Dark Horse Observatory
Synthetic Luminance Investigation - Browse Images - 5
You Are Here > Astrophotography Studies > Synthetic Luminance Investigation
This is an attempt to recreate the white light (luminance channel by combining the RGB exposures to recreate the luminance signal. It has obvious benefits of saving the time required to take additional luminance exposures. I read several imagers offering opinions that it enhanced their RBG images. Since I was interested in doing star clusters, this had some appeal to me. I determined to investigate it.
There are several ways to do create a synthetic luminance channel. One way that is quite easy is to create the RGB image in a program such as Maxim DL or CCDStack or other similar product. Afterwards, the RGB image is brought into Adobe Photoshop. Under the Image menu, you can select Mode and then select Grayscale. This reduces all the color data to a simple luminance image.
For this study I used images that I had taken of IC 434, B33, and NGC 2030. The subframes were taken with 5 minute duration and were all binned 1x1. There were 12 luminance images and 6 subframes for each of the RGB. This gave a total exposure of L:R:G:B = 60:30:30:30 minutes. These were combined into mean exposures for each channel and then the RBG were combined in CCDStack. Finally they were taken into PhotoShop and after stretching and level adjustments a luminousity combine of the L channel with the RBG channels was performed.
The final LRGB, although decent, appeared to need additional exposure time; however the weather was not cooperative. This when I became interested in investigating whether a synthetic luminance image could improve the image. As much as is humanly possible I tried to process comparative images identically and end up with similar histograms.
While the results were not useful to me due to decreased color saturation and increased color noise, I decided to look at the results more closely.
I leave the images for you own conclusions. One observation I made was that the color noise was introduced into the synthetic luminance and the combination did not appear to reduce the noise much if a t all. I can't say whether increase RGB exposures would have worked better by virtue of better S:N. This would be an additional chapter to this investigation. If the clouds ever go away and if I have an imaging lull I may try this in the future.