WinJupos Comparison

 

Having tried a series of 5-11 sets of stacked images de-rotated in WinJupos, I think it's fair to say that you really need like 15-16 sets of stacked images to improve the image.  While the de-rotated image isn't as good as a really good 3 min RGB stack from the previous night, it's a definite improvement.  

I should clarify by saying that not only are we reducing noise in the WinJupos image, but we are getting smaller scale detail which is not available in the single stacked image.

My slight frustration with WinJupos is that most of my attempts have only cleaned up the noise but haven't done much to reveal more small-scale detail.  Had I done this comparison back in August, I would have then instituted a policy (lol) where I always collected 15+ runs of any planets.  For Mars, that's like 45-60 minutes of data.  For Jupiter, that's 45 minutes at most.  

I have derotated a few Jupiter runs, but since I didn't know how many runs would make a difference, I was just doing like 5-7 runs on Jupiter, taking a long break (an hour or longer), then doing another 5-7.  

Chrisotpher Go has mentioned that he does like 5-7 runs in excellent seeing and many more in comparatively worse seeing.  

Also, it's probably worth mentioning (again) that you really need to pay attention to how WinJupos is aligning the images.   Sometimes if you don't pay attention, WinJupos will move the planet slightly from one stack to another which I'm sure is compromising the nature of the stacking.

There are some weird color issues that I couldn't quite fix.   When I did the first Registax wavelet sharpening, I sorta pushed it and in the process, I began to see odd colors (bluish-red).  I didn't really pay attention but then in the final sharpening, I had to deal with it.  And after fifteen minutes of goofing around with the colors, I was too burnt out to go further.  Another color related-issue is that the atmospheric blue "halo" was more orangish.  Again, I can fix this, but I just thought it was interesting that pushing wavelets too hard and de-rotating in WinJupos gave me this "false" color.