A possible solution to mitigate chromatic aberration

Somewhere to share your expertise in using SharpCap
Post Reply
User avatar
oopfan
Posts: 722
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

A possible solution to mitigate chromatic aberration

Post by oopfan » Thu Mar 05, 2020 2:24 pm

GRGB - A solution for fat blue stars perhaps.jpg
GRGB - A solution for fat blue stars perhaps.jpg (199.75 KiB) Viewed 1077 times
One of the benefits of a monochrome camera is having the ability to explore creative solutions to systemic problems. My William Optics ZenithStar 20th Anniversary Edition OTA has serious problems focusing blue light. You can see it in the right-side panel of the attached image. M35 is a cluster of predominantly blue stars. It might just be me but I don't mind the overall width of blue stars, I mind the large central disc. What can be done about it?

Up until now I've practiced standard LRGB astrophotography: bin1 mode using the Luminance filter, bin2 mode for the Color filters. You can see it on the left-side panel of the attached image of a blue star. Notice how the width of the star in blue light is wider than in red and green. One reason is that the blue star emits more energy in blue and less in red and green. The other reason is due to my refractor's poor performance focusing blue light.

Turn your attention to the top image labeled "L" for Luminance. Notice that the width of the star is nearly identical to the blue star! Why is that? It's because it lets in all visible light from deep red to deep blue. The image is being affected by the poor focusing performance I mentioned.

In the coming weeks I am going to experiment with replacing the Luminance filter with the Green filter. As results roll in I'll append it to this topic.

Brian

mAnKiNd
Posts: 307
Joined: Sun Jun 11, 2017 4:55 pm
Location: Long Island, NY
Contact:

Re: A possible solution to mitigate chromatic aberration

Post by mAnKiNd » Thu Mar 05, 2020 7:39 pm

I've come across these headaches in the past, so I will share some thoughts I've had.

Consider using a baader fringe killer/semi-apo/contrast booster as your luminance for future acquisitions, because although green is the luminance detector for the human eye, the above filters certainly contain more wavelength data than just green.

Another trick might be to convolute red and green to match blue, sacrificing sharpness and tight stars in the process. This can be slightly mitigated by later deconvolution and star reduction techniques on the LRGB image as a whole.

Good luck, I look forward to seeing what you come up with GRGB.

Minos

User avatar
oopfan
Posts: 722
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: A possible solution to mitigate chromatic aberration

Post by oopfan » Thu Mar 05, 2020 8:39 pm

Some encouraging preliminary results!
LRGB-GRGB-initial-comparison.jpg
LRGB-GRGB-initial-comparison.jpg (205.03 KiB) Viewed 1051 times
In order to get a quick look at what I might expect I used my existing M35 image data that I captured last month in LRGB, and copied the 300 frames captured with the Green filter into a new folder called GasL (stands for Green as Luminance). Then I reprocessed using APP using the GasL frames. The final step was to combine the stacks for GRGB to make the final color image.

In the attachment the GRGB image on the right is poor quality because it is bin2 resolution. The LRGB image on the left was scaled up by APP to bin1 because the L frames were capture in bin1. The GasL frames are bin2. Nevertheless it supports my hypothesis that capturing luminance in green will result in more compact stars. Notice also that the blue halo still extends as far as it does in the LRGB image but the fact that the central disc is tighter makes for a more pleasing image.

Next step is to image a new star cluster. Unfortunately M35 is behind trees now. I'll have to find a substitute having plenty of blue stars and some reds. If I can find one.

Brian

User avatar
oopfan
Posts: 722
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: A possible solution to mitigate chromatic aberration

Post by oopfan » Thu Mar 05, 2020 9:42 pm

Hi Minos,

I see your point about the Green filter not picking up Ha regions in galaxies. Certainly the Red filter will pick it up but only having bin2 detail. I think that GRGB will work fine for star clusters where the stars themselves are the focus of attention. For all other targets then LRGB is best.

Thanks,
Brian

User avatar
oopfan
Posts: 722
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: A possible solution to mitigate chromatic aberration

Post by oopfan » Tue Mar 10, 2020 1:46 am

All right, I had a challenging session last night (98% Moon and high clouds) but managed to get in one hour of data. I chose a loose cluster of stars in Canes Venatici. Red stars and Sun-like stars, unfortunately no blue stars.

