 3 now out

Post by happy-kat »

If I'm working with an image with strong CA then I exclude the stars with strong CA from the colour module and leave then in luminance colour.
User avatar
admin
Site Admin
Posts: 2866
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: StarTools 1.8.512 public alpha/preview 4 now out

Post by admin »

Some small fixes now available in 1.8.512.
  • Fixed (dumbed down) color preservation in Contrast module
  • Fixed disappeared Help button in SVDecon
  • Fixed color artifacts sometimes introduced when using Mask Fuzz in Color module
  • Fixed SV Decon PSF Load dialog not appearing until after completion of processing (abort early introduced)
  • Fixed SV Decon PSF Load only working first time
Clear skies!
Ivo Jager
StarTools creator and astronomy enthusiast
Mike in Rancho
Posts: 124
Joined: Sun Jun 20, 2021 10:05 pm

Re: StarTools 1.8.512 public alpha/preview 4 now out

Post by Mike in Rancho »

Awesome. Thanks! :obscene-drinkingcheers:

Is there much in the way of instruction yet on the PSF load? Or save?

I know the logs are coded, and there's an online converter to go through some steps and end up with a graphics file for masks. Of course the easy way to avoid that is -- just save your masks as a jpg from ST, dummy!

But I don't think I've seen a save PSF button. And in the log it's pretty small, like one line so far from what I've done (short enough to just copy and paste if the PSF load allowed that). It does say Base 64, though not png, but does that mean use that online converter to create a graphics file?
User avatar
admin
Site Admin
Posts: 2866
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: StarTools 1.8.512 public alpha/preview 4 now out

Post by admin »

Mike in Rancho wrote: Sun Aug 29, 2021 8:24 pm Awesome. Thanks! :obscene-drinkingcheers:

Is there much in the way of instruction yet on the PSF load? Or save?

I know the logs are coded, and there's an online converter to go through some steps and end up with a graphics file for masks. Of course the easy way to avoid that is -- just save your masks as a jpg from ST, dummy!

But I don't think I've seen a save PSF button. And in the log it's pretty small, like one line so far from what I've done (short enough to just copy and paste if the PSF load allowed that). It does say Base 64, though not png, but does that mean use that online converter to create a graphics file?
This needs documentation indeed.
You can just grab the BASE64 string, stick it into a new text file with a text editor, and open that in the SVDecon module. That should work.
Ivo Jager
StarTools creator and astronomy enthusiast
Mike in Rancho
Posts: 124
Joined: Sun Jun 20, 2021 10:05 pm

Re: StarTools 1.8.512 public alpha/preview 4 now out

Post by Mike in Rancho »

admin wrote: Mon Aug 30, 2021 2:02 am This needs documentation indeed.
You can just grab the BASE64 string, stick it into a new text file with a text editor, and open that in the SVDecon module. That should work.
It did! :D

Now, it doesn't remember the blue box size, so that still needs to be reset. But, I also tried to trip it up by erasing one of the PSF stars from the apod mask before loading PSF, and there was no problem - it was simply left out.

Should work great for following a prior log or maybe even the Replay app (which I haven't tried yet). However, for in-session usage, the PSF code isn't written to the log until keep is hit on SVD. So if one wants the chosen PSF's to be memorialized, one must make sure to do that, which can be undone if not really ready to keep, or if one wanted to hit cancel to do something else before deconv.
Mike in Rancho
Posts: 124
Joined: Sun Jun 20, 2021 10:05 pm

Re: StarTools 1.8.512 public alpha/preview 4 now out

Post by Mike in Rancho »

As a heads-up, I have recently noticed some SVD wild pixels again in 512. :cry:

Far more rare in both occurrence in any particular image, as well as amount generated as it is quite minor compared to the first time. So pretty hard to detect especially if you aren't looking closely.

I haven't yet put time into finding repeatable patterns or whether it was in 511 also, but I will start doing that and looking for them to see what I can learn.
Mike in Rancho
Posts: 124
Joined: Sun Jun 20, 2021 10:05 pm

Re: StarTools 1.8.512 public alpha/preview 4 now out

Post by Mike in Rancho »

I did a little more work on wild pixel testing. Not a lot, I still haven't checked across many data sets, so it might just be this particular one. That said it has happened in multiple iterations of it, i.e. as I've added more data and stacked.

Here's a screenshot where you can see the two bright stars have little pixels dangling off of them, in more or less the same orientation. There is a third bright star off-screen that has the same thing.

ST Screenshot SVD pixels again.jpg
ST Screenshot SVD pixels again.jpg (385.36 KiB) Viewed 163 times

I have tried multiple settings to eliminate them, to no avail, including selecting and unselecting various sample stars, varying the number of sample stars used (this changes the pixels around somewhat), altering the mask around the affected stars, altering the apod mask around the sample star, trying the higher precision apod mask, etc. etc. I tried the data in 511 and it happened there too. Basically, at 2x iterations and above, the pixels show up. The deringing settings, intra-iteration/centroid, spatial error, and so on do not seem to work either (for fixing the pixels, that is).

