Astro Pixel Processor work flow for use in StarTools?

[PaulInNorthMichigan]

I am new to StarTools so I don't have a really good processing order determined yet.

I use the Astro Pixel Processor stacking software. I don't see a whole lot written on the do's and don'ts associated with preparing an APP image for use in StarTools.

Presently I have been using all of the defaults between Step 1 and 6. I have been using the un stretched integrated image after step 6 is completed.

Ivo assisted me in deciding to go with StarTools due to the good results he helped me obtain in processing Andromeda with StarTools. In his recommendations he suggested that Normalization (which is step 5 in APP) should be avoided. I tried avoiding Normalization but as my image was being processed I could see that the Normalization was being applied at step 6 Integration. In searching the APP forums I saw a thread in which Mabula from APP said Normalization could be avoided through unchecking a setting but that this would not be recommended.

That is just one of the questions which I have not found a resolution to yet.

Another question which I have not resolved yet is whether I should or shouldn't apply Gradient reduction from step 9 in APP. The Gradient Reduction tool in APP is top notch. This is a manual setting which you exit when you have the desired effect. I have not been applying Gradient Reduction since after my Andromeda project. Since that time I have one project in particular in which I used a UHC filter that I could not remove the red cast to the stars. Here I am wondering perhaps whether I might find better results with working out APP's manual Gradient Reduction rather than using StarTools almost automated Gradient Reduction.

Ivo pointed out some anomaly's from my Andromeda image that I believe he suggested could have been injected from my processing of Andromeda in APP. The suggested work flow in stacking prior to StarTools that I've heard is to minimize anything that might deviate the image from it's natural state.

I'm trying to filter out what settings would be beneficial applied in APP and what settings would not. Are there any experienced APP users that might share their work flow?

[admin]

This is a great questions and I'd really love to hear from some APP users.
I wish there was some documentation for this application, as this would really help nail down the right settings. Right now it's a bit of guessing game as to what goes on with the signal and what each option does.

Ideally, you would want to end up with a stack that, on first AutoDev inspection, looks very similar to a DSS or PI stack in terms of "in your face" characteristics. E.g. for most DSLR or OSC data it should show a strong green, teal or blue bias (depending on filter characteristics) with some of the brighter stars (and possibly the brightest nebulosity) poking through. E.g. something like this;

integration.jpg (317.08 KiB) Viewed 23698 times

or this;

m31-integrated.jpg (59.17 KiB) Viewed 23697 times

(note you can only see this green/teal color in AutoDev if choosing the first option when importing your stack, otherwise it's just mono. Importing such an ideal/good/virgin stack with the first option is of course the opposite you'd want to do - you'd really want to use the second option which separates reweighted luminance and color, but this won't show the green bias for the purpose of this exercise!)

If anyone knows how to reliably accomplish the above in the latest versions(s) of APP that would be amazing!

[hixx]

Hi,
These are the settings in the APP tabs I used successfully for stacking:
0) / RAW/FITS: Bilinear or Adaptive Airy Disk (I did not see a disadvantage of AAD vs Bilinear in StarTools) AAD seems to produce a slightly better result.
1) / LOAD: default settings
2) / CALIBRATE: default settings except: disable "adaptive pedestal / reduce Amp glow"
3) / ANALYSE STARS: default
4) / REGISTER: default settings
5) / NORMALIZE: default settings except: disable "neutralize background" - selecting "none" in method dropdown should disable normalization but I did not see any improvement for Star Tools operation and result. Probably Normalization is mostly impacting Outlier handling so I would not recommend the 'none' setting.
6) INTEGRATE: default settings. Though you may use Local Normalization Correction (1st degree,2-3 iterations) and Multi Band Blending (if You need to stack images from multiple nights). That would not provide the real raw data StarTools would operate on best, but I found this "unlinearity" not too huge for StarTools Tracking to derail on it.
9) Tools: Do not use this tab - these tools are creating strong unlinearity and Tracking will have an issue with these. Use the ST Wipe tool to do background color calibration and light pollution removal instead.

