Ha-OIII color algorithm?

In the 0-tab where you select the Ha-OIII algorithm, this indeed takes out either the Ha or OIII data automatically which you then register, normalize and integrate. When done with one, you go back to the 0-tab, select the other and APP wil ask to re-normalize the subs and goes on to make the other integration. You then combine the integrations yourself.

Yes. I did this and extracted both Ha and OIII and successfully combined them along with a mono lum. My question is about the Ha-OIII color algorithm. I’m curious about what it is for. When I use this algorithm it appears to give me a composite of Ha and OIII that I can then separate into channels in the combine tool in order to map them to different colors. Is this correct or is this algorithm for another purpose?

Ahh I'm sorry, I misunderstood your question. So the color algorithm gives an RGB output as-is recorded by the sensor, not separated and processed as one. You can indeed adjust them later, but the "extract Ha" or "extract OIII" are there to get the Ha and OIII data out straight away and giving you more control on post-processing (you can then adjust these channels to your liking completely before combining). The "mono" is meant to give you a "luminance" result.

Vincent—Thank you very much. This completely answers my question. APP is great for us OSC users!

I agree! 🙂

@vincent-mod

Hi Vincent and Ahbok,

May I ask an additional question? What is the difference between the Hydrogen Alpha algorithm and the HA O3 extract HA?

Also what would be the difference between O3 algorithm an the HA O3 extract O3?

Is it a complete other algorithm (for example O3 an the HA O3 extract O3). Ore is it the same with a other name?

Hope you can give us a bit more insight?

Clear skies Jan-Willem

Hi Jan-Willem,

I think they are the same algorithm, not 100% sure but I can't think of why it wouldn't be really. In all cases APP gets the RGB data from an OSC sensor presented and it doesn't really matter what filter was used. It also doesn't know, which is why you manually select it. I think it's just a simpler way for users to pick the right one, it would confuse people if we only had the "OIII, extract Ha" as the only option to extract Ha.

@vincent-mod

Hi Vincent,

Thanks, so if I understand correctly (for example) the O3 algorithm and the HA OIII extract O3 are both the same? It is only for "user friendly" use that the "extract" algorithm was made?

If so couldn't it be possible to make a new "algorithm" that would automatically makes to stacks? 1 of Ha and 1 of O3? 

Clear skies Jan-Willem

I *think* so. 🙂 I'm not 100% sure, so for that I'll tag @mabula-admin.

@vincent-mod, @mabula-admin

Hi Vincent,

Thanks again, i hope Mabula can give us a bit more insight how this algarithm work and if it would be possible to make an extra algortihm that could make an HA and O3 integration in one session.

Cheers Jan-Willem

Hi All,

I agree with the "in one session" request. 

As for the explanation on the algorithms used, Mabula did a write up in the release notes for 1.068:. Note the section on the difference with the AAD algorithm, even for color images. I have found with the STC-DUO in that going straight for a color output is an efficient route. It gives good star colors and good nebulosity colors. Only when you are dealing with significant OIII contributions, you would look at doing it manually. Simply because then you would want to increase the OIII signal in comparison to the HA signal, because the HA signal is so dominant.

Below the write up:

Astro Pixel Processor 1.068

• ADDED/IMPROVED, Bayer CFA debayer algorithmes, I have added 4 new debayer algorithms, especially for so-called Multiple NarrowBand filters like the
  • STC Astro Duo Narrowband filter : 
    

     

  • OPT TRIAD Tri-Band Narrowband filter :  https://optcorp.com/products/opt-triad-tri-band-narrowband-filter

These filters enable you to capture narrowband data from Hydrogen Alpha and from Oxygen III at the same time with a One Shot Color camera (OSC = DSLR or Astronomical Color camera) camera. The OPT filter has a bit broader bandpass so it also includes Hydrogen Beta which is close to OIII.