I also tried at both the standard 50% bin as well as heavier bins. Did not yet try 71% or just leaving it full res. The wild pixels survive and must be healed out at the end (if you watched and know what and where they are).

Here is the dataset used (maybe it has problems?) - 9.5 L'eNhance hours of Tulip, stacked in DSS, kappa 3 (trying to get the Cyg X1 shock wave).

https://drive.google.com/file/d/1sDw9DH ... sp=sharing

On the last test I just ran through a typical default workflow the way one commonly would but here's a log anyway to the point where I closed it. The offset crop is due to D5300 sensor edge glow which the darks didn't remove. It was a hot night for imaging, and Nikon darks are very hit and miss. You have to get lucky, and in this case, I didn't. :lol:

Code: Select all

-----------------------------------------------------------
StarTools 1.8.512alpha
Sat Sep 11 11:58:56 2021
-----------------------------------------------------------
Loading red channel data
File loaded in LRGB module [F:\ASTROPHOTOGRAPHY\DSO SH 2-101\SH2-101 Sept 2021 processing\Tulip 9pt5hr duo k3.fts].
Loading green channel data
File loaded in LRGB module [F:\ASTROPHOTOGRAPHY\DSO SH 2-101\SH2-101 Sept 2021 processing\Tulip 9pt5hr duo k3.fts].
Loading blue channel data
File loaded in LRGB module [F:\ASTROPHOTOGRAPHY\DSO SH 2-101\SH2-101 Sept 2021 processing\Tulip 9pt5hr duo k3.fts].
--- Compose
Parameter [Luminance, Color] set to [L + Synthetic L From R(2xG)B, R(GB)(GB) (Bi-Color from OSC/DSLR)]
Parameter [Color Ch. Interpolation] set to [On]
Parameter [NB Accents Type] set to [Ha/S-II from NB filter]
Parameter [Lum Total Exposure] set to [Not set]
Parameter [Blue Total Exposure] set to [1h00m (60m) (3600s)]
Parameter [Green Total Exposure] set to [1h00m (60m) (3600s)]
Parameter [Red Total Exposure] set to [1h00m (60m) (3600s)]
Image size is 5840 x 3891
Type of Data: Linear, was not Bayered, or was Bayered + white balanced
--- Crop
Parameter [X1] set to [362 pixels]
Parameter [Y1] set to [250 pixels]
Parameter [X2] set to [5810 pixels (-30)]
Parameter [Y2] set to [3861 pixels (-30)]
Image size is 5448 x 3611
--- Rotate
Parameter [Angle] set to [180.0]
--- Bin
Parameter [Scale] set to [scale 50.00% / +2.00 bits / +1.00x SNR improvement]
Image size is 2724 x 1805
--- Wipe
Parameter [Synthetic Dark/Bias] set to [Off]
Parameter [Gradient Edge Behavior] set to [Absorb 50%]
Parameter [Synthetic Flats] set to [Off]
Parameter [Sampling Precision] set to [256 x 256 pixels]
Parameter [Dark Anomaly Filter] set to [1 pixels]
Parameter [Gradient Falloff] set to [0 %]
Parameter [Synth. Bias Edge Area] set to [100 %]
Parameter [Gradient Aggressiveness] set to [75 %]
Parameter [Correlation Filtering] set to [0.5 pixels]
--- Auto Develop
Parameter [Ignore Fine Detail <] set to [3.0 pixels]
Parameter [Outside RoI Influence] set to [15 %]
Parameter [RoI X1] set to [0 pixels]
Parameter [RoI Y1] set to [0 pixels]
Parameter [RoI X2] set to [2724 pixels (-0)]
Parameter [RoI Y2] set to [1805 pixels (-0)]
Parameter [Detector Gamma] set to [1.00]
Parameter [Shadow Linearity] set to [50 %]
--- Contrast
Parameter [Expose Dark Areas] set to [Yes]
Parameter [Brightness Retention] set to [Off]
Parameter [Precision] set to [256 x 256 pixels]
Parameter [Shadow Detail Size] set to [10 pixels]
Parameter [Locality] set to [50 %]
Parameter [Shadow Dyn Range Alloc] set to [50 %]
--- HDR
Parameter [Small Detail Precision] set to [Max]
Parameter [Channels] set to [Brightness Only]
Parameter [Algorithm] set to [Reveal All]
Parameter [Dark/Bright Response] set to [1.00]
Parameter [Detail Size Range] set to [1000 pixels]
Parameter [Strength] set to [1.2]
--- Wavelet Sharpen
Parameter [Structure Size] set to [Large]
Mask used (BASE64 PNG encoded)
--- SNR-aware Wavelet Sharpening
Parameter [Protection] set to [Shadow/Highlights]
Parameter [Scale 1] set to [100 %]
Parameter [Scale 2] set to [100 %]
Parameter [Scale 3] set to [100 %]
Parameter [Scale 4] set to [100 %]
Parameter [Scale 5] set to [100 %]
Parameter [Mask Fuzz] set to [4 pixels]
Parameter [Amount] set to [300 %]
Parameter [High SNR Size Bias] set to [85 %]
Parameter [Low SNR Size Bias] set to [0 %]
Parameter [Dark/Light Enhance] set to [50% / 50%]
Mask used (BASE64 PNG encoded)
PSF samples used (1 PSF sample locations, BASE64 encoded)
VFMAAAEApAoNB8YJTQQ=
--- Spatially Variant PSF Deconvolution
Parameter [PSF Resampling] set to [None]
Parameter [Synthetic PSF Model] set to [Circle of Confusion (Optics Only)]
Parameter [Sampled PSF Area] set to [19x19]
Parameter [Synthetic PSF Radius] set to [1.5 pixels]
Parameter [Synthetic Iterations] set to [Off]
Parameter [Spatial Error] set to [1.00]
Parameter [Deringing Fuzz] set to [20.0 pixels]
Parameter [Deringing Detect] set to [65 %]
Parameter [Dyn. Range Extension] set to [1.00]
Parameter [Linearity Cutoff] set to [85 %]
Parameter [Sampled Iterations] set to [2x]
Parameter [Deringing Amount] set to [0.80]
Mask used (BASE64 PNG encoded)
--- Color
Parameter [Bias Slider Mode] set to [Sliders Reduce Color Bias]
Parameter [Style] set to [Artistic, Not Detail Aware]
Parameter [LRGB Method Emulation] set to [RGB Ratio, CIELab Luminance Retention]
Parameter [Matrix] set to [HOO Duoband 100R,50G+50B,50G+50B]
Parameter [Dark Saturation] set to [4.0]
Parameter [Bright Saturation] set to [2.0]
Parameter [Saturation Amount] set to [200 %]
Parameter [Blue Bias Reduce] set to [1.00]
Parameter [Green Bias Reduce] set to [1.00]
Parameter [Red Bias Reduce] set to [1.25]
Parameter [Mask Fuzz] set to [1.0 pixels]
Parameter [Cap Green] set to [0 %]
Parameter [Highlight Repair] set to [5 pixels]
firebrand18
Posts: 28
Joined: Tue Feb 04, 2020 1:43 pm

