SPCC and Monochrome Imaging (LRGB, LRGB+Ha, and Narrowband???

Comments

• BlockHead

  December 2022

  You are combining two different things. 

  SPCC is designed to calibrate the ratio of  fluxes (intensity of light) of the wavelengths you observe based on your filters. 

  So when you observe with a RED filter of a given bandwidth and transmission- the calibration gives you the proper flux ratio (compared with other filters) based on GAIA data as a reference and your input filter parameters after the the correction. 

  
  

  This also happens with any narrowband filters. 

  
  

  In the case of RGB filters.. the ratios once calibrated can be assigned ratios based on a desired white reference.

  
  

  With narrowband filters only the first part happens. You get the accurate astronomical ratios between the filters. This is the true line strengths of emission. If the nebula is emitting strongly in H-alpha that line will be brightest compared to the SII or OIII.

  
  

  This is what the narrowband option gives you. Nothing more.

  There is no "white balance" step for narrowband. There is no other predefined blending for the results. That is up to you. I do not think *anyone* takes the true emission line strengths and makes a narrowband image. Everyone boosts the weaker channel (lines). 

  
  

  Thus the NB "mode" of SPCC is a little misleading in my opinion. It doesn't do anything other than allow you to specify NB filters. That is it. There is an "optimize" mode that takes into account the width of the NB filters and attempts to curtail excess continuum color for wider NB filters. I didn't understand your point concerning setting up NB filters for either RGB,SHO etc. This isn't an option?

  
  

  You would not "boost" the Blue channel using SPCC. The point is to attain the color ratios. 

  
  

  Concerning your results with the monochrome imager and Optolong RGB filters- if the results are rough... and the published values are correct- you might want to consider quality control on the filters is suspect. 

  Do you have a good example?

  
  

  -the Blockhead

  
  
• Wade Hilmo

  December 2022

  Adam,

  
  

  Is your comment that the NB mode of SPCC "doesn't do anything other than allow you to specify NB filters" accurate?

  
  

  I took some recent data for IC410 and used ChannelCombination to combine my Ha data into the red channel, the OIII data into the green channel, and the OIII data again into the blue channel.  The initial result was far from reasonable in terms of color accuracy.

  
  

  I then ran SPCC in NB mode, specifying HOO with 3nm for each of the NB filters (which is correct for my filters).  The result looked exactly as I would expect with correct color calibration.

  
  

  I have attached the calibrated image for reference.  This image was processed with MureDenoise, ChannelCombination as above, SPCC in NB mode, and a slightly dimmed STF autostretch.  Nothing else was done.

  
  

  
  

  

  

  

  ic410_spcc_nb_test.jpg

  1092 x 849 - 85K
• George Yendrey

  December 2022

  Blockhead,
  I don't think your description of what SPCC does is entirely accurate.  In the limited documentation it mentions entering the filter information in NB RGB to match the palette (RGB, SHO, HOO for example).  With no prior data entered, when the NB option is clicked for the first time, it displays an HOO palette filter assignment in the filter information boxes.
  
  With the excpetion of attempting to create a straight RGB image from NB data, I can sort of agree there is no white point.  Or least not one that corresponds to the 'standard' RGB white point.
  
  Maybe it does nothing with the data as you suggest, but there seems to be a lot of programming effort involved in the GUI if nothing is being done with data.
  
  I did note that the graphs provided by SPCC differed from palette to palette.  When provided an HOO image (with HOO filter data+ camera data), the only data graphed was the R/G graph.  All the graphs were very tight lines.  Which indicates, if I understand Adams description correctly, that SPCC IS creating a white point reference indicated in the graphs.
  
  For NB data/filters assigned for RGB and SHO palettes, both R/G and G/B graphs were created.  When I performed this test, no pixel math algorithms were used to weight the NB into the RGB channels.  I dropped the respective NB masters into the ChannelCombination tool based on the palette output (RGB, SHO, or HOO) desired.  If you include Wade's results above, it seems that SPCC is doing "something", which may or may not be more than just calibrating flux...
  ???
• Wade Hilmo

  December 2022

  Adam and I had an off-board discussion about this topic, after which I went off to do some investigation.

  
  

  My response and image above were the result of a cursory examination of the tool, with the results of applying it to the data that I had on hand.  I've gone back and done some reading, some experimentation and watched some videos.

  
  

  Before getting into my findings, I wanted to express my appreciation for the Fundamentals videos on this topic.  In particular, the Fundamentals: Color Calibration : SPCC The Process is excellent.  It is - by far - the best explanation of the math behind this kind of calibration that I've ever seen.  Before watching it, I had a general idea of how the math could be done, but now I feel like I understand it at a much deeper level.

  
  

  At this point, I believe that everything that Adam said in his response above is accurate.  I'm not sure that I initially understood his points.  I'm going to reply in my own words, but I think that I am now in agreement with him.

  
  

  With that out of the way, I want to start by saying that I believe that SPCC in narrow band mode is doing something interesting.  Further, I believe that if done correctly, the color - specifically in the emission nebulae - is accurately calibrated.  And I'll note that I didn't quite do it correctly for the image that I posted above.  More on that below.