Since you will capture data, shot with these filters, with your OSC camera in the Red, Green and Blue channels, there are different ways to process the data. You can simply interpret the data directly as it is recorded in RGB (Ha-OIII color). You can interpret it as luminance data, so monochrome (Ha-OIII mono). You can even directly extract the Hydrogen alpha (Ha-OIII extract Ha) and Oxygen III (Ha-OIII extract OIII) data directly from these exposures, which will give you much more control in post-processing. The purpose of the new debayer algorithmes is to enable you to do this directly  by choosing the suitable debayer algorithm:

The 4 new debayer algoritmes are:

1. Ha-OIII color
2. Ha-OIII mono
3. Ha-OIII extract Ha
4. Ha-OIII extract OIII

The Ha-OIII color algorithm is an improvement of processing data shot with these kind of filters, over the default Adaptive Airy Disc (AAD) debayer algortihm.

The OIII data quality and sharpness is the same, since the AAD algorithm is perfect for the OIII data that has data in both the green and blue channels.

For the Hydrogen alpha data however, the AAD algorithm is not suitable. The HA-OIII color algorithm will treat the Hydrogen alpha data totally separate from the OIII data and will therefore give a much better reconstruction of the Hydrogen alpha data versus the AAD algorithm. The noise in your H-alpha data will be much lower ;-). In integrations of several frames the Signal To Noise Ratio will be much higher as a consequence.

Thank you very much for this. I am primarily interested in combing NB data with the RGB data taken with my OSC. It appears that using the Ha-OIII color algorithm will be an efficient way to get the NB data  to integrate with my RGB data.

Hi, just following up on this topic, despite it being 5 years since the last post! I use an OSC, normally in Bortle 3 conditions, so with only UV/IR filter.  Sometimes however, I use an L-Pro filter to bring out more Ha & OIII on some targets, as well as when there is a moon - but am unsure how best to process the mix of shots with UV/IR and L-Pro filters. Is it best to:

1. Simply combine all using AAD (tab 0) and call them all RGB?

2. Process the L-Pro shots as some other channel (which?) in parallel with the UV/IR shots?

3. Use one of the Ha-OIII Color or Ha-OIII Mono alternatives instead of AAD in Tab 0? If so, which one???

4. Follow the now well-described process for using Ha-OIII extract Ha (and extract OIII) to get separate channels?

If the suggested answer is #3, what do I do afterwards? Do I process the UV/IR filter shots separately then re-load both that processed file and the Ha-OIII color result for the L-Pro filter, and re-integrate them? Or something else?

Tx again!

Hi Mark @mfc,

Thank you for your question.

I would advise to load the Uv/IR filter data with a corresponding filter/channel description with Multi-Channel/Filter processing enabled in 1)Load. Same for the L-Pro data. If you shoot targets with both filters, the results will be a bit different as you will have experienced, since Uv/IR will let the RGB spectrum come through fully, whereas the L-Pro filter will block parts of the RGB spectrum to reduce light poillution in your data, right?

For both filters, simply use the AAD demosaic algorithm, that is perfectly fine.

In 6) Integrate, simply let APP create a stack per filter and also one for the 2 filters combined with the setting at the top op 6) Integrate integrate per channel & all.

Depending on the object, the stack with both the UV/IR data and L-Pro data might look nice/okay and an improvement over the single filter results.

But... it could also happen that the combined stack of the 2 filters looks a bit weird. I guess you have seen that happen? In those cases, I think the best solution will be to use the UV/IR data as RGB data, which it really is. And use the L-Pro data as a luminance channel, which is usually the best for light pollution filters since their RGB interpretation will be off.

You can combine the RGB/UV/IR data with the L/ L-Pro data in the RGB Combine tool. Simply assign the RGB data per channel as R,G,B and when you load the L-Pro data, assign it only as L and then see how things will look if you play a little bit with the L sliders. Normally you want the R channel of the L-Pro data to contribute like 20% to L, green about 70% and blue only 10%.

Let me know if this helps and if you have more questions about this.

Mabula

@mabula-admin thanks for your (as always!) comprehensive response. You’re correct that I have had some unusual results when treating L-Pro files the same as UV/IR or no-filter files. But I assumed it was just my technique that was the issue. I’ll try what you suggest on my latest targets and give a response later on…

Great Mark @mfc, you are most welcome !