From: Alan Friedman <alan@greatarrow.com>

Date: July 30, 2008 3:12:35 PM MDT

To: Astro_IIDC@yahoogroups.com

Subject: Re: [Astro_IIDC] Choosing the proper confidence level percentage

As Milton has said, the seeing as displayed in the video file is quite good - nice to have Jupiter at a decent altitude.

A couple of observations on the reference you posted:

There is a minimum number of frames needed to get a decent result on any subject - this varies based on the signal/noise of the subject which will vary depending on the brightness of the object and the need to amplify the signal with gain. For a subject of ample brightness (the sun and the moon at a short EFL a single frame will suffice. These images are mosaics constructed of single DMK video frames:

http://www.avertedimagination.com/img_pages/fulldisk_072808.html

http://www.avertedimagination.com/img_pages/moonrise022407.html

For jupiter/saturn with my 10" scope, a stack of approximately 200 frames is my target and I will generally look to adjust the CI to deliver this target (+- 20% or so.) The sharpness levels must also be adjusted - using the same settings regardless of the frame count will allow the noise levels to jump all over as displayed in your sample images. A flat line setting as in your example (using the same levels at each pixel radius) applies way too much contrast enhancement to the image IMO. I would experiment with a curve that sets the largest radius very close to 0 and increases to the right as the pixel radius decreases. I might be up as high as your settings at radius 1.5. Often I find that a high setting at radius 1.0 adds a fine noise pattern - I often set the 1.0 slider back down to half the setting at 1.5.

There is a ying/yang sort of balance between  >the number of frames selected with the CI setting/ loss of sharpness by lowering the CI and adding more frames/ and the ability to get it back by sharpening further because you have a smoother stack using more frames< that makes the correct CI level somewhat hard to establish. By varying the processing (something not done in your example) you can achieve a very similar result over a wide range of frame counts. Using a gaussian blur to the stack will increase the signal quality and allow more sharpening to be applied before the noise level gets objectionable in much the same way as adding frames to the stack.

I would choose the planetary gross alignment method for Jupiter rather than lunar/solar. I would also use the manual selection for alignment regions with the 64 pixel box size. Your image has enough contrast that you can eke out additional detail with MAP technique, choosing the GRS and sections of the NEB with festoons for alignment. I have not had success sharpening the frames prior to stacking and do not recommend it. You want to reduce noise as much as possible by stacking prior to sharpening, else you will be amplifying the noise along with the fine detail in your image.

Thanks for taking the time to share your techniques with this great tutorial - it is very helpful for discussion.

clear skies,

Alan

On Jul 29, 2008, at 12:01 PM, doodlebun wrote:

Jupiter imaging requires the proper setting of confidence levels or
the result will be noisy and oversharped. I have created a set of
confidence levels from the same movie taken under good seeing. As I
seemed to prove to myself 50% is the way to go. Which percentage would
you choose? look at:

http://web.mac.com/davidbleser/AncientDocuments/CONFIDENCE_LEVELS.html