I set out to test the possibility of replacing the luminance filter with green. After processing the images I discovered that it is not a viable solution. Star colors are bland and uninteresting. In a later post I'll talk about that. Thankfully all was not lost. Last night I squeezed in a stack of luminance frames so that I could compare and contrast LRGB with GRGB.

I processed the LRGB and once again as usual my stars suffered from the blue fringe around the bright stars. I've read that Photoshop has a fringe killer algorithm but I don't use PS only AstroPixelProcessor (APP). I visited APP's forum and searched for "fringe". A post popped up between a user and Mabula Haverkamp, APP's architect and programmer:

https://www.astropixelprocessor.com/com ... #post-3227

There is a checkbox in the Calibration tab called "Align Channels" that attempts to adjust the diameter of star discs to approximate the performance of an APO. My refractor does a poor job of focusing blue. Red and green are better but green is the best.

I chose three stars that had slightly different temperatures: a Sun-like star, a slightly redder star, and a much redder star. (Color can be determined by subtracting the "V" magnitude from the "B" magnitude. "B" and "V" magnitudes are available in many planetarium software packages.) I meticulously "white balance" my RGB filters so there is no need to rely on APP's color calibration tool. All of my images are perfectly color balanced from the beginning.

Columns A, B, C, and D show the result of applying Mabula's recommendation for healing chromatic aberration. The differences are minor from one column to the next but I think there is a significant improvement between column A and D.

One thing I want to experiment with is using bin1 for all filters. Right now I am using bin2 for RGB and bin1 for Luminance. My camera's pixels are a bit too large for the telescope's focal length so I am undersampling a little at bin1 but really undersampling at bin2. My exposures will be longer at bin2 but not too bad since stars are bright subjects.

Brian
Attachments
APPs Fringe Killer.png
APPs Fringe Killer.png (30.15 KiB) Viewed 958 times

AussieHowie
Posts: 25
Joined: Fri Aug 11, 2017 12:23 am

Re: A possible solution to mitigate chromatic aberration

Post by AussieHowie » Tue Mar 10, 2020 11:22 am

Yellow #12 has long been touted to be equal or even better than most fringe killer filters. And much less cost. Be sure to include that in your tests. Several forums have had folk test #8 and #12 yellows with and without UV/IR cut filters co-fitted and comparisons made against high end fringe killers.

User avatar
oopfan
Posts: 722
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: A possible solution to mitigate chromatic aberration

Post by oopfan » Tue Mar 10, 2020 11:26 am

As promised, here are the results of LRGB vs GRGB:
LRGB vs GRGB.png
LRGB vs GRGB.png (18.75 KiB) Viewed 941 times
At first glance they look quite similar. The diameter of the GRGB stars are a bit smaller which I see as a plus. The problem is with the red star in the middle. It has become paler and even smaller in size. Why is this?

Stars do not emit all wavelengths equally. It is somewhat like a Bell Curve where a lot of energy is concentrated in and around a particular wavelength. For hot stars the peak wavelength is in the blue. For cool stars the peak wavelength is in the red.

A standard Luminance filter passes all wavelengths from red to blue (approximately 400-700nm). So the luminance filter accurately captures the brightness of both hot blue stars and cool red stars.

A Green filter, on the other hand, passes only green wavelengths from 500-600nm. Hot blue stars and cool red stars have peak wavelengths outside of that bandpass. If I use the Green filter as Luminance then those stars will appear less bright than a Sun-like star whose peak is within the green bandpass.

In the attachment I am showing a red star having a Color Index of 0.941. That is Spectral Class K which is one class redder than a G star like our Sun. There are two class even redder: M and C. A Carbon star has a Color Index greater than 2.0. The Green filter will capture some of the energy of a Carbon star but very little. It will appear quite dim in a GRGB print. The same applies for hot blue Spectral Class O stars.

In conclusion, GRGB is not a viable solution if you are interested in accurately depicting the color as well as the brightness of a wide range of stars however it does do a good job of mitigating chromatic aberration.

Lesson learned: Beware of marketers who claim "APO-like" performance. Buy a true APO.

Brian

User avatar
oopfan
Posts: 722
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: A possible solution to mitigate chromatic aberration

Post by oopfan » Tue Mar 10, 2020 11:38 am

AussieHowie,

Thank you for suggesting Yellow #12. Yellow crossed my mind yesterday but there will still be a problem with accurately capturing the brightness of hot blue stars since it falls outside of the Yellow bandpass. The lesson is "buy a true APO". I might also give a reflector a try.

Brian

Post Reply