Posted 06 January 2013 - 11:17 AM
I think I'm largely echoing zAmbonii here, but my "normal" regime starts with the capture itself. If the seeing is 2/5 or less, I don't usually even bother trying for hi-res. On those nights where I can get a sharp definition on the NEB or SEB, or can easily see the GRS hollow, I try to capture as much data as I can.
Like zAmbonii, I also use FireCapture with its cutout, histogram and autoguiding functionality. If I'm trying for a longer capture in good seeing, I'll also enable derotation, but I'm unconvinced this helps much. I do try for a large image scale, though - somewhere in the F20-F30 range.
Then I'll stack in Autostakkert. I estimate the level of detail and on a good night will set the Noise Robust to 2. Then I'll set my stack count to only stack the frames above a 70% quality level (in an ideal world). Some nights there aren't too many like that, so I have to compromise to bring in as many frames as necessary to have a positive enough impact on the S/N ratio. I really try to not have less than about 800 frames per stack.
Once I have a stack (it usually takes a few iterations to get an optimal one), I move onto Registax, where I auto-balance the colors if doing OSC, sometimes stretch the histogram a little and then apply sharpening. On a really good capture, I only diddle with the top wavelet setting, typically moving the slider along to an initial level of 80%, then increasing the Sharpen until noise becomes too evident, at which point I increase the Denoise, then after that it's a matter of adjusting all three to optimize the detail without too much intrusive noise.
If the seeing was less than good, I sometimes skip the top slider altogether as it has no impact. In that case, I apply the same sort of adjustments to the 2nd-to-top slider. On very rare occasions, I've seen a benefit from using more than one slider. Your data above, funnily enough, was one of these occasions, just like zAmbonii found.
Once the detail is maximized in Registax6, I save as a TIF and move onto either PS, MaximDL (or GiMP, if I'm not out in my observatory). Since GiMP is free, I'll describe the approximate process for that one.
Using the G'Mic plugin, I play with the Sharpen (deconvolution) settings to bring out as much detail with only a little noise. Then using the Smooth (anisotropic) adjustment in the same plugin, I smooth out the noise just a tad, trying not to take away any of the fresh detail I just brought out.
Then I unsharp mask at something like (3,0.6) followed by a Gaussian blur at 1.0, then another unsharp mask at (1,0.5). I also execute any slight tweaks in gamma or color here.
On my own images, my last move is to rescale down (I prefer to capture at large scales), followed by a final unsharp mask at (1,0.5). Tada.