  
  

  The data that I used here includes only the Ha and OIII that I collected.  When you think about it, there is an implication here that color calibration for the stars can never be accurate.  As a simple thought experiment, consider a blue star.  To be blue, the blue channel must be dominant over the red and green channels.  With an HOO mapping, like I have here, blue and green will always be equal.  That prevents any possibility of getting a blue star in the results.  So we know for certain that SPCC can never accurately calibrate a full image with HOO.  I'll mention SHO mapping in a few minutes.

  
  

  I would like to present some quotes from the relevant documentation on SPCC, which you can find here.

  
  

  The narrowband working mode in PCC has been discontinued since the narrowband emission lines can be calculated with much more accuracy in SPCC. This is because SPCC calculates the brightness of each measured star exclusively within the small interval of wavelengths transmitted by narrowband filters. Thus, we no longer need to rely on imprecise star brightness comparisons between broad and narrowband filters.

  
  

  My takeaway from this, particularly the last sentence, is that SPCC can use the data from the GAIA catalog to determine the Ha and OIII contributions for included stars.  This ratio is even more accurate than what you get in RGB mode, because the behavior of a narrow band filter is quite specific.

  
  

  When you run SPCC in narrow band mode, it removes the option to specify your filters by name.  Instead, it allows you to specify the filter bandwidth and frequency directly.  If you were to graph the filter responses in the same way that Adam does with the RGB filters in his video, you would not get large areas under the curve.  You'd get very narrow notches, almost like vertical lines with 3nm filters.  But the math works the same.  SPCC can use the data to accurately determine the contribution of each filter.

  
  

  Where it diverges from calibrating RGB images, is that you cannot use a white reference to calibrate the channels.   This follows from what I mentioned above about being unable to accurately calibrate star colors when the green and blue channels are identical.  Instead of selecting a stellar spectral class, or a galaxy, as a white reference, you select "Photon Flux" as the reference.  I believe that this tells SPCC "Don't try to do a white balance.  Just make it so that the relative contributions from each filter are accurate."  The documentation calls this out:

  
  

  Since we want an image where the intensity of the nebula in the different filters corresponds to the actual proportions between the emission lines, the white reference in this mode should be set to Photon Flux.

  
  

  I am notoriously bad at reading documentation, and like to experiment and analyze to figure out how things work.  One of the dangers is that I sometimes miss subtle points, and this happened here.  When I said above that I didn't do it correctly before, this is what happened.  I used "Average Spiral Galaxy", instead of "Photon Flux".

  
  

  And here is another quote from the documentation, where it says, pretty directly, that the emission ratios are accurately calibrated with this process:

  
  

  In some cases, the difference from the original image can be subtle. Still, the critical point here is that we know that the proportion between the emission lines in the processed image corresponds to reality.

  
  

  I'm including the report that SPCC generated when I went back and recalibrated the data (correctly selecting Photon Flux this time).  The thing to notice is that the R/G flux shows that the stars in my image fit the function pretty well.  Also interesting in this case, is that there is no data in the B/G flux graph.  This is because it's redundant.  The B and G channels are identical in my data.  The SPCC process seems to recognize that the B and G channels are the same, by virtue of the fact that I have them both set to 500.70nm and 3nm width.  As a quick test, I changed the value of the blue filter to a different wavelenth.  That caused SPCC to populate the B/G graph, but all the ratios were 1 (since I still used identical data for the B and G channels).

  
  

  So why do I assert that the color calibration for the emission nebulae is correct?

  
  

  The wavelength of Ha puts it squarely in the red channel, so that's pretty straightforward.  The wavelength emitted by OIII is 500.70nm, which puts it almost exactly in the middle between green and blue.  By putting an identical copy of my OIII data into each of the green and blue channels, so combining those two channels should accurately represent what OIII would look like if it were bright enough to trigger color detection in your eyes.

  
  

  I still have some questions, and planned experiments around what would happen with a filter whose wavelength were, say, 25% of the way between green and blue.  I don't have time today to dig into that, but I wanted to mention that I am thinking about it.

  
  

  So is this a useful thing or not?

  
  

  In my opinion, this is quite useful.  Unless I've missed something, I believe that the color in the emission areas of the image are accurate.  Since there are now tools to easily remove the stars, it opens up the possibility to combine accurately calibrated RGB stars with the calibrated emission nebulae.  That is actually where I'm planning to go with this data.  In addition to Ha and OIII, I also collected R, G, B and SII.  So it's a good set of data to experiment with for SPCC.  I'll probably also try SHO with narrow band mode.

  
  

  So now it's a good time to mention SHO.  The documentation for SPCC says that it will always give a poor rendition of the Hubble Palette (which is synonymous with SHO).  I think that this wording puts people off a bit.  I believe that SPCC will accurately calibrate SHO data.  The problem is, that's not what we want.

  
  

  Accurately calibrated SHO data will almost always be heavily green tinted.  This is because Ha is generally much stronger than OIII or SII.  In order to get the golds, yellows and reds that are so much associated with SHO, you need to boost OIII and SII far beyond their natural ratios.

  
  

  So that's my current thinking on this.  I have more playing to do.  Also, I welcome comments or constructive criticism.

  
  

  Thanks,

  -Wade

  
  

  

  

  SPCC_Report.jpg

  514 x 971 - 85K
• BlockHead

  December 2022

  Nice job on all of thinking.

  Just a quick point- the issue of where emission lines fall in terms of a channel assignment is a tricky one. This is especially true for scientific filters (like Sloan) where the Ha line is actually in a "green" filter!

  
  

  -the Blockhead