Driveway Astronomy

 

 

NGC281: Variations on adding Ha and RGB data

After reading much discussion on the Starlight Xpress Yahoo group, the PixInsight Forum and the AstroMart Deep Sky Imaging group about best ways to add Ha data to an RGB image, I decided to try various methods. I should also cite Robert Gendler's article on the subject as a must read. (All processing done with PixInsight and AstroArt. Camera: Starlight Xpress MX716. Telescope: William Optics ZenithStar 66 SD. Schuler 10nm Ha filter, Astronomic Type II RGB filters.)

My original processing used a straight Ha image as the Luminance, then added RGB to that for color. The result to my eyes was  rather "thin" as the Ha Luminance definitely preserved detail, but overwhelmed the "richer" data of the RGB. The result had color, but still a very washed-out look in my opinion. It appears as a "colorized" Ha as opposed to a true color image. (Note: there was a large dust mote on my chip, I believe, that is visible on the top of many of the images that was hard to flatten. I might have clipped a bit much on the black end to overcome it.)

Original HaRGB Processing 104/24/30/30 (some background noise reduction done in PixInsight)

Note the big difference between the Ha and the other color channel masters as shown below, both in background and star size.

Ha 104 Minutes

As expected, the color channels show the nebula faintly with the Red being the strongest. Remember, these were fairly short total exposures of  6 minute subs. I tried to stretch them fairly equally. No other postprocessing was done.

R 24 Minutes

G 30 Minutes

B 30 Minutes

Adding the Ha and the R does seem to make a nice compromise. This is a concensus recommendation on the forums to replace the pure Red channel. Certainly additional attempts at different percent conributions from each channel could be pursued. Other possible combination techniques were brought to my attention by members of the Pixinsight Forum and can be performed in the Pixinsight math function.

Ha and R added together (1:1 ratio)

The bigger issue is what to use for the Luminance if not the Ha? Two possibilities exist: a pure luminance, or a mix of luminance and Ha in some proportion. (Again, given the meager color data here, a fair test doesn't seem possible.) Below are the two version of the new Luminance possibilties. But I do think that the Ha+L(R+G+B) makes a nice compromise between improved SNR and enhanced detail over the straight L while bringing it more in line with the RGB channels that it will be combined with. Even with the short times involved in these images, there is enough data to work with so that even finer detail can be brought out through various filters, wavelets, deconvolutions or whatever other techniques are called for. It would seem that it would take some long exposures indeed to get the same workable data without adding the Ha.

Luminance made up of R+G+B 24/30/30 added in Pixinsight

Luminance of 50%Ha+50%combinedRGB added in Pixinsight

Finally the result of all this messing around, an image that used the Ha twice: added to both the luminance and the R channel of the RGB. To make comparison easy without all that scrolling, the original HaRGB is below the L((Ha+R)HaRGB. I can't vouch that my processing was the same in both cases, but overall I do feel that the color of the nebula is richer and star color is more pronounced. In fact, there seems to be far more stars overall. Not surprising as the Ha would miss stars. The downside is definitely more noise and loss of fine detail along with some star bloat that was hard to tame. (ACDNR noise reduction was used in both images below.) While there was insufficient data in all channels to do this great nebula justice, I can see where the method used in the top image below has real potential. I do plan to shoot luminance in addition to Ha and RGB in the future on emission nebula. (Addendum: I added a revised HaRGB at the bottom with more saturation at the suggestion of a SX Forum member...definitely richer color.)

L(Ha+R)HaRGB

And again, for comparison, the original HaRGB image:

Above image with more saturation (thanks for the suggestion, Adam)

A reference on the PixInsight Forum to a presentation by Chris Schurr sent me back to the computer once again (thanks, Jack!). His paper talks about adding Ha to each of the RGB channels for more detail in galaxies after first subtracting the red channel from the Ha for a "pure" Ha image. I decided to try with this nebula. This "pure" Ha was added to the RGB channels 100%, 5% and 10% based on his recommendations. I also added (yet!) one more image: the same Luminance of Ha+R with a straight RGB, no Ha to any of the channels. I added equal saturation to both of the following images.

(Ha+RGB)RHaGHaBHa

(Ha+RGB)RGB

Here's a summary of the final four reprocesses as a screen capture from AstroArt. It's a toss-up to me on the two on the right, both of which have Ha added to the Lum and one or more channels of the RGB. If forced I'd probably take the upper right with the Ha in all channels.