It would be really great if APP information on LNC and MBB could be passed into ST "Tracking" somehow. This would enable StarTools to leverage the really advanced features of APP stacking.

Hope this helps

[PaulInNorthMichigan]

Ivo,

That does sound like a good check to try in experimentation by temporarily using the 1st import option. This would be similar to testing flat frame exposures for their effectiveness. I could perhaps try some different settings to see which methods best achieve the blue, green or teal coloring.

APP isn't a perfect match with StarTools but I'm sure there must be some best combination to obtain repeatable quality. I already have been very happy with a couple projects, it will be a joy to see similar results with my other images.

Thanks for your idea for me to help filter out good and bad results! Perhaps there will be others who could help in establishing a good stacking procedure with APP through their trial and error.

[PaulInNorthMichigan]

Hixx,

Thanks so much for sharing your stacking procedure with APP! I appreciate your number by number settings listing.

I most certainly will process my next project using these settings.

That is good that you are not seeing a derail with applying settings such as LNC and MBB when stacking a combination of multiple night sessions.

I will trust you in the fact that I won't need to use APP's light pollution removal tool in number 9. Using the StarTools equivalent instead.

I have one image in specific (the Veil Nebula) that I used an UHC filter that I can't remove the red cast. I will try your setting suggestions and see if it achieves a closer to expected result.

[admin]

Hi,
These are the settings in the APP tabs I used successfully for stacking:
0) / RAW/FITS: Bilinear or Adaptive Airy Disk (I did not see a disadvantage of AAD vs Bilinear in StarTools) AAD seems to produce a slightly better result.
1) / LOAD: default settings
2) / CALIBRATE: default settings except: disable "adaptive pedestal / reduce Amp glow"
3) / ANALYSE STARS: default
4) / REGISTER: default settings
5) / NORMALIZE: default settings except: disable "neutralize background" - selecting "none" in method dropdown should disable normalization but I did not see any improvement for Star Tools operation and result. Probably Normalization is mostly impacting Outlier handling so I would not recommend the 'none' setting.
6) INTEGRATE: default settings. Though you may use Local Normalization Correction (1st degree,2-3 iterations) and Multi Band Blending (if You need to stack images from multiple nights). That would not provide the real raw data StarTools would operate on best, but I found this "unlinearity" not too huge for StarTools Tracking to derail on it.
9) Tools: Do not use this tab - these tools are creating strong unlinearity and Tracking will have an issue with these. Use the ST Wipe tool to do background color calibration and light pollution removal instead.

It would be really great if APP information on LNC and MBB could be passed into ST "Tracking" somehow. This would enable StarTools to leverage the really advanced features of APP stacking.

Hope this helps

Fantastically useful as always!

The #5 part about normalization is particularly enlightening. I have indeed been worried about having no normalization happen at all for the purpose of outlier rejection - disabling the neutralize background option is I think/hope the missing piece.

From the few datasets I have seen, Local Normalization Correction appeared to be quite destructive to faint signal, while missing (and thereby "introducing") faint medium scale detail discrepancies.

The gradient modelling algorithm in StarTools is quite a bit different (by design) from the arbitrary manual sample-setting in PI, APP, Siril, etc.
The latter methods are little bit like using a shovel for an archaeological dig. It may be intuitively faster to use something that "attacks" what you see, however it can/will badly damage anything delicate that lies underneath.

Wipe hands you a brush, not a shovel - it is overcautious and gentle. It exploits the fact that actual real celestial gradients (and vignetting) have a minimum frequency (governed by the Aggressiveness parameter) by which they can undulate across the image. If a detected frequency is found the be below this minimum, Wipe determines that the "gradient" must actually be nebulosity/detail and not a celestial gradient or vignetting. As a consequence, it will to its utmost to preserve it. Arbitrary sample setting without knowing your target well is how you would completely kill the faint signal from, for example, Integrated Flux Nebulosity found around some galaxies. Similarly, you may introduce something that looks like IFN but is just the result of poor sample setting.

