Upload & Sell: On
| p.1 #20 · Hi iso NR technique for night sky’s. |
Well, this probably won't be very helpful but here is what I was able to do with some limited time this evening.
First, I'll just repeat again that I think you might want to try increasing exposure. You can always pull the exposure back down in post so it looks like you want but by exposing more to the right, you will get more tonal values out of the noise floor. You may run some risk of clipping the stars and losing some subtle color in them but as it is now you are probably losing that subtle color anyway with your noise reduction.
Here is just a reminder of how dark this image is using linear gamma of 1, the "natural" gamma of the raw file. Those darkest 2 or 3 stops of the image have many fewer tonal levels than the brightest stops. I'm sure you've probably read several of the "Expose to the Right" articles on the web, Ben, so I won't repeat all that here but just take this as a reminder. I think it could really help this type of photography and it may be the reason you are seeing different results from some of the sites you referenced.
Here is a comparison of different Noise Reduction packages. It is not perfect by a LONG shot but maybe it will give you a rough idea. Comparing NR packages is much like comparing raw developers - there is no common baseline for types of settings and setting amounts so it's almost impossible to do an apples-to-apples comparison. It's also easy to favor the package you use most because you are more familiar with its controls and how it works. In my case, that would probably be Imagenomic's Noiseware although these days I really use Lightroom most of the time.
I tried to select 3 representative patches from your image: dark blue sky, clouds/sky/mountain, and some foreground trees/brush. Everything was ran through ACR for raw development but ACR noise reduction and sharpening were set to zero for all tests except the ACR NR sample.
All of these packages have a variety of settings that can be tweaked and then you can do selective masking like you're looking at with BIF so it is almost certain that none of the samples above reflect the best of what could be accomplished with each package. I obviously haven't made the case but I still feel that you are going to be able to achieve better results with a specialized package over ACR/LR, particularly when you start adding sharpening but I could be wrong. Perhaps if someone wants to give LR or ACR their best shot, I will do a "best shot" approach with one of the other packages and we can compare. I may not have time to do that until the weekend though.