Combining OSC with Red only (Ha) from L-extreme

[dx_ron]

Hi

I have stacked 228' of OSC data for the Cocoon Nebula from when there was no moon interfering. Then, with the near-full moon, I collected 312' of data using the L-extreme filter. I am combining them, using the OSC+Ha guide from Guy's 'Compose Module use' post. I registered the L-extreme stack to the OSC stack before proceeding.

I made a mono file from the Red channel of the OSC stack
I made a mono file from the Red channel of the L-extreme stack

When I combined those, I set the Red exposure to 228' and the Green/Ha exposure to 312'. Was I right in setting those weighting factors?

Now, when I go to load this new combined OSC+L-extreme Red channel along with the OSC Blue and Green channels, should I weight the exposures again? Red=540' / Green and Blue = 228'?

I don't fully understand the effect of these weights on later processing. Well 'not fully' is false bravado - "I have no clue" is more accurate.

Thanks,
Ron

[Carles]

Hi Ron!

As I also got l-extreme recently, I'm very interesting in this topic!
by the way, when you make the mono files from Red and H-alpha, what "luminance, Color" mode do you select?
it is something that is missing in the manual /user notes from @Guy and not sure about it..

Tried playing with it, but never touched the exposure times and no idea what they would do.
@admin could you give some hints ?


Regards

Carles.

[Mike in Rancho]

Me too!!!

I have a substantially similar post started the other day in the image processing subforum, hoping for clarification.

Indeed the instructions, website, Guy notes, etc., are all not quite fully transparent as to what we are doing here with the settings and combinations, and why, so that we can make informed choices. So I've just been trying different permutations of everything, but it's hard to tell differences apart.

For one small part of it -- I am moderately sure that the exposure time settings in compose are only used to weight the channels for creation of the synthetic L. And that giving less exposure time to a channel relative to the others will make that one be weighted up.

That said, I think there are other tricks involved here so it's not the whole story. Wasn't it recently mentioned that setting a channel to zero exposure time will utilize it for detail/luminance purposes (though at what weight I am uncertain) but not for color? Or is that backwards?

The manual might need to be 500 pages...

[admin]

I don't fully understand the effect of these weights on later processing. Well 'not fully' is false bravado - "I have no clue" is more accurate.

Thanks,
Ron

The weighting is purely for the purpose of luminance generation;

Code: Select all

Luminance = red * sqrt(exposure time) + green * sqrt(exposure time) + blue * sqrt(exposure time)

...or in the case of an OSC/DSLR with a Bayer matrix (giving us twice the green samples for every blue and red samples);

Code: Select all

Luminance = red * sqrt(exposure time) + green * sqrt(exposure time * 2) + blue * sqrt(exposure time)

Signal (versus noise) improves by the square root of the amount of exposure time (or the amount of frames stacked, etc.)

The point of the weighting, is to establish the exact blend that yields the cleanest possible luminance signal and the lowest amount of noise (this is also the main reason why processing a white balanced stack yields a much noisier result, as the multiplication factors used have no bearing on the actual quality of the signal in the three channels and were purely used for the purpose of coloring).

I have stacked 228' of OSC data for the Cocoon Nebula from when there was no moon interfering. Then, with the near-full moon, I collected 312' of data using the L-extreme filter. I am combining them, using the OSC+Ha guide from Guy's 'Compose Module use' post. I registered the L-extreme stack to the OSC stack before proceeding.

I made a mono file from the Red channel of the OSC stack
I made a mono file from the Red channel of the L-extreme stack

When I combined those, I set the Red exposure to 228' and the Green/Ha exposure to 312'. Was I right in setting those weighting factors?

You are mixing very different signals here. The L-extreme is a narrowband filter, so you if you are using the red channels only (there is no reason for that really - the OIII is useful too!), you are mixing Ha and wide-band red. Adding these together could be used for luminance, however you will get rather strange coloring if you also use the "new" red channel for coloring. Wideband green and blue, combined with some strange mix of Ha and wideband red will yield strange color casts due to the relative weakness of the red (since "normal" red is attenuated in areas where there is no Ha).

So that's a problem. However, if you really want to combine narrowband and visual spectrum data, I would create mono files from both stacks (including all channels, e.g. just use two sessions of the Compose module to load each stack into Luminance, using the "L + Synthetic L Frin R(2xG)B, Mono (Mono from OSC/DSLR)" setting. Save each session. You will now have two synthetic luminance datasets.

You can then create a blend of the two synthetic luminance sets in the Compose module ("abusing" the two of the red, green or blue channels and using "L + Synthetic L From RGB, Mono") or Layer module (calculate blend amount; 228 / (228+312) = ~42% red vs ~58 Ha) according to their exposure times.

You now have a single new synthetic luminance. Now combine this with the coloring of the just the visual spectrum dataset and you should have the "best" of both worlds without any strange color skew. "Best" between quotes because, though you are making full use of the signal, you are attenuating detail outside of the Ha spectrum. This may or may not be desirable. It can be rather disastrous in objects that are Ha poor (say a galaxy) for example.

Of course, there are a million other ways to make use of narrowband data, on of which being the NB Accent module.

The question is to what end did you record the data (and that specific amount) that you did?

[Mike in Rancho]

Likely it was done for the CN monthly challenge on the Cocoon, as I and quite a few others were doing also.

In my case I obtained 5 hours of RGB, and later 3 hours through my L'eNhance. I stripped the red from the filter data. The Cocoon is not much of an OIII target. It may be in the stars, but I didn't want to risk turning them all blue/teal tinted. So I left whatever OIII I got out.