Adaptive Airy Disc debayering under #0 is something I really like about APP; it is innovative thinking specific to astrophotography. We need more of that!

I'd have to read more about MBB to see what it does... Does anyone have any pointers to documentation, video or a forum post?

[richbandit]

Hi there!

I've not made many posts on the forum so far, so this is an opportunity for me to address that.

Great advice and workflow from hixx, with a few 'exceptions' I'd like to add to the mix which might help to assist someone.

Brief history - I started in astrophotography around the beginning of last year, and REALLY struggled to get my head around things. You name the software, I tried it, all the usual and well known suspects, I couldn't get any of them to work for me as I liked. I then discovered APP and Star Tools at about the same time and 'bingo', that was it - a terrific combo that just worked together! With a little assistance from Ivo (thanks Ivo!) and also Mabula over at APP, I was finally off and running ...

Unable to post an image it seems, otherwise I'd provide a humble attempt by me of what APP and Star Tools can achieve together!

So, to explain my 'exceptions' :

- If using a mono 'astro' camera (I have ASI183mm Pro Cooled) then I believe hixx's point #0 doesn't apply, as it's intended for DSLR and /or colour camera users. I don't think I've ever actually looked at that tab before!
- Because the 183mm is a CMOS chip, I MUST enable the 'adaptive pedestal/reduce amp glow' - it DOES make a difference due to the severe amp glow present on this (and similar) cameras.
- I totally agree with hixx's point #5 - definitely deselect 'neutralise background'. Not only does using this seem to have little effect, but on occasions I've also noticed some loss of detail .... so don't use it.
- Finally, and perhaps surprisingly - I generally find that applying 'light pollution removal' on the 'tools' menu actually does seem to work for me! I would have also suspected that this might have a subsequent effect on Star Tool's tracking or it's own light pollution removal processes, but actually I tend to find it cleans the whole image up nicely. If you do use this feature, only run it through once though, and be very careful where you place the selection boxes (away from any objects of interest) - then save the image and move straight on to Star Tools to carry on ....

The latest APP version also now has an 'automatic' option to select on the 'integration' tab, which instructs the software to analyse the image and apply what it believes to be the optimum settings for LNC, etc. during integration. I must say, for me this 'auto' setting has worked brilliantly - as usual I guess YMMV, but it's certainly worth giving it a try.

Hope this helps someone, happy imaging!

[PaulInNorthMichigan]

RichBandit: Thank you for sharing your experience between APP and StarTools.

Presently I am starting up a processing per hixx shared settings. If my issue with background persists I may again try applying APP's Gradient Reduction tool in which you've seen good results when using.

[PaulInNorthMichigan]

I haven't fully processed yet in StarTools my modified APP processing of the Veil Nebula but there is a dramatic coloration difference when applying Hixx setting recommendations and using the display from the 1st import option as Ivo suggested.

Notice the difference between images using the 1st import option and opening it up in Auto Develop. (The modified image uses hixx recommendations.)

Veil Default.jpg (441.21 KiB) Viewed 23659 times

[richbandit]

Hi Paul,

Only use APP's 'remove light pollution' on the tools menu, nothing else on that tab really benefits over what Star Tools can then subsequently offer - works for me, at least.

That said, I have on occasion used APP's 'calibrate star colours' and 'combine RGB' tools when I have some data that appeared to be difficult (not sure why Star Tools didn't like it?) with some success. As always, it's a question of trying different options to see what works for you ....

At the end of the day, nothing seems to beat Star Tools for providing the most straightforward route to achieving a great final result! I also tend to finish off with a little Photoshop curves/levels tweaking as well (and Astronomy Tools actions plug-in is also very useful).

All we need now are a few decent clear nights after what seems an age of cloud/rain (in UK), well for me anyway!

Good luck!