Re: StarTools 1.8.512 public alpha/preview 4 now out

Post by firebrand18 »

You are not alone with the wild pixels; I recently processed 9 separate globular cluster stacks taken over the last several months using the Optolong L-Pro filter and intensive processing in release 512 with great results (love 1.8xxx). I first used SVDecon module in Synthetic mode (default) and no wild pixels appeared. As soon as I started sampling anywhere from 1 to 25 stars, a few random pixels sprouted in some images and not in others, and no amount of tweaking the options in the module removed them. Much better decon with sampling however so simply used the tried and true Heal to eliminate them (easily).

Curious as well to see what makes these gremlins show up. My guess is it depends on what stars are used for sampling and even though most are good candidates (green) shooting from heavy light polluted skies (Bortle 8) where I am, lots of background noise may cause variants bleeding into the stars that affects the stellar profile when spatial variant decon kicks in.

Really love what's going on with 1.8 and won't go back!

Clear skies
jackbak
Posts: 16
Joined: Sun Jun 14, 2015 3:38 pm

Re: StarTools 1.8.512 public alpha/preview 4 now out

Post by jackbak »

I am having a devil of time with masks (see below) I have binned the image to 71%, it is an OpenSuse Tumbleweed Current as of today,
with an Intel i7 with core GPU and 16Gb of Ram.

I saw this behavior in 1.7 but this is just maddening, if I try to manually generate the mask I get the same thing or worse.

Thoughts?
Attachments
Laggon SV PSF Mask
Laggon SV PSF Mask
StarTools_ SV_PSF_1_8_512_Mask.jpg (244.86 KiB) Viewed 111 times
User avatar
admin
Site Admin
Posts: 2866
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: StarTools 1.8.512 public alpha/preview 4 now out

Post by admin »

Wild pixels
Thank you for reporting (and sharing the dataset @Mike in Rancho!)

I have not seen any at all on Linux, so I'm wondering if it might be something specific to the Windows version... :think:

As "usual" I cannot replicate this on Linux (the reference platform/OS for ST). So I'm going to have to do some wider testing. Do they appear on the non-GPU version as well?
jackbak wrote: Sun Sep 12, 2021 5:04 pm I am having a devil of time with masks (see below) I have binned the image to 71%, it is an OpenSuse Tumbleweed Current as of today,
with an Intel i7 with core GPU and 16Gb of Ram.

I saw this behavior in 1.7 but this is just maddening, if I try to manually generate the mask I get the same thing or worse.

Thoughts?
I have seen this behavior on Linux when using the GPU version on an Intel iGPU. The Intel iGPU drivers are rather unusable :( If this is the case, then, frankly, I don't have any solution but to recommend the non-GPU version, try to add a well-supported discrete GPU, or to try different drivers if you can find any that will work with your distro. FWIW, I have personally had much better luck on Ubuntu 20.04-based distros, which seem to come with better Intel drivers.
Ivo Jager
StarTools creator and astronomy enthusiast
Post Reply