My first try was to use the Ha as L, but that turned the cocoon way too pink for my liking no matter how I tried to massage it. Ultimately I stripped R from RGB and combined it with Ha to make new R, also using the Guy/manual steps. This seemed to add nice extra cocoon detail, and after balancing back the red strength the star colors seemed normal, even if the nebula was red heavy. I am still uncertain quite what the steps in the instructions created though, as noted. Maybe LLHaRGB? lol.

What I was really hoping my filter data would acquire is some deep Ha structures in the region just adjacent to the cocoon, but alas, I didn't manage to get that.

[admin]

Maybe LLHaRGB? lol.

We seriously need a (new?) nomenclature.
I have actually been thinking about this in earnest, given StarTools' formalised approach to narrowband, luminance and chrominance, as well as the proliferation of duo/tri/quadband filters these past years.
It would really help people understand what they are looking at and how the image came about.

A simple three-comma separated system would do;

L,RGB,Accents.

In your case (if I understand correctly) that would be;

HaRGB,HaRGB,

[Mike in Rancho]

We seriously need a (new?) nomenclature.
I have actually been thinking about this in earnest, given StarTools' formalised approach to narrowband, luminance and chrominance, as well as the proliferation of duo/tri/quadband filters these past years.
It would really help people understand what they are looking at and how the image came about.

A simple three-comma separated system would do;

L,RGB,Accents.

In your case (if I understand correctly) that would be;

HaRGB,HaRGB,

I dunno, that nomenclature confused me also. We'll be needing more documentation.

Mostly though I'm just trying to figure out what I'm doing, what ST is doing, and whether it makes any sense to be doing so. Now, I did get a better image, but why, and was it an accident?

One part I'm confused on in the guide steps is tossing the original OSC file into L. I presume that turns it into a grayscale luminance based on all channels. Perhaps the same way ST would create the synthetic L if we just put the OSC file into R, G, and B and kept the default L (of which there is none) + synth L? BUT, on top of that, the final composite, again using the default L+synth L, presumably makes said synth L out of newR, OSC, OSC.

Whether this is useful or not I guess depends on just what ST is doing when it has both an L and an synth L, and that I'm not sure of.

Overall I guess, there's a little clarity missing when it comes to the settings in compose, particularly what to choose for a) Luminance, Color, and b) interpolation. Many of the steps for various techniques in the manual and user notes are silent as to one or both of those.

Though I think I have a better handle on interpolation setting now after playing with it some, even though it seems counterintuitive from the documentation (i.e. ST can interpolate missing channels....).

Anyway, not critical, just something that even after a fair bit of usage and trying to read up about it, I still don't have a handle on.

[dx_ron]

Mine would be a Ha,Ha

And yes, this is a widefield view of the Cocoon, so a small fraction of the image is the Ha region, the rest is a rich star field. My understanding is that there is little-to-no Oii. I shot the L-extreme data under a full moon, so my thought was that including the G and B from that would add a bunch of gradient.

I can see that the OSC+Ha leads to trouble getting the color balance right. While the stars actually do have "correct" colors, I'm not sure anyone is going to look at my image and check the color temps.

Here's a compressed jpg of one of my versions with the Red L-extreme channel mixed in

Cocoon_OSC_plus_L-extreme_ST_1600.jpg (466.79 KiB) Viewed 6733 times

[admin]

Mine would be a Ha,Ha

Took me a while to process that, but ended up with L,O,L

And yes, this is a widefield view of the Cocoon, so a small fraction of the image is the Ha region, the rest is a rich star field. My understanding is that there is little-to-no Oii. I shot the L-extreme data under a full moon, so my thought was that including the G and B from that would add a bunch of gradient.

That's a fair assessment of course.

I can see that the OSC+Ha leads to trouble getting the color balance right. While the stars actually do have "correct" colors, I'm not sure anyone is going to look at my image and check the color temps.

The simple solution is to not use Ha in your coloring at all, and either add it to your luminance or use the new NBAccent module in 1.8.

[Guy]

Hi Ivo,

I'm trying to capture the technique you describe here and I have a few questions.

- If I just load an OSC/DSLR file into the luminance channel (and nothing into R,G or B) and set 'Luminance, Color' to 'L + Synthetic L from R(2xG)B, Mono (Mono from OSC/DSLR)' does it extract the RGB from the OSC file to create the synthetic luminance data?

- I do this with the wide band OSC file and with the filtered Ha data file (presumably also from the OSC) and create luminance data files.
- If I use the Compose module to combine the two files - say I use R channel for Ha & G channel for Wideband and I set set 'Luminance, Color' to 'L + Synthetic L From RGB, Mono':

• Do I set Channel Interpolation to Off since we don't want a synthesised B channel adding to our synthetic luminance ?
• Do I set 'Red total exposure' to 312 and 'Green total exposure' to 228 and 'Blue total exposure' to 'Not Set' to get the weighting of the two sets of luminance data correct?

So I then have the luminance file and load it into Luminance channel and load the wideband OSC file into the R G and B channels individually.

• Do I set 'Luminance, Color' to 'L,RGB' and leave exposure times to 'Not set' as R,G and B all have same exposure time?
• Is there any sense in setting 'Luminance, Color' to 'L+ Synthetic L from RGB,RGB' since we have already extracted the synthetic luminance from the RGB?

One final unrelated Compose question:
- If I set total exposure time in a channel to zero it ignores the channel when creating synthetic luminance data - useful if you want to put the same (narrowband) data into 2 channels (e.g. OIII/Hb in G & B)- is that right?

Thanks

Guy