HaRGBCombination script installation

[sixburg]

Do the AIP-related multi-channel processing scripts supercede HaRGBComination or are they just alternative approaches to the same problem?

Secondly, how in the heck do you load the HaRGCombination script?  I downloaded a version (no idea if it's the latest or not as I don't know of the existence of a master database of scripts), copied it to the PI /src/scripts directory, used Scripts | Feature Scripts, pressed Add and selected the proper directory to which PI says zero additional scripts found on directory. 
What's the secret?

Thanks.

[jerryyyyy]

You may have to execute the script. 

C:\Program Files\PixInsight\src\scripts

You will find Harry's in there and some of the others.  If if you install, would put them there.  Some seem to be able to be added to the menus, but this is an obscure magical process to me and does not always seem to work. 

You should also find some under the Scripts mutlichannel synthesis subdirectory.  I call these the French ones. 

[sixburg]

Thanks.  Yes, execution under script | execute script file seems to work.  I thought there was a way to add it to the file menu. 

[sixburg]

While I can get the script to execute I find that it does absolutely nothing.  After making a correction to line #128 from "true" to "false" that was suggested (I don't know on what date it was suggested, but found it in one of the forums and I tried it anyway), all I end up with is a blank screen called "haemission".  Without changing line #128 the script doesn't even execute.  What am I missing?

[pfile]

do we know if this script works right with 1.8RC7?

where's the thread with the script?

thanks

rob

[sixburg]

http://pixinsight.com/forum/index.php?topic=3401.msg23690#msg23690

Based on what I'm doing it doesn't work in 1.8RC7.  This thread is a couple of years old and I've no way to telling what is 'state of the art'. 

[topboxman]

That's pretty old. Try this one:

http://pixinsight.com/forum/index.php?topic=3446.msg40130#msg40130

Peter

[Ioannis Ioannou]

Download the latest (Peter pointed to it above), unzip and save it somewhere (I avoid the official directory to prevent loss of my scripts during major upgrades, I use a separated personal folder for my scripts), add the folder with Scripts->Feature scripts ->Add, and it should appear under Scripts -> Utilities. Mind also that this is a user-contributed script, not an official supported, so no central repositories or official support or similar.

[sixburg]

Thanks for the inputs.  I have run two scripts (NARGB and HARVP-AIP) in an attempt to combine ha with RGB.  The texture and detail of the ha is hard to resist.  However, neither of the methods underlying these scripts quite get at what I'm looking for, but the AIP is much closer.  As I understand it and from what I can see the NARGB adds the ha to the R channel and recombines so you have (HaR)GB.  I've seen and experimented with this method before and while it works pretty well to boost the red in an image it has not captured the detail in the ha. 

My example below is 1 of an 8 panel mosaic.  The images on the left are only stretched (RGB received standard processing with BN-CC-DBE) and in the case of the ha (lower left) treated with HDRC-HDRMT-LHE.  They are the source images for each algorithm.  Working with these data as linear didn't get good results at all.

The upper right is NARGB with defaults and the lower right is AIP with 15% Ha in R and 15% R in Ha (approximately).  While the color balance doesn't fit my taste for the AIP method the nebulosity detail of the ha shows up very nicely.

Is this result consistent with what you've experienced? 

[topboxman]

I thought you are supposed to run the script right after RGB combine and before any processing like DBE-BN-CC. In other words, do not process of any kind to Ha and RGB when running the scripts like the following.

1) Calibrate all raw and unprocessed images
2) Star alignment of all calibrated images
3) Image Integration of star aligned Ha, R, G, & B images
4) RGB combine of Image Integrated R, G, & B images
5) Run NARGB or HARVP-AIP scripts using Image Integrated Ha and combined RGB images

Peter

[sixburg]

Perhaps you're correct.   However, my experience says otherwise.  See the attached for the results.  NBRGB results are similar or the same for linear or non-linear  But, for experiment's sake, here are the results with calibrated and combined RGB (color uncalibrated), calibrated Ha...no other processing at all. 

Note that NARGB results are identical to the prior post (standard STF, RGB linked).  Therefore, it seems to not matter whether or not the images are linear.  The AIP processing is a mess.  Maybe I did it wrong (even though it's pretty straightforward), but in either case, I'll stick with non-linear unless someone can demonstrate a superior method. 

I'm on my way to collecting what will be in the neighborhood of 40 hours of ha and ~16 hours of RGB for a haRGB mosaic...I intend to learn the proper way to combine for the effect I'm looking for as in my last post.  The idea is that my understanding of processing the combination will coalesce before I use all that time for this one project.  Worst case I have a great ha mosaic. 

If only I could treat the Ha as Lum to get all that detail...however, I don't believe the standard LRGB method will work using Ha as luminance (as has been amply documented in these forums).  My understanding is that luminance effectively "lights up" and provides detail to the underlying RGB.
 
Much of the interesting nebulosity present in the M42 area is prominent in Ha, but does show up as much in RGB.  So, there's effectively no RGB in many places for the ha to "light up".  This is the understanding I've gained here.  Please correct me if I'm wrong. 

I will post another experiment on LRGB with the same data which will attempt to demonstrate how well, or not LRGB works with Ha as L + RGB and Ha as L + R +G + B.  I believe these provide very different results.  That's my hypothesis which I believe has already been demonstrated here.  Results to follow shortly...

[sixburg]

Nope!  Doesn't matter whether it's L+R+G+B or L+RGB (L=Ha in this case). 

[sixburg]

FITs for anyone interested.

http://sdrv.ms/17wZbQC