|
Thanks for the comments.
Robert, I'm following most of the same steps but i'm getting a better understanding of how to tweak the most out of each of them without overprocessing, and i'm getting much better at adjusting the colour balance so I don't end up with a yellow Mars, like I did a month or so ago.
My usual processing run is:
1) (Optional step) - Using virtual dub, save the avi as individual bmp frames, and then use ppmcentre to centre the planet in each frame (and reduce the frame size to 400x400 while i'm at it). This makes it easier for registax to align the frames, makes it easier to do a 2-stage alignment run in registax, and the reduced frame size makes the whole processing in registax quicker.
2) Load the bmp stack (or avi) in registax, find a fairly sharp frame (without being too pedantic) and align. I use the "Gradient" estimator, and I don't pay any attention to what the % is set to.
3) If the align curve doesn't follow the right pattern (heading up at the far end), then I scroll the slider all the way to the left to pick the sharpest frame registax found, and then click on the centre of the planet and click align again. This usually gets the shape of the curve I want.
4) I then just drag the slider left or right to reject the poor quality frames that look blurry or that I know won't add anything to the stack. That's why I ignore the % value cause I adjust how many frames I "Limit" by manually.
5) In the optimise tab, I "create reference frame", and do a slight wavelet adjustment (only layer 3 -> 5.8).
6) Optimise, usually takes 3 passes.
7) Stack tab - press "Stackgraph" and slider the vertical slider down until I think enough frames are being stacked. How many frames I stack, is a choice I make each time depending on a) How good each frame is, b) How many frames in total I have to choose from.
If the seeing was really good, and I rejected a lot of bad frames already during the "Limit" phase, then I might stack 75% of the frames that are left, by dragging the slider down and cutting off the peaks. If I had 1000 frames, I might stack 700.
If the seeing was only average or bad, I'll reject even more frames by dragging the slider lower and only stacking a few hundred, sometimes even less.
Obviously if you had gain high or transparency was bad, the image can be quite grainy if you don't stack enough frames, but i'd rather stack fewer good frames, than more bad ones which will blur the features.
8) Go to wavelets, click "RGB Align" and hit "Estimate".
9) Adjust wavelets, I usually go from Layer 6 down to Layer 3, with Layer 6 being the highest (depending on image quality/seeing), sometimes around 50-60, then less for Layer 5 etc, until layer 3 .
10) Save as TIFF
11) Bring the image into AstraImage.
12) Split into RGB planes
13) Then do a LR or ME deconvolution on each colour, usually a value of 3/1.3 depending on the quality of the image and how many frames I stacked. If I didn't stack many frames (because of bad seeing), then it won't stand up for much deconvolution without getting hard rings around the edge.
14) Recombine the RGB planes, sometimes adjusting the colour weight to give a better look. Screenshot the result and save as TIFF.
15) Take the image into NeatImage and do some noise reduction/sharpening. Sometimes this step doesn't help, it depends on how many frames I stacked and the quality of the image. Just have a play.
16) Take the image back to Photoshop, do an unsharp mask (if needed), and adjust the colour balance, contrast, add text etc.
And that's it!
HTH's someone.
|