White-balancing Optolong LRGB filters with Altair 290M

Anything that doesn't fit into any of the other forums
Post Reply
User avatar
oopfan
Posts: 1321
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

White-balancing Optolong LRGB filters with Altair 290M

#1

Post by oopfan »

I wanted to try my hand at color astrophotography using LRGB filters and my monochrome camera. White-balancing is the process of determining the exposure ratio for each color channel so that an image can be rendered accurately and true-to-life. For my filters and camera I determined that the proper balance is 1.0 unit of Red, 1.01 units of Green, and 1.2 units of Blue. Here is how I did it.

A common method is carried out at the telescope and then followed up at the computer by measuring the flux of a Sun-like star in red, green, and blue using specialized software. Note that ZWO sells an LRGB filter set that is specifically balanced for the ASI1600MM therefore eliminating the need to white-balance (this is my understanding.) Since I don't have that camera I decided upon Optolong filters.

Al Kelly has a great tutorial for white-balancing at the telescope here http://kellysky.net/White%20Balancing%2 ... ilters.pdf Unfortunately for me the weather is terrible and has been for too long. I needed a method that I could perform entirely with mathematics. I couldn't find anything online so I invented it:

I began by taking a snapshot of the filters' spectral response from the retailer's website (see attachment "Optolong LRGB Spectrum.jpg".) I converted it to FITS file format using PIPP and then opened FITS Liberator. I placed the cursor over the curves and entered the (x, y) coordinates into Excel every five pixels in the x-axis. Finally I sampled the 'x' and 'y' minimum and maximum values of the grid so that I could convert the (x, y) coordinates to 'wavelength' and 'percent transmittance'. I did the same for the Quantum Efficiency graph from the camera manufacturer (see attachment 'QE of Altair 290M.jpg')

Lastly I needed a white-light source. Our Sun is said to be the standard. I found an online calculator that enabled me to graph and to download to a CSV file the radiance of a Black Body at 5772 degrees Kelvin, the temperature of the Sun's chromosphere (see attachment 'Solar Luminosity vs Wavelength.jpg')

Getting those data points was 99% of the work. The last step was to multiply three numbers for each wavelength: the filter's transmittance percent, the Quantum Efficiency, and the relative radiance of the Sun. Finally for each filter I summed those calculated values over all wavelengths, essentially taking the "area under the curve". Those values are 18.8 for Red, 18.6 for Green, and 15.7 for Blue. With simple rearrangement I converted it to 1.0 units of Red, 1.01 units of Green, and 1.2 units of Blue, in order to achieve a balance of 1:1:1. A practical example might be 30 minutes in Red, 30.3 minutes in Green, and 36 minutes in Blue.

(See next post for Part 2)...

Brian
Attachments
Optolong LRGB Spectrum.jpg
Optolong LRGB Spectrum.jpg (89.82 KiB) Viewed 2387 times
QE of Altair 290M.jpg
QE of Altair 290M.jpg (30.51 KiB) Viewed 2387 times
Solar Luminosity vs Wavelength.jpg
Solar Luminosity vs Wavelength.jpg (47.84 KiB) Viewed 2387 times
User avatar
oopfan
Posts: 1321
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: White-balancing Optolong LRGB filters with Altair 290M

#2

Post by oopfan »

Part 2:

I was eager to try it out. One night I struggled for over four hours with broken clouds to obtain one minute of imaging data. Using C2A I discovered a tight grouping of four relatively bright stars in Hercules having different spectral types: B, G, K, and M. I took 16 seconds for each of Luminance, Red, Green, and Blue. For fear of losing the opportunity to collect all data I chose not to refocus after filter changes. I took Darks and Bias frames but no Flats. Less than ideal but at least I got something to work with.

Processing in Deep Sky Stacker was challenging. Usually my frames have dozens of stars useful for alignment purposes but in this circumstance DSS could only find the bare minimum of 8 stars for the Luminance and Red filters but not enough stars for Green and Blue. So I was forced to use "Comet Mode" for Green and Blue. This required me to select one star, the same star, on each frame. So again, less than ideal. I wasn't too hopeful at this point.

I used StarTools to complete post-processing. By default StarTools uses the ratio of 1:1:1 in its LRGB Module. I don't have that image to show you but believe me that the colors were truly awful. Thankfully StarTools allows me to enter a new ratio, so I entered 1:1.01:1.2. Like magic I got a life-like rendition of the true colors of the stars (see attachment).

To obtain this image I zoomed into each star and took a snapshot and pasted it into Paint. The top row shows accurate colors. Notice the left-hand column says "Green: Cap to Brown." I discovered that by watching a YouTube video of a person processing a globular cluster. I went to the StarTools website to see if I could find an explanation. All that I could find was the statement that "not too many objects are green, so we need to subdue the green channel." Well, I don't know about you but I am not completely satisfied with that explanation. I am beginning to think that it has more to do with the output media. Perhaps color monitors require "cap to brown" and printers require "cap to yellow". Any ideas?

Brian
Attachments
Star Color Comparison.jpg
Star Color Comparison.jpg (60.96 KiB) Viewed 2385 times
mAnKiNd
Posts: 308
Joined: Sun Jun 11, 2017 4:55 pm
Location: Long Island, NY
Contact:

Re: White-balancing Optolong LRGB filters with Altair 290M

#3

Post by mAnKiNd »

Awesome!
User avatar
oopfan
Posts: 1321
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: White-balancing Optolong LRGB filters with Altair 290M

#4

Post by oopfan »

Thanks, Minos
User avatar
oopfan
Posts: 1321
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: White-balancing Optolong LRGB filters with Altair 290M

#5

Post by oopfan »

I decided to try out the empirical approach to find white balance as described here:

http://kellysky.net/White%20Balancing%2 ... ilters.pdf

Last night I took 100x 2s frames with each RGB filter of a magnitude 8.23 G2V star in Ursa Major: HIP 60268. I chose the best 10 frames for each filter, calibrated them, measured the flux using AstroImageJ, and adjusted for atmospheric extinction.

Here are the ratio of exposures needed to achieve white balance:

R: 1.0, G: 1.0, B: 1.4

The ratios that I've been using since last June when I purchased the filters are:

R: 1.0, G: 1.0, B: 1.2

I derived those ratios through mathematical reduction of the relative transmittance of the filters, the QE of the sensor, and the black body radiation of a star like the Sun. Apparently there are other factors that I did not account for, particularly in the blue channel.

I've always suspected that my blue channel data was too noisy. The difference is not huge but the new ratios allow the blue frames to "bake" a little bit longer to increase the SNR.

I was using these exposures:

R: 30s, G: 30s, B: 36s

I should have been using these:

R: 30s, G: 30s, B: 42s

Lesson learned: Always back up theory with evidence.

Brian
User avatar
turfpit
Posts: 1779
Joined: Mon Feb 13, 2017 8:13 pm
Location: UK
Contact:

Re: White-balancing Optolong LRGB filters with Altair 290M

#6

Post by turfpit »

Now that is thorough Brian.

Good to hear from you Minos.

Dave
